Plans Jet Engine Pulsejet Book

Page 1

Second revision (May, 2004)

Copyright Notice: This book is copyright 2004 to Bruce Simpson and all commercial rights are reserved. Anyone discovering a violation of these terms is requested to contact the author through the webpage at http://aardvark.co.nz/contact/

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 2

Foreword Modern jet engines, like the ones found on large passenger aircraft or military fighters are incredibly complex and expensive to make. Built from thousands of individual parts, many of which are made from exotic alloys like titanium and Inconel, these engines are a masterpiece of modern engineering. But what if I was to tell you that there is at least one type of jet engine that has been around for almost 100 years, can be built out of plain old steel using simple tools, and in some cases has no moving parts at all? Well it’s true and I am, of course, talking about the pulsejet. In this book I’ll do my best to explain how these engines work, how to build them, how to improve on the basic designs and how they can be used to power all manner of vehicles from model airplanes to gokarts. In an attempt to make the information contained in this book accessible to the widest range of people, I’ve taken a few liberties in explaining some of the more complex concepts. I’m sure there will be physicists, engineers and mathematicians who will throw up their arms in disgust when they look at how I’ve presented some of this information. My justification for this is that the majority of readers are probably not equipped nor interested in wading through pages of complex mathematical formulas in order to understand some aspect of a pulsejet’s operation. In such cases, I believe, it’s better to replace all this complexity with a simple analogy or basic calculation that hopefully anyone can follow. Another are of contention is the very explanation of how a pulsejet works. Although there is some consensus on the basic mechanisms behind the operation of a pulsejet, much of the detail is a topic of hot debate and disagreement. Wherever possible I’ve tried to present all sides to an argument but obviously I favor my own opinions which are based on years of empirical observation and experimentation.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 3

A Note About The Author Who is Bruce Simpson? Well I’m a middle-aged guy who has always had a strong interest in technology and things that go bang. From an early age you’d find me out in the garage playing with my chemistry set, building all manner of weird and wonderful devices from old discarded radios, or just reading books about science. Since the age of about seven, I’ve also been an avid builder of model airplanes, mostly of my own design. Over the years I’ve created all manner of odd-ball flying creations including flying wings, flying saucers, flying lawnmowers, flying carpets, and many others. It was only natural therefore that eventually my fascination with things that go bang, chemistry, physics and aerodynamics would collide and produce a strong interest in jet engine technology. It was also inevitable that, rather than focus on currently fashionable small gas turbine technology, I’d instead concentrate my efforts on the almost forgotten pulsejet. Over the past couple of years I’ve built dozens of different pulsejet designs, mainly to test my own ideas. As a result of this experimentation I’ve developed a reputation for being at the leading edge of this almost forgotten technology and have come up with a number of innovations such as the blast ring and a novel fuel-injection system that significantly extends the valve life of small pulsejet engines. For about a year I was actively involved in the commercial manufacture of several of my pulsejet designs but unfortunately I rapidly found myself extremely embarrassed at being unable to keep up with the unexpected demand. As a result, I have sold the manufacturing rights for these engines and am again focused on pure research and development in this area, working on several projects including some commissioned by clients in the aerospace and defense industries. I am also performing some design work on a new generation of ultra-low-cost high-speed pulsejet-powered UAVs designed for reconnaissance and other applications.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 4

Contents: How do pulsejets work? The world’s simplest pulsejet Pulsejets for models How to design a pulsejet Comparing intake valving systems Making reed valves last Fuel systems Constructional techniques Powering things with pulsejets Schmidt’s contributions Ignition systems How to start a pulsejet Valveless pulsejets The design of valveless pulsejets Improving pulsejet performance Accidents and failures A simple guide to anodizing Making reed valves with electrochemical etching Newton’s third law The Reynolds effect The Bernoulli effect The Coanda effect Plans Wacky ideas Afterburning augmentors A Little History and a Few Important People The future of pulsejet technology

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

5 8 10 12 20 23 29 37 43 46 48 51 54 59 66 69 76 82 88 92 91 93 95 97 99 103 105

Page 5

How Do Pulsejets Work? The honest answer to this question is that nobody’s really 100 percent sure of all the mechanisms that drive a typical pulsejet engine. While most of the basic principles of operation are understood fairly well, there are many small details that are still the subject of debate amongst engineers and experimenters to this very day. It is safe to say however, that the primary effect behind the function of a pulsejet is the fact that gases are compressible and tend to act like a spring. This “springiness” is crucial to the way a pulsejet draws in a fresh mixture of air and fuel then expels the hot burning gasses that are generated when that fuel is ignited. The Kadenacy Effect The effect of this springiness has been labeled the Kadenacy effect and here’s how it works: Take a regular 12 inch rule and lay it over the edge of a table so that just two or three inches are held firmly against that table. Pull the free end down an inch or so and release it. Did you see what happened? The ruler, acting like a simple spring, quickly flicked back, away from your hand – but it didn’t stop once it became fully straightened – it continued to move and actually bent the other way very briefly. This flexing back and forth probably continued for a second or so with the magnitude of each swing being slightly smaller than the previous one. Now something very similar happens when we take a sealed container and fill it with a compressed gas such as air. If we suddenly release that pressure by popping the cork, the compressed air will rush out but (and here’s the surprising bit) – even once the pressure inside falls to match the pressure outside, the air will continue to flow out. It’s pretty easy to see that this will cause the pressure inside the container to fall below the pressure outside – and then the gas will flow back inwards. This cycle of increasing and decreasing pressure will repeat a number of times, decreasing in magnitude each time. That’s the Kadenacy effect in action driven by the springiness of air.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 6

For a practical demonstration, find an empty bottle that has an opening about the size of your finger or thumb. Wet your finger or thumb and slide it into the neck of the bottle, allowing the air to escape as you do. Now remove your finger quickly. Hear that sound? That’s the air rapidly bouncing in and out – just like the springy ruler vibrated when you let the free end go. See how easy this jet engine stuff is? Now that we’ve seen how Kadeancy works it’s time to explain how it becomes the driving force behind a pulsejet engine. The Pulsejet Engine Operating Cycle Let’s assume that a mixture of finely atomized fuel and air has just been ignited inside our pulsejet. The rapidly burning fuel generates gases such as carbon dioxide, carbon monoxide and watervapor. These gases take up far more room than the air and fuel alone did so pressure build up inside the engine. The heat generated by the combustion causes those gases to expand so the pressure is increased even more. Our pulsejet has become a container filled with pressurized gases – just like the one described previously. Those gases now rush out the opening at the end of the engine’s tailpipe and in doing so, they create thrust which pushes the engine (and whatever it’s attached to) in the other direction. It was Benjamin Newton who first described this effect when he said “for each and every action there is an equal and opposite reaction.” A fraction of a second after the air/fuel is ignited and the hot gases have started flowing out the tailpipe, the pressure inside the engine has dropped to match that of the outside air. However, thanks to the Kadenacy effect, the gases keep flowing down the tailpipe and a partial vacuum is created inside the engine. At this point, two very important things start happening. Firstly, the valves at the front of the engine open. They’re pushed open because the air outside the engine is at a higher pressure than that inside the engine. This pressure difference pushes on the valve and causes them to move aside, thus allowing fresh air and fuel to enter.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 7

At the same time, those hot burning gases that were travelling down the tailpipe stop for an instant then start travelling back towards the front of the engine – driven by the air outside which is at a higher pressure than that inside. You can see that at this point, we have fresh air and fuel coming in the front and still-burning gases coming back down the tailpipe. Can you guess what happens when the two meet? That’s right – the whole cycle starts all over again when the flames and hot gases from the tailpipe ignite the highly flammable mixture of air and fuel that has been sucked in through the valves in the front. And of course, as soon as that fuel ignites, the pressures generated push the valves at the front of the engine closed, leaving the hot gases only one way to go – out the back. When a pulsejet is running, this whole process is repeated many times per second and it is this repeated blasting of hot gases out the tailpipe that gives the engine its characteristic noisy bark. Wasn’t that simple?

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 8

The World’s Simplest Pulsejet Here’s a chance to make your own ultra-simple pulsejet using nothing more than a hammer, a screwdriver, and a few readily available materials. This simple design was first dreamt up by one of the grandfathers of pulsejet engines, a Dutchman by the name of Francois Henri Reynst back in the first part of last century. Although this pulsejet won’t hurt your ears or produce massive amounts of thrust, it’s still a good idea to include a few warnings at this point. Safety Pulsejets use explosive mixtures of air and fuel to create power. They also produce lots of burning hot gases when running. For these reasons, you should never attempt to run a pulsejet (not even this very simple one) indoors or near anything that could catch fire. Also be aware that because a pulsejet generates pressurized gases, there’s always a small risk that part of the engine could fly off and strike anyone standing nearby. This is particularly true if you’re using a glass jar in the following experiment. The glass can (and will eventually) crack and break due to the heat and pressures involved. Here are some basic rules that will help keep you safe: 1. 2. 3. 4.

Always wear eye protection Keep a safe distance from a pulsejet when it is running. Keep a garden hose and/or bucket of water nearby at all times Use hearing protection

Now on with the fun. Materials In order to build our simple demonstration pulsejet you’ll need to find the following items: 1. A small jar with a screw-top metal lid (75mm or 3” diameter) 2. A screwdriver or nail. 3. Some methanol or model airplane fuel Here’s how we go about building our engine. With the nail or screwdriver, make a hole in the center of the removable metal lid. You can then enlarge this hole to around a half-inch (13 millimeters) in diameter. Now pour some methanol or model airplane fuel into the bottom of the jar to a depth of about a quarter inch (5 millimeters).

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 9

Replace the lid and swirl the liquid around in the bottom of the jar a few times. Remember that you need to run this little engine well away from anything that could catch fire and one suggestion I have is to dig a small hole in the ground so that the jar can be inserted, leaving the lid slightly above ground-level. This will protect you and reduce the fire risk if the jar should crack. Now bring a lighted match or flame near the hole in the jar’s lid – keeping your face and hands well away from the area directly above that hole – because a large flame may come shooting out with a whooshing noise that can give you a bit of a fright. It is highly recommended that you wear eye-protection and at least a long-sleeved cotton shirt to protect yourself when performing this experiment. If you’re lucky, your simple little “jam jar” engine should start puffing away – producing a series of little pulses of hot gas and perhaps a gentle purring noise. IMPORTANT: do not let this simple jam-jar engine run for more than 5-6 seconds at a time or the glass will crack, possibly spilling burning fuel. You can stop it by covering the hole in the lid with a small piece of wood or even a suitably sized coin. How Does It Work? Now the more observant reader will have noticed that this pulsejet has no valves – so how does it work? The answer is simple – when you ignited the air/fuel mixture that was originally in the jar it burnt, expelling that large jet of flame and making that whooshing noise. Because the hot gases rush out of the jar with great speed, the pressure inside the jar quickly drops below normal atmospheric pressure and a weak vacuum is created – just as described in the previous chapter. When this happens, a fresh gulp of air is sucked in through the hole in the lid and that air then mixes with highly flamable methanol vapor that is rising from the pool of fuel still sitting in the bottom of the jar. The jar once again contains a combustible mixture of air and fuel – but how does it ignite? After all, we had to use a match to ignite it the first time didn’t we? Well, also inside the jar are some remnants of the hot gases generated from the last combustion cycle. Eventually the hot gases and the air/fuel mixture run into each other – whereupon ignition occurs, pushing its hot gases out the hole –starting the whole cycle all over again. How’s that? We’re not even a quarter way through this book and you’ve already built your own pulsejet engine!

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 10

Pulsejets for Models

Forty or fifty years ago there were a number of manufacturers producing small pulsejet engines designed specifically for use on model aircraft. Most of these engines are very similar in design and construction, consisting of a lightweight stainless-steel body and tailpipe, with a machined aluminum head and thin spring-steel valves. Quite a few of these engines were made in Eastern-bloc countries and at least one was made by the OS model engine company in Japan. However, perhaps the most recognized model pulsejet of all time is the Dynajet. The Dynajet Thousands of avid model airplane enthusiasts have owned, seen or lusted after this icon of the pulsejet era. The Dynajet was so popular, and so many were sold, that it rapidly became the benchmark against which all other small pulsejets were compared. Its simple design and lightweight construction made it perfect for use on model airplanes and well suited to the U-control speed models of the era. Even today the Dynajet is a popular item on auction website such as eBay, often producing bids of several hundred dollars or more. This picture is of an early model Dynajet that had the spark plug located right at the rear of the combustion chamber and didn’t have an anodized head. In later models the sparkplug was moved forward and the aluminum head was anodized a rich red color which resulted in the engine being known as the Dynajet Red-head. The body of these engines is made from two pressed stainless steel shells that are welded along an upright seam. This technique results in a nice smooth contour between the combustion chamber and the tailpipe, which probably helps the performance somewhat. This picture shows the front of a later-model Dynajet with a more deeply finned valve-head section. You can clearly see the effects of the red anodizing on the aluminum and the more forward location of the sparkplug. The company that used to make these engines, Dynafog, is still in business but now focuses on industrial fogging machines which use the pulsejet principle to atomize and spray insecticide and other chemicals.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 11

Another company, Bailey Machine Service, is still making a clone of the Dynajet engine which it sells for about US$250 although the supply is said to be somewhat erratic. Many of the other pulsejet designs you’ll see from this era are very similar to the Dynajet in size, dimensions and performance, although there are still plenty of weird and wonderful variations on the basic theme. The OS pulsejet This engine was manufactured in the 1950s and 1960s by the Japanese company OS. There were apparently two slightly different versions of this engine, one being a little larger and more powerful than the other but they were both obviously very similar to the Dynajet in both form and function. However, unlike the Dynajet which used a single machined piece of aluminum for its valvehead and venturi, the OS unit consists of two separate parts which screw together. There was also a myriad of other pulsejet designs manufactured about the same time and sold under a raft of different names such as TigerJet, etc. An even wider range of designs and ideas were published as plans for a generation of enthusiasts who, in a post-war era, were eager to build one of these magical “jet engines” for themselves. As a result, many magazines such as Popular Science and Popular Mechanics were littered with advertisements for such plans. It is unlikely that many of those who purchased these plans ever managed to construct a working engine and at least a few of the designs were so fatally flawed that the publisher was obviously just trying to cash in on a craze.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 12

How to design a pulsejet One of the more interesting and more readable texts on pulsejet design and theory was written by a C.E. Tharratt while he was a staff scientist at the Chrysler Space Division in the late 1950s. Titled “The Propulsive Duct”, this paper condensed much of the known pulsejet theory into a few simple formulas and constants. Using these calculations Tharratt claimed that “we are in the surprising position of being able to determine the dimensions of a duct capable of developing a given thrust literally on the back of an envelope and without knowing anything about its gas dynamics!” Sounds great doesn’t it? Unfortunately, while Tharrat’s formulas have stood the test of time and his understanding of the mechanisms behind the pulsejet (or “propulsive duct” as he called it) remains valid, when it comes to designing a powerful, reliable pulsejet engine, the devil is in the detail. However, here is the simple formula that Tharratt proposed to be the core of pulsejet design (note: Tharratt’s papers and constants are presented in the imperial measurement system so that’s what I’ve used in this chapter. I’ll update with metric versions in the next release of this book): V/L = 0.00316F Where:

V = engine volume (cu ft) L = effective acoustic length of the engine (ft) F = thrust (lb)

The validity of this formula has been verified against a wide number of different and proven pulsejet designs including the Argus V1 and Dynajet. Let’s take a look at what this formula actually means in terms of the way that the dimensions of a pulsejet affect its power output. We can see that if we kept the volume of the engine (V) constant but increased its effective acoustic length (L) then the power would reduce. Note that in order to do this, the diameter (and cross-sectional area) of the engine would need to be reduced – so it would appear that there is a definite relationship between cross-sectional area and power. Now, if we keep the length (L) constant but increase the volume (V), the power would increase. To accomplish this however, we’d have to increase the diameter (and cross-sectional area) of the engine. This confirms that relationship between cross-sectional area and power output.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 13

If we manipulate that simple formula a little more, we come up with a new formula containing a very interesting constant: F = 2.2A Where: F = thrust (lbs) A = mean cross-sectional area (square inches) Let’s just make an important point here – this 2.2lbs/sq-in constant is derived from a formula that includes the engine’s total volume as a factor. This is why it’s not just the cross-sectional area of the tailpipe that is important (as many would have you believe), but the mean (average) cross-sectional area of the entire engine along its total length. However, it should be added that Tharratt didn’t expect engines built to his formulas to have a huge bulbous combustion area at the front so don’t expect that such a feature will significantly increase an engine’s performance over a straight pipe. So now we can plug in our required power output of 10lbs thrust and get this: 10=2.2A which simplifies to: A=10/2.2 A = 4.545 square inches From this we can calculate the mean diameter of our engine as follows: D = 2√(4.545/π) D = 2.4 inches Now we need to decide on a length for the engine. Remember that making the engine longer or shorter won’t actually increase or decrease its power – only a change to the cross-sectional area will do that. However, the length does have a bearing on the frequency at which our engine will operate. The only reference I can find from Tharratt in respect to the suggested length of a pulsejet engine is that it be at least eight times the mean diameter. This length to diameter ratio is usually expressed as L/D and Tharratt notes that “with geometrical ratios of L/D < 7 the development problems become particularly challenging” and that in an engine with an L/D < 7 “combustion with chemical fuels is difficult to sustain, let alone develop maximum thrust”

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 14

It makes sense therefore, to use an L/D somewhat greater than 7 and it’s been my experience that a figure of about 14 is a good place to start for relatively small engines like this. By way of comparison, the Dynajet has an L/D of 15 and the Argus V1 has an L/D of 9.6. As a rule of thumb, the smaller the engine, the higher the LD needs to be in order to get reliable operation and good power levels. So now we can calculate the length of our engine as follows: L = 14D L = 14 x 2.4 L = 33.6 inches Okay, so now we know that to create a 10lb-thrust pulsejet engine we’ll need a piece of pipe that is 33.6 inches long and 2.4 inches in diameter -- but wait, there’s more! Another key formula presented by Tharratt was one for calculating the valve area for an engine of a given size and power: Valve area = 0.23 x mean cross-sectional area Or

Valve area = 0.1045F sq in

It is interesting to note that this 0.23 (or 23 percent) figure contrasts sharply with that proposed by other pulsejet “experts” of the era who suggest a figure of 0.4-0.5 is better. Let’s use Tharratt’s formula and constant to work out the size of the valve area we need for our 10lb-thrust engine. Using the first formula we get: Valve area = 0.23 x mean cross-sectional area Valve area = 0.23 x 4.545 Valve area = 1.045 sq in Using the second formual we get: Valve area = 0.1045F Valve area = 0.1045 x 10 Valve area = 1.045 square inches Yep, we get the same answer both times so we now know that the effective valved area for our l0lbs-thrust engine is just over 1 square inch. Now Tharratt’s formulas assume that the intake is an open hole, with no losses due to the presence of valves. Unfortunately, the presence of spring-steel reed valves will significantly impact the flow of gas so we need to take those losses into account when designing our intake valving system. Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 15

The precise efficiency of a valving system depends on many factors and I suggest you read the chapter on intake valving for more information – but in the case of our little design, let’s use a simple petal valve and assume that it is just 50 percent efficient. To get the actual area of the valve required to achieve an “effective” area of 1.045 square inches at 50 percent efficiency we simply divide by 0.5 to get a figure of 2.090. So let’s see how our pulsejet is looking like so far: Thrust Length (from valves to end of tailpipe) Mean diameter Total valve area (assuming 50% efficiency)

10lbs 33.6 inches 2.4 inches 2.090 square inches

So what about those valves then? A petal valve system consists of a ring of holes over which a spring steel valve, consisting of a matching number of petals, is laid. If we were to use 10 holes, spaced at 36 degree intervals, each hole would need to have an area of: 2.090/10 = 0.209 square inches which requires a diameter of 0.516 inches. Alternatively, if we used a ring of 12 holes spaced at 30 degree intervals, each hole would need to have an area of: 2.090/12 = 0.174 square inches which requires a diameter of 0.470 inches It’s been my experience that a half-inch hole is the upper limit for petal valve holes. Once you go beyond this size the pressure on the valves themselves cause them to be bent so that they begin to dish into the hole and this adversely affects their operation. For this reason, we’ll go with the 12-hole option. You’ve probably already noticed that most small pulsejets have a much larger diameter section at the front, from where they funnel down to a narrower tailpipe. Many people mistakenly assume that this bulbous front is a combustion chamber, designed to contain the burning air/fuel mixture. While it may be true that much of the air/fuel is burnt inside this bulbous front section, the reason small engines are shaped this way is actually quite different and we’ll see why when we do the next set of calculations.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 16

We’ve decided to use a petal valve system consisting of 12 holes, each of 0.47 inches in diameter. If we assume that the holes will be placed in a ring around the edge of the pipe, and that we allow a certain amount of space between the holes so that there’s room for the springsteel valves to rest, then we have a problem. Lets assume we need a ¼ inch gap between the holes – that means the total circumference of a circle drawn through the center of each hole will be: NumOfHoles x diameter + NumOfGaps x GapSize And when we plug in our numbers we get: 12 x 0.47 + 11 x 0.25 = 8.39 inches That represents a circle with a diameter of: 8.39/π = 2.67 inches What’s more, that figure is the diameter of a circle that runs through the center of each hole so we need to add two times the radius of the holes to get the size of a circle that will run around the outer edge of the ring of holes. 2.67 + 0.47 = 3.14 inches Clearly, the absolute minimum diameter of our valve system (3.14 inches) is larger than the calculated diameter of our engine’s pipe (2.4 inches) This disparity grows even further when we build in a bit of extra space so that the valves don’t scrape against the side of the engine when they swing open and closed. It’s been my experience that you should allow a space between the outer edge of the valve holes and the side of the engine which represents an area equal to the total area of our valve holes. Or in other words, we need to leave 2.090 square inches of space around our ring of holes. That can be calculated as follows: The area of a single circle covering our ring of valve holes would be: πR2 or 3.1415 x 1.57 x 1.57 = 7.74 square inches Add 2 .090 square inches to get the size of the larger circle and we get 9.830 square inches From this we can calculate the diameter needed to provide that extra 2.090 square inches of space around the edge of the valve-ring: Diameter = 2√(9.830/π) Diameter = 3.53 inches Is your head sore from all these calculations yet? Don’t worry, we’re nearly done.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 17

Now that we know the diameter of our “combustion chamber” we need to calculate how long this section of the engine should be. Since Tharratt didn’t use petal valves, he didn’t need this bulbous front section so has no advice (that I can find) for calculating this dimension. However, we can look to the work performed by Schmidt (the guy who designed the Argus V1 engine) and my own empirical research that indicates the following: During the intake phase of a pulsejet’s operation it will draw in a fresh charge of air equal to 15%-20% of the total engine volume. I see no reason why we shouldn’t design the engine so that this front section is just large enough to hold this fresh charge of air/fuel. In that case we need to do some more calculations to determine its length: If our engine were just a straight pipe of 33.6 x 2.4 inches then it would have a volume of 152.7 cubic inches. Such an engine would suck in 152.7 x 0.2 = 30.54 cubic inches of air during each cycle – so our front section needs to have a volume of 30.54 cubic inches. We’ve already calculated the area of this section as being 9.83 square inches we can calculate the required length as follows: 30.54 / 9.83 = 3.11 inches. Of course more alert readers will notice that this 30.54 cubic inches no longer represents 20 percent of the engine’s total volume. This is because by making this front section wider without reducing the overall length of the engine, we’ve actually increased its total volume by an additional 16.4 cubic inches. The total volume of our engine is now nearer 169 cubic inches so the 30.54 cubic inch front section only represents 18 percent of the total volume – but this difference is so small as to be unimportant. The only thing remaining now is to join the front section of the engine to the tailpipe using a cone with a diameter of 3.53 inches at one end and 2.4 inches at the other. How long (ie: what angle) should this cone be? Well if we look at the Dynajet we can see that it’s hardly a cone at all – more of an abrupt transition. By comparison, the Argus V1 engine uses a very long, shallow angled cone to join the two sections. So how do we decide which is best? Let’s look at the effects that the angle of this cone might have on an engine’s operation. Coming up with a suitable angle for this cone requires balancing a number of factors. To examine them, let’s look at extreme examples:

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 18

1. If we simply used a flat plate to join the two sections of the engine then the hot exhaust gases would have a rather torturous path to follow. Some of those gases would have to travel around two 90 degree bends to get from the combustion chamber to the tailpipe and that would potentially reduce the speed at which they were able to exit from the engine. Remember, the speed at which the gases leave the engine affects the thrust – we want those gases leaving as fast as possible. For this reason, a flat plate is obviously not such a good idea. However, this configuration is not quite as bad as you might think, after all, the Dynajet has a very steeply angled cone that must constrict the flow of exhaust gases right? The very fact that it is so hard for the combustion gases to get into the tailpipe means that immediately after ignition, pressure will build up inside the combustion chamber as all those gases try to rush around a tight bend and down the tailpipe.. Those higher pressures can improve combustion efficiency and actually increase the speed of the gases in the tailpipe. This is called post-ignition confinement (PIC). It’s also worth noting that in the case of a flat plate, the hot gases that return from the engine's tailpipe and ignite the fresh air-fuel charge may do so far more efficiently. This is due to an interesting effect that occurs in the way they form a narrow jet that reaches deep into the chamber rather than a larger diffuse front that ignites the fuel more slowly. Remember that the faster the fuel burns, the more power our pulsejet will develop because it will have less time to expand as it burns – thus producing the higher internal pressures that will, in turn, result in higher tailpipe gas velocities.

This diagram shows how ignition differs based on the angle of the cone between combustion chamber and tailpipe. Note that in the second diagram, the distance between the hot gases and the engine body is far less than in the first – this is important. The speed at which the combustion flame-front travels through the fresh air/fuel mixture is relatively slow (just a few tens of feet per second) in a low-compression engine like the pulsejet. Because of this, the mixture in the second diagram will be burnt far more quickly than that in the first, since the flame-front will be wider with a much shorter distance to travel. 2. If we used a very long cone that had a shallow taper all the way to the end of the engine (ie: no tailpipe as such) --then it would obviously be much easier for the combustion gases to Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 19

flow out under pressure. However, we’d also be significantly reducing the ability of the engine to create a vacuum after combustion is completed because a much smaller percentage of the exhaust mass would be travelling at maximum velocity inside the engine. Remember that the “force” exerted by the escaping gases is equal to their mass times the velocity to which they’re accelerated (F=MA). For a given size of engine engine, the mass will always be the same but the velocity to which those gases are accelerated will depend very much on the design of the tailpipe. We need plenty of velocity to get the force required to establish a strong Kadenacy effect. In fact, tests conducted by the NACA during the 1950s indicated that an engine designed with just a long convergent cone instead of a straight tailpipe was very difficult to get running at all. Once again it seems that a compromise is in order so we’ll chose an angle of 30 degrees for the section between the combustion chamber and the tailpipe. This will provide some postcombustion confinement to increase the internal operating pressures while ensuring that the engine still has good internal mass-flow speeds to provide maximum Kadenacy effect. A 30-degree cone will be a fairly short section – just 1.84 inches long. So here are the final dimensions of our pulsejet engine, wasn’t that simple? It should be noted that this is a very basic engine and there are still a few tricks we can use to improve its performance – but more of this later.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 20

Comparing Intake Valving Systems

One of the most critical components of a traditional pulsejet engine is the intake valving system. The demands placed on the intake valves are amazing. They have to open and close several hundred times a second while being exposed to the thermal stresses associated with being alternately blasted by searing hot combustion gases and cold incoming air. At the same time, these thin strips of spring steel must resist metal fatigue and fracture resulting from the high mechanical stresses imposed. What’s more, they have to do all this while providing a 100 percent seal against combustion gases when closed, and allowing the smooth, unimpeded flow of fresh air when open. To make life even harder, the only power available to open them is the tiny difference in pressure between the outside air and the small vacuum created inside the engine by the kadenacy effect of escaping exhaust gases down the tailpipe. (just a few psi). It’s no wonder therefore, that no aspect of pulsejet design and construction has caused more sleepless nights, scratched heads and frustration than the valving. Lets examine the alternatives: Petal Valves Small engines almost always use a petal-valve. These valves offer the following benefits: 1. simplicity. The valve can be etched or cut from a single piece of spring-steel. 2. Low cost. As a side effect of their simplicity, petal valves can also be very economical to manufacture – especially when you consider that the valve plate consists of a simple piece of aluminum with a ring of holes drilled in it. Unfortunately, the petal valve also has a number of disadvantages: 1. poor aerodynamic performance. Since the air passing through a petal valve must negotiate two near-90 degree bends on its way into the engine, the efficiency of such a system is not particularly high. 2. low durability. Because the tips of the petals are directly exposed to the hot combustion gases, petal valves often suffer from premature tip cracking or fracture. 3. High maintenance. Since petal valves are usually made as a single piece, the failure of individual petal requires the replacement of the entire spring-steel valve.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 21

Despite their drawbacks, petal valves are generally the best option for small pulsejet engines, although I wouldn’t recommend them for any engine larger than about 20lbs of thrust. The V or multi-V valve Generally only seen on larger engines, these valves are generally more efficient than petal valves because they produce less deflection of the airflow when they’re in an open position. There are two basic methods of constructing such a valve system – one involves the use of two or more flat metal plates with holes in them, joined at an angle (45 degrees is a good starting point). The other method of forming a V valve is the one used in the Argus V1 where a cast or machined spacer with multiple ribs is used to hold the valves in position and limit their movement as in the diagram below:

V valves provide the following benefits: 1. Higher efficiency than a simple petal valve. Since the incoming air has a far straighter pathway into the engine, more air is able to flow for a given size of valve opening when compared to a petal-valve. 2. Lower maintenance costs. Since the individual spring steel valves in a V-valve system can be replaced as/when they fail, maintaining the engine becomes a less expensive task and all valves can be used to the full extent of their lifespan. 3. Scalability. Unlike the petal-valve, a V-valve can be easily scaled to create the required valve area by simply increasing the length or number of V-valves in the array.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 22

Of course there are downsides too: 1. Greater complexity. A V-valve generally requires more machining steps and a higher component count than a petal-valve setup. 2. Increased expense. As a side effect of this complexity, the production cost for a V-valve system is significantly higher than for a petal-valve. This is another reason why most cheap model engines don’t use V-valving. Less commonly used valving systems Petal and V-valves are not the only systems that have been used on pulsejets but they are by far the most common. Perhaps the only other practical valving system for a pulsejet is: The Rotary Valve These generally consist of either a spinning disk containing a hole that controls the flow of gas by covering and uncovering a matching hole in the front of the engine, or a spinning butterflytype valve that alternately blocks and allows the flow of gas. Rotary valves can be made very robust and thus have the potential to create very reliable, longlived pulsejets. Unfortunately however, they are fraught with hidden complexities, not the least of which includes the issue of timing. In a conventional pulsejet valving system, the valve timing is automatically controlled by the changing pressure inside the engine. When the internal pressure goes up (because the fuel has ignited) then the valves close. When the pressure falls (due to the Kadenacy effect) then the valves open). This results in a very simple and quite reliable system that automatically compensates for any fluctuations in the engine’s operating frequency or phase. Rotary valves on the other hand, have no such intrinsic timing control and therefore require a very sophisticated system to control their rotational speed and phase relationship to the engine’s basic operating cycle. This immediately negates the pulsejet’s two single most endearing qualities – simplicity and low cost. Research done in the USA during the 1940s cited engines using the rotary valve as offering “very long useful operating periods” along with “good thrust and specific fuel consumption” but also mentioned the complexity associated with driving such a valve in a synchronous manner. Never the less, rotary valves are being considered as a viable option for the new generation of pulse detonation engines (PDEs) currently under development. Since these PDEs already require a significant number of ancillary control systems anyway, the overhead of the rotary valve adds little to the cost or complexity of these engines.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 23

Making Reed Valves Last Longer The thin spring-steel valves normally used to control the flow of air into a pulsejet and stop the hot combustion gasses from escaping out the front are the most highly stressed part of a conventional pulsejet engine. The life of the reed valves in most pulsejet designs is measured in minutes rather than hours and they must be considered a “consumable” part of any conventional pulsejet engine. It's not hard to understand why these fragile little pieces of metal don't last long. They're slammed back and forth between the intake and retainer plates with great force, several hundred times per second. What's more, they're usually exposed to extremely hot combustion gasses There are three factors that contribute to reed-valve failure: • • •

heat-damage impact damage fatigue due to flexing

A well designed valve system attempts to minimize all these factors so as to provide maximum valve life but (woudn’t you know it) there are always compromises involved. Let’s look at the simplest and easiest to solve issue first: Fatigue due to flexing This picture shows the effect of valve failure due to metal fatigue brought about by the flexing motion of a petal valve. Note that one of the petals has completely broken near the root. Close inspection of this valve showed that stress cracks were starting to appear at the root of the other petals. In this case, failure was due to a poorly designed valve-retainer which had an uneven radius of curvature. Obviously, reed valves must flex in order to operate properly. The key to avoiding premature failure however, is to limit this flexing to a bare minimum and try to keep the flex radius as large as possible. To this end, the conventional petal-valve arrangement with a curved valve-retainer is quite good – providing the curvature of the retainer is of a constant radius. If the valve retainer doesn’t have an even, large radius curve to it, most of the valve flexing will be concentrated over a small area near the root of the petal. In a fairly short space of time, the stresses caused by this flexing will cause the spring steel to crack and fracture.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 24

This diagram shows the right and wrong way to design the valve retainer for a petal valve system. Note that when a small radius is used near the base of the valve retainer, most of the reed valve remains unbent so all the stress is concentrated at the root. There are several ways to make a valve retainer that has a large, even radius but because it’s so much easier to make a straight-sided retainer, people often make the mistake of creating what amounts to a shallow cone instead– with predictably bad results. Impact Damage Most small pulsejets run at somewhere between 180 and 250 cycles per second. This means that the valves must open and close as often as 15,000 times per minute. Each time the valves close, they slam into the valve-plate at quite high speed and therefore with significant force. All the energy that is contained in these fast-moving valves has to go somewhere – and some of it is absorbed by the valve material itself. This constant hammering eventually causes minute cracks to form at the tips of the valves after which they begin to fray and small fragments will eventually flake off. If you’re running your pulsejet at night, these small fragments can be seen as impressive sparks flying out the tailpipe. This picture shows a badly frayed petal valve that has certainly reached the end of its useful life. It is a very good idea to replace valves long before they get to this state because the sharp, ragged ends will soon damage the comparatively soft material of the aluminum valve-plate against which they impact. There are a few techniques that can be used to reduce impact damage to reed valves. •

Reduce the amount of valve travel. If the valves can open too far then they will reach a much greater velocity when they’re closing and this will increase the forces applied to them as they impact the valve seat. Of course limiting the valve-opening will also tend to reduce an engine’s power as it means that less air can be drawn in during the intake phase.

•

Use a softer material for the valve-plate. Most small pulsejets already use an aluminum alloy for the valve plate so there’s not a lot of room for improvement here. However, the NACA did conduct tests on engines which had a think coating of neoprene on the valveplates. This was said to almost double valve-life by reducing the impact shock experienced by the closing valves.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 25

Note that although it might seem like a good idea to simply increase the amount that a valve overlaps the hole it covers so that the air trapped between the valve and the plate acts as a cushion to soften the impact, this is actually a bad idea. If the overlap area is too great, the air can’t get out of the way quickly enough and the tips of the valves are actually bent backwards by this trapped air. As a result, tip fraying is dramatically increased because any cracks that form grow very quickly due to this additional stress. Determining the ideal overlap area is something best done by trial and error. If the overlap is too great then you’ll get premature fraying and poor engine performance (due to late closing of the valves). If the overlap is too small then the valve plate will be damaged by the high pressure loadings and this will ultimately affect engine performance because the valves will no longer seal properly. Heat Damage Reed valves are usually made from hardened, tempered spring steel because it’s strong and will return to its original position after flexing. The problem with spring steel is that the hotter it gets, the softer it gets. If it gets too hot then it will lose much of its strength and some of its springiness. Unfortunately, the inside of a pulsejet engine is a very hot place and it is the pressure generated by extremely hot (1,500 deg C) combustion gasses that actually cause the valves to be closed. So why don’t the valves simply get red hot and go all floppy? Well fortunately, the valves are only exposed to the hot combustion gasses for part of the operating cycle. During the intake cycle they’re cooled by the incoming charge of fresh air. In a petal valve setup, the valve retainer also provides a measure of protection from the heat of combustion by shielding most of the valve from direct exposure to the hot gases. However, the tips of the reed valves will still get hot and, as a result, they will become softer than the rest of the valve. What’s more, the valve retainer will itself heat up once the engine is running and some of this heat will be transferred to the valves when they’re open. A number of solutions have been proposed to the problem of valve heating but most of them will reduce an engine’s performance to some degree. One of the simplest solutions is to place a flame-trap in the engine directly behind the valves. This flame trap consists of little more than a mesh of stainless steel or some similar heatresistant metal. It’s well known that a metal mesh with suitably sized holes will not allow a flame to pass through it – but it will allow air and other gases to do so. This was the principle behind the Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 26

old miner’s safety lamps. Before the days of battery-operated flashlights, miners still needed some form of illumination while working underground. A naked flame such as that from a candle or oil lamp would pose a very real danger as it risked igniting underground pockets of explosive gases such as methane. By enclosing the flame in a metal mesh, the flame could not extend beyond that mesh so the “safety lamp” could provide light and still be used without risk of sparking an explosion. Now, the problem with a using a flame-trap mesh in a pulsejet is that in order to be effective, the size of the holes in the mesh must be quite small. As a result, the mesh represents a significant obstacle to the flow of the incoming fresh air charge. This means less air is drawn in during each intake cycle so less power is produced.. Never the less, a flame-trap mesh is one way of making a relatively low-powered engine that will run for far longer between valve-changes than a regular pulsejet. Over the past two years I’ve given quite a bit of thought to the issue of extending valve life and have come up with two ideas of my own. The first is the Blast Ring concept which works by providing a physical shield between the hot exhaust gases and the valve tips. By blocking the direct path of the combustion flame to the valve-tips, the operating temperature of the valves is significantly reduced. However, because the Blast Ring has a very large hole in the middle it doesn’t restrict the flow of the incoming charge of fresh air to the same degree as a flame-trap mesh. This picture shows my PJ15 design running with a Blast Ring in place. You’ll notice that the ring itself is glowing red-hot, an indication that it is indeed absorbing much of the heat that would otherwise be reaching the valves. However, even this system imposes about a 15%-20% performance penalty on the power levels that would otherwise be obtained from an engine. Another method for reducing the valve heating is to design the engine so that there’s a buffer of cold, dense air between the valves and the combustion gases. The easiest way to do this is to inject the fuel into the combustion chamber some distance from the front.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 27

In such an engine, the air in front of the injection point will not contain any fuel thus will not actually take part in the combustion process. It will however, act as an insulating buffer between the hot combustion gases and the valves. Unfortunately, as with the other methods mentioned so far, there’s a performance penalty associated with this method. Since a pulsejet normally only draws in a fresh charge of air equal to about 15-20 percent of its total volume, creating a buffer zone which contains no fuel leaves less available for the combustion process. That means less fuel can be added and, as a result, the engine produces less power. There are very few free lunches in the world of pulsejet design. More recently however, I have come up with what appears to be a system that imposes no performance penalty, yet significantly improves valve-life. This system works by creating a two-layer valve retainer that is cooled by the incoming fuel. As you can see in this diagram, the fuel (purple) passes between the two thin dished valve-retainer plates before mixing with the incoming air. As you can see, this method allows virtually all the air in the combustion chamber to be mixed with fuel and therefore provides good power. This provides multiple other benefits over the traditional system. 1. The fuel is pre-heated and/or vaporized before it mixes with the incoming air. This provides a much better (and more combustible) mixture than is normally achieved either by direct injection or atomization. 2. As it atomizes, the fuel absorbs a tremendous amount of heat from the two dished plates, this cooling them to a much lower temperature than they would otherwise run at. 3. The two disks, with the small gap between them, act as a far more efficient heat shield than does the normal one-piece valve retainer. 4. The small gap between the retainer disks tends to absorb some of the hot combustion gases that would otherwise reach the valves.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 28

Here’s how I’ve implemented this design concept on my PJ15 engine. The two disks are formed from thin 0.020” (0.5mm) stainless steel which is spun to shape on a lathe. I’ve actually observed a small power increase after changing from a conventional one-piece valve retainer to this new concept and valve-life has been almost doubled. Unfortunately, building a valve system like this takes more time and more time and skill than the traditional one-piece valve-retainer – remember what I was saying about free lunches?

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 29

Fuel Systems Next on the list of critical elements of a pulsejet must surely be the fuel system. Atomization Smaller engines such as the Dynajet have traditionally used a very crude form of carburetor that using the incoming air to create a spray of rather coarsely atomized fuel droplets. This atomizing process occurs right at the front of the engine when the incoming air is forced through a slight venturi. An Italian by the name of Bernoulli discovered that the faster air flows, the lower its pressure becomes. This observation was promptly labeled (wait for it…) the Bernoulli Effect. The atomizer on these small pulsejets uses a venturi to squeeze the incoming air through a narrowing in the intake. As it squeezes through, it has to speed up. As it speeds up – the pressure drops. Now, if we stick a pipe carrying some fuel into the middle of this low-pressure area, the fuel is literally sucked out and turned into a fine spray of droplets. What could be simpler? Unfortunately, although this system does work, the magnitude of the low-pressure area created in the pulsejet’s venturi is quite small and this means that there’s not much energy available to suck that fuel through. A Note About Atomization and Vaporization Another problem with the simple atomizer is that the fuel droplets created tend to be very large and therefore do not vaporize particularly well. It should be remembered that liquid fuels themselves don’t actually burn – only the vapors that they emit will ignite. In order to obtain good vaporization, the goal should be to create the smallest possible droplets because this results in the largest surface area (from which vapor is emitted) for a given volume of liquid. Fortunately however, the inside of a pulsejet engine is a very hot place so, despite the fact that the simple atomizer does a poor job of converting liquid fuel into a nice fine spray, the high internal temperatures of the engine greatly assist the conversion of those large droplets of fuel into vapor. [endnote] The end result is that most of these small pulsejets are extremely sensitive to just where the fuel tank is placed relative to the atomizer assembly. If you place the tank too low then the engine won’t have enough “suck” to pull the fuel up to the atomizer nozzle. Place the tank too high and gravity will draw the fuel through – effectively flooding the engine. Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 30

What’s worse, even if you do get the engine running nicely, moving the fuel tank up or down by even an inch or two can cause it to stop because the fuel flow is affected. There are ways to reduce this sensitivity to fuel-head however and perhaps the simplest is to use a pressurized fuel tank. By delivering the fuel under pressure, the effect of a changing fuel-level is dramatically reduced. The big problem is how do we generate this pressure? One option is to simply pump some compressed gas into the fuel tank then seal it up. In order for this to work, the tank should only be filled with fuel to only about 25 percent of its capacity otherwise the pressure inside will drop significantly as the fuel is drawn off.

Alternatively, the compressed gas can be stored in a separate container and fed into the fuel tank through a regulator. This is how the fuel system for the Argus engine that powered the V1 flying bomb was configured and is illustrated in the diagram above. Rather than rely on a large reservoir of compressed air inside the tank, it is possible to tap into the pressure produced by the combustion of the pulsejet itself. This diagram shows how some of that pressure can be directed into the tank to keep it pressurized. Note the small reed valve that stops the pressure from leaking back into the engine during the intake phase.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 31

In practice, the reed valve should be placed in the pipe that leads from the engine to the tank rather than in the tank itself. Surprisingly, there’s little risk that the hot gases from inside the engine will ignite the fuel in the tank. This system can be used with both atomized and injected fuel systems. Another simple way to achieve fuel pressurization is to use something like a small balloon for a fuel tank. This configuration is called a “bladder tank”. The elasticity of the balloon will automatically pressurize its contents – but be aware that some fuels will quickly break down the rubber from which normal balloons are made and if it goes “pop”, you’ll have a very real fire danger. Some of those using pulsejets in model airplanes often use these bladder tanks to ensure good pressurization and reliable fuel feed under varying G-forces. It’s worth noting however, that the rubber tank is usually contained inside another leak-proof container such as a plastic soda bottle. This way, if the bladder bursts, the fuel remains contained Most flyers of pulse-jet powered model airplanes also use a device called a Cline regulator to ensure not only that the fuel pressure remains constant but also to automatically shut off the fuel flow if the engine stops unexpectedly. You should also be aware that any leak in a fully pressurized fuel system can result in large amounts of flammable liquid being dumped onto the ground or in the general area of the engine. This is an obvious fire risk. What’s even worse is that if the engine stops for any reason, the flow of fuel will continue to flood into what is now a red-hot steel tube. That can result in a very impressive fireball that could also be very dangerous. Injection Virtually all engines over 20lbs of thrust use direct fuel injection rather than atomization. In such a system, the fuel is squirted directly into the engine’s combustion chamber under some form of pressure. This makes the engine’s operation far more reliable and adds the additional benefit that by varying the amount of fuel being injected, the engine’s power can be varied. Yes, a throttleable pulsejet! The Argus V1 engine used direct injection but, to the best of my knowledge, no attempt was made to provide any form of throttle control – not that it would have been of any use on a flying-bomb anyway. The downside of fuel injection is that you need some method of pressurizing the fuel to force it into the engine in a fine spray. There are really only two options – use a fuel pump or pressurize the entire fuel tank.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 32

The V1-flying bombs used the latter option and the fuel tank was pressurized using the same compressed-air source which drove the missile’s gyroscopes and other onboard systems. Most of my injected engines use propane as a fuel because this has the advantage of being selfpressurizing. Your common BBQ tank has around 100psi of pressure in it so you can use this for direct injection without the need for a supply of compressed air or a fuel pump. Using such a system, the pulsejet remains a stand-alone engine that requires no extra bits and pieces to keep it running. The simplest injection system for a petal-valved engine simply involves locating a cross-drilled injection nozzle directly behind the valve-retainer plate. This nozzle is drilled so that the incoming fuel is sprayed out directly towards the side of the combustion chamber. This ensures optimum mixing with the air and (in the case of liquid fuels) means that any droplets of fuel that aren’t vaporized by the incoming air will be instantly flashed into vapor when they hit the hot combustion chamber walls. A more recent innovation I’ve come up with however involves placing an additional disk behind the valve retainer, separated by just a small space. By injecting the fuel in the same radial pattern as with the previous system but between the two disks, the fuel is not only vaporized more effectively but also serves to cool down the valve retainer disk (and the valves). Building a system like this does however, require access to a lathe in order to turn up the key component which is this radial injector nozzle. Using this double-disk setup I’ve been able to double the life of the reed valves used in a petal valve engine while also slightly increasing the engine’s performance and throttle range. Timed Injection One disadvantage of direct fuel injection is that simple systems such as the one used in the Argus V1 engine tend to spray fuel throughout the engine’s operating cycle. Fuel will only burn efficiently (or at all) when mixed with exactly the right amount of air. This combustible mixture of air to fuel is referred to as the “stoichiometric ratio” and it varies depending on the type of fuel being used.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 33

It makes little sense therefore, to waste fuel by injecting it when there is no incoming air to mix with it as that fuel will be unable to burn inside the engine thus contributes nothing to the thrust being generated. Back in 1947, the guys at Princeton University came to this same conclusion and suggested that using timed fuel injection would be a way to improve the fuel-efficiency of pulsejet engines. Now there are two ways in which timed fuel injection could be done: the simple way and the complex way. Given that the simplicity of a pulsejet is its single greatest virtue, I’m all in favor of keeping a timed fuel injection system simple too. I regularly get email from people who think it would be a good idea to use an electrically driven fuel injector like the ones used in modern car engines – but I disagree. In order to make one of these injectors work you’d need a rather complex system that involved a battery to drive the injector, sensors to measure the pressure inside the combustion chamber for timing, and some electronics to tie the whole thing together. This setup, although I’m sure it could be made to work, would be costly, complex and offer only minimal benefits over the system I use to obtain timed fuel injection. Fortunately it is a simple job to synchronize the injection of fuel into the engine with the intake of a fresh air charge. This is because the pressure inside the engine falls to below 1 atmosphere (14.7psi at sea-level) during the intake phase and rises to as much as twice atmospheric (30psi+) during combustion and exhaust phases. A valve placed over the fuel jet is sufficient to provide a degree of injection timing and the addition of this mechanism can provide a noticeable improvement in the fuel-efficiency of a large pulsejet. I’ve experimented with a number of different valved injectors ranging from a simple bolt drilled length-wise with a flap of spring-steel over the end like the one illustrated here… To this carefully machined injector made from stainless steel and nickel-plated steel components I fitted to the 100lbs-thrust engine on my gokart. I noted a very definite improvement in the fuel-efficiency of this engine after fitting the timed injector system. What Fuel is Best? One of the great advantages of pulsejet engines is that they can, at least in theory, be made to run on almost any type of combustible liquid or gas.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 34

Pulsejets aren’t limited to liquid or gas fuels however – on at least two occasions, coal dust has been used as a fuel. It is rumored that the Germans attempted to run the Argus V1 engine on coal dust when liquid fuel supplies became almost unobtainable near the end of WW2 and some of Reynst’s pulsed combustors were designed specifically to use this unusual fuel. Before you start worrying too much about what is the best fuel, it’s worth citing part of a report published by Princeton University in 1947 that summarized a large amount of the research done into pulsejet engines up to that time. It said “the pulsating jet engine of contemporary design ran on almost any common fuel with negligible variations in performance.” The only caveat the report included was that “principal [sic] differences were in the degree of body heating and the rapidity of valve destruction.” They found that even the use of exotic fuels such as nitropropane or nitromethane offered only a slight power increase at the expense of doubling an engine’s fuel consumption. It makes sense therefore to choose your fuel on the basis of whatever’s cheapest or most convenient to use. For most of us however, the choice of fuels is fairly simple and boils down to one of these: Gasoline This has the advantage that it’s relatively cheap, very easy to obtain, and is pretty clean burning. It’s also quite volatile so atomizes easily to promote easy starting. Note that, contrary to what you might think, higher-octane gasoline is not going to produce any more power than regular gasoline. In fact (in theory) it may produce slightly less power. If you plan to use gasoline, just use whatever’s cheapest. Propane (LPG) Thanks to the popularity of gas-fired BBQs, propane has also become quite easy to obtain and suitable 20lb refillable tanks can be bought for well under $50. In some countries, propane is even cheaper than gasoline and it burns very cleanly indeed – leaving no smell and very little residue at all. Despite the fact that it’s stored under pressure, it is actually quite a bit safer to use than gasoline because its vapors dissipate very quickly in the open air. Since the boiling point of propane is well below normal room temperature, it either comes out of the tank as a gas (thus avoiding the need for vaporization) or, when drawn off as a liquid, instantly boils into a vapor. This makes a propane-powered pulsejet one of the easiest to start. Note that bigger engines will almost certainly demand to be fed with liquid propane because an average BBQ tank simply can’t provide gas at a sufficiently high rate to keep up.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 35

If you’re planning to use a BBQ tank of propane as a fuel, you’ll have get rid of the regulator that is normally used to limit the flow of gas. This regulator reduces the pressure of the gas to just a few psi, far too low for a pulsejet’s needs. To give you an idea of just how much gas is needed to run a pulsejet, my own 15-lbs-thrust engines (PJ15) will drink all the propane gas you can feed them – with the regulator removed. The Lockwood valveless engines will drink all the liquid propane you can feed them without any regulator in place. If you try to use propane without removing the regulator then all you’ll get from a pulsejet is a few bangs and pops – it won’t run. However, you will still need some form of control over the flow of gas into the engine and for this I recommend buying a cheap propane/air torch – of the type often used for soldering or brazing. These torches are available from almost any hardware store and cost just $25-$30. Note that depending on the exact make/model of torch you buy, you may need to purchase an additional adapter fitting so that it can be screwed directly onto a 10lb or 20lb propane tank. To use a torch like this as the gas-control valve for your pulsejet, simply unscrew the burner fitting on the end and slide your propane-certified plastic fuel pipe over the end, securing it with a small hose-clip. The gas-flow knob on the torch will now enable you to control the amount of gas that is delivered to your engine. If you invert your BBQ tank of propane, the torch will still serve as a very simple way to control the flow of liquid propane to larger engines. Very simple, very inexpensive, and very effective. Another method of controlling the flow of propane to your engine is to simply use a device called a needle-valve. These valves are readily available from a number of sources and, just like the gas-torch, offer a very fine degree of control over fuel-flow. Butane It should be noted that although it is also often sold for use on small camp stoves, butane is not a good substitute for propane. It contains less energy and doesn’t produce as much pressure as propane at room temperature. In short – don’t waste your time or money trying to use butane as a fuel for pulsejet engines. White Spirit/Coleman fluid This is simply a very low octane unleaded form of gasoline which has no fancy anti-knock or combustion “improvemnet” additives included. It’s actually a better fuel than high-octane

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 36

gasoline for pulsejet use. Many of the early small pulsejet engines such as the Dynajet run best on this fuel. Methanol This is my second-favorite pulsejet fuel. It has the advantage that it will burn over a very wide range of rich/lean mixture settings – making an engine less sensitive to fuel head or starting conditions. It also burns very cleanly with no smelly or oily residue and creating little more than water vapor and some C02 as combustion byproducts. On the downside, methanol is more expensive than gasoline, your engine will burn more of it for a given amount of power, and it can be very dangerous if spilled because it burns with an almost invisible flame. Many people have been burnt because they’ve walked straight into a methanol fire without seeing it. Despite the downsides, I prefer to use methanol for all my aspirated engines because it generates a little more power, allows the valves to run cooler, and doesn’t leave my hands stinking of gasoline. Note that you shouldn’t use pre-mixed model airplane fuel instead of straight methanol. Model airplane fuel contains up to 20% oil that will leave significant deposits inside your pulsejet and also affects the vaporization of the mixture. It’s also a lot more expensive than plain old methanol so you’ll be wasting money. Your local hot-rod or drag-racing club ought to be able to help you find a source of methanol but if all else fails, try one of the major oil companies like Mobil – they sell me 5-gallon drums of the stuff when I want it. Another thing to watch when using methanol as a fuel is that it is very hydroscopic – which is to say that it tends to absorb moisture out of the air. If you leave a can of methanol uncapped then it may well absorb so much moisture that its combustibility is affected and this can result in hard-starting. Also be aware that when you use methanol as a fuel, one of the combustion byproducts is water (albeit as water vapor). This means that the spring-steel reed valves used in an engine run on methanol are prone to rusting unless you oil them lightly before storing your engine after each run.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 37

Constructional Techniques Once you’ve calculated the dimensions for a pulsejet, how do you then go about building one? Of course It really helps if you’ve got access to a workshop or some basic metalworking tools such as a hacksaw, drill, welder, etc. but don’t be put off if your resources are a little more modest. You’d be surprised how helpful your local welder or engineer can be when you explain that you’re building a jet engine and would be happy to demonstrate it to them when it’s done. It’s also amazing what you can do with a minimum of tools – if you’ve got enough patience. The Engine Body/Tailpipe Most commercial pulsejets are made from thin stainless steel sheet that is rolled or otherwise formed into tubes and cones before being welded together. This results in a durable engine that is light enough to be practical for such uses as powering model airplanes. These cones and tubes are formed using a device known as a slip-roll which looks rather like an old washing-machine wringer and consists of three rollers that can be adjusted to both grip and curve the metal sheet as it’s wound through.

A hand-operated slip-roll like this one is limited to rolling stainless steel that is no more than 1mm thick – and even then it’s damned hard work if you’re rolling a piece the full 600mm long which is the maximum this set of rolls can handle. For larger engines it really pays to find someone who has a set of motorized rolls that can handle the thicker material and longer lengths you’ll need to use.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 38

Unfortunately, stainless steel is not only expensive but can also be very difficult to weld when it’s very thin. For this reason, many enthusiasts prefer to make their engines from cheaper and more easily worked materials Providing weight isn’t a problem you can save a lot of time and significantly simplify construction by using mild steel pipe for the body of your engine. The best stuff is exhaust tubing which is usually protected from rusting by a thin layer of aluminum on the surface. This stuff is relatively cheap, available in a wide range of sizes and can be cut and welded easily using MIG, arc or oxy-acetylene equipment. Unfortunately it’s also quite heavy, but that’s usually unimportant to the eager enthusiast who simply wants to build an engine and get it running ASAP. Here’s a picture of my gokart with an engine built from exhaust tubing. It only produced just enough thrust to get the kart moving on a smooth surface but it was an interesting experiment none the less. As you can see, this pipe is quite thick-walled which accounts for its weight and easy welding characteristics. The Valving System Exactly how you construct this depends on the type of valving system you plan to use. If you have access to a lathe then you can easily make a petal-valve system with a nicely turned aluminum head. If you’re not lucky enough to have a lathe at your disposal, all is not lost. Instead of making the front of the engine from a single, large piece of aluminum rod, you can cut a circle from a sheet of plate. Even on small engines it pays to use a piece at least 3/8” (8mm) thick for this. Because aluminum is such a soft metal, you can actually do a pretty good job of cutting a circle from a flat sheet by using a jig-saw. You can even use a small coping saw or jeweler’s saw to cut the valve holes (after drilling a starting hole first). To create a circular valve-plate from a flat piece of aluminum, simply mark out your circle using a compass then cut it slightly over-sized. It can then be filed down to a precise fit into the front of your engine. However, before you cut out the circle, it pays to mark out and drill the valve holes. By doing this before you cut the circle to shape, you have a larger piece of metal to hold onto when drilling the valve holes and this makes the job much easier.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 39

Here’s a valve-plate that was made from flat-sheet, cut to shape with a jigsaw and with valve-holes that were cut with a jeweler’s saw. In this case, a lathe was used to finish the surface of the disk and it was then anodized to provide a hard, corrosion resistant layer and a pleasing gold color. See the chapter on anodizing for details of how to perform the anodizing process. Note that there’s no reason why the valve-holes should be circular and a trapezium shape as in the picture actually allows a greater valve area for a given size of valve-plate. Having made the valve plate a snug fit in the front of your engine, you can then drill 4-6 small holes around the circumference of the pipe so that they also go through into the edge of your valve plate. By choosing the correctly sized drill, you can then fit self-tapping screws to hold the valve-plate firmly in place. Any leaks in this area can be fixed by the liberal application of some muffler-sealant – the type that comes in a small tube and is designed to block-up gaps in exhaust systems. This stuff will expand slightly as the engine heats and seal any small gaps between the valve plate and the engine body. If you don’t have a lathe then the other option is to build a V-valve system instead of a petal valve one. The V-valve can be created from flat pieces of steel as in this drawing. Making Reed Valves Since the reed-valves are the heart of a conventional pulsejet engine, it’s important that they are well made and that the right materials are used. Most valves are made from high-carbon spring steel of between 0.006” and 0.012” thickness. If you can’t find a source for this material locally then I suggest you take a look at some of the online mail-order metal supply companies. Once you’ve got the right material, the next problem is cutting it to the required shape. Spring steel sheet is incredibly brittle and will split or crack very easily if you try to cut complex shapes with regular metal snips.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 40

While it’s simple enough to cut the rectangular shaped valves used in a V-valve system, creating the intricate shape of a petal valve represents more of a challenge. There are two methods you can use to fabricate a petal valve from spring-steel sheet. The first involves using a Dremel or similar tool fitted with a cut-off tool as shown in this picture. This technique requires a bit of practice and is most suitable for smaller valves. The preferred method for making petal valves involves the use of a process known as electrochemical etching. Full details of how to make reed valves using this process are provided later in this book. Welding There are really only two welding options when it comes to joining al the pieces together: • •

TIG (tungsten inert gas) MIG (metal inert gas)

The ideal welding process is TIG, since this allows total control over the amount of heat used and how much (if any) extra filler metal is added to the weld seam. However, since TIG welders are more expensive than simple MIG units, and most people find MIG welding much easier than TIG, I’ll describe both processes. Welding with MIG I won’t turn this into a welding tutorial so I assume you’ll already be moderately competent with a MIG welder. Instead, I’ll focus on the points specific to welding the various parts of these engines together. It is important to use the right filler-wire. If you’ve cut and rolled the parts out of 304 stainless steel then you can use 308 or 316 stainless filler wire. If you’ve cut your parts from 316 stainless then use 316 filler rather than 308. I’ve found that 0.8mm (0.032”) wire is about the right thickness and you should only need a small spool to build an entire engine. While you can use plain steel wire, the resulting weld will rust very quickly and become weak, so it’s not recommended. Stainless steel also has a much higher rate of thermal expansion than plain steel so you’ll get additional stresses set up as the engine heats and cools.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 41

When welding, adjust the current, wire-speed and stickout used to try and get the flattest weld bead you can without burning through. I find when working with very thin material (< 1mm/0.040”)that sometimes it’s easier to overlap the material along the weldseam. If you plan to do this then don’t forget to allow for that overlap when marking and cutting your pieces to size. It also helps to have a “chill bar” behind the weld seam whenever possible. A length of 25mmx25mm (1”x1”) aluminum bar does a good job here, with a length of 12.5mmx25mm used along the top surface to ensure intimate contact between the two sides. A pair of two C-clamps can be used to hold everything together as in this picture. Welding with TIG TIG welding thin stainless is not as easy as you might think. The use of a chill-bar (as described in the MIG welding section above) is absolutely essential. Without this, the stainless will tend to sag and form ugly black “danglers” on the back-side of the weld. This not only produces a weaker weld but those danglers will also interfere with the gases that flow at very high speed through these tubes. Unfortunately, unless you want to go to the bother of turning up conical chill-bars to match the radius of each circumferential weld, you’ll have to just do your best without one. Backpurging the weld with argon will certainly help here, as will making sure that the fit-up of the pieces is as near to perfect as you can get it. Another option is to use a special high-temperature flux paste designed for use with stainless welding. This usually comes as a powder that is mixed with methyl alcohol then spread on the backside of the weld to stop oxidation during the welding process. The brand I’ve used with some success is “Solar”.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 42

Unless you’ve turned up those conical chill-bars, you’re probably going to have to overlap the cones slightly to get a good fit-up, since the rolling process combined with the stresses produced when welding the length-wise seam, will almost certainly mean that the cones themselves are not perfectly circular in cross-section. When welding, try at all times to keep the weld bead the same thickness as the sheets being joined. It’s obvious why you would not want the weld to be too thin – but read the piece on heat-stress later in this bookto find out why a thick weld bead is just as bad. Getting a Good Fit-up A good fit-up is essential to the success of any welding job and there are some simple methods of getting a good fit-up when assembling these engines. Butting a cone up to a matching tube in such a way that the edges match perfectly so as to ensure a good weld is a very difficult (some would say impossible) task when using very thin material – but fortunately there are alternatives. I always flare or flange edges of my cones to ensure the maximum contact area between the two parts and to ensure that the springiness of the stainless steel actually works to keep the two surfaces in intimate contact for welding. There are two ways of flaring and flanging – you can use a rotary machine (sometimes called a “jenny”) or you can just beat the snot out of it with a hammer and dolly. If you’ve got a jenny then you’re probably already aware of the techniques used to create flares and flanges – but if you don’t, here’s how you do things the manual way: First you’ll need a good hammer (or two) and a dolly. Although I use the term “dolly”, I use nothing more than a 100mm (4 inch) long piece of 25mm (1 inch) steel bar. I’ve ground a fairly large (6mm - ¼ inch) radius onto one end and a small recess onto the other [PICTURE]. This piece of metal is then used as a backstop as you beat the stainless into submission, carefully working around the edge of the cone until the desired flare/flange is obtained. If you do a good job, the cones and tubes should stay together before they’re welded without the need for clamps or other mechanical devices.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 43

Powering things with pulsejets Pulsejets have been used to power just about every type of vehicle you can think of. Gokarts My own simple gokart has been pushed along at various speeds by a simple exhaust-pipe pulsejet, a much larger 100lbs-thrust pulsejet similar to the Argus V1 design, and a Lockwood valveless engine. What I learnt from this is that on a smooth, flat surface a gokart is going to need at least 30lbs of thrust to be much fun. If the surface is a little rougher or if there’s any kind of slope then you’re going to need at least 50lbs of thrust or more. Once you get to 100lbs of thrust, riding on a pulsejetpowered gokart becomes a very “interesting” experience and not something recommended for the faint of heart. The picture above is that of the pulsejet-powered kart I helped build when I was expert for a team of Royal Navy Engineers on the popular Scrapheap Challenge TV series. Model Airplanes But perhaps the most common type of craft to be powered by pulsejet engines are model airplanes. Back in the 1950s, many Dynajet engines were sold for use on control-line speed models, many of which reached speeds well over 150mph. These days there are a growing number of radio controlled models being fitted with pulsejet engines as people are attracted to the simplicity and comparatively low cost of these engines when compared to more expensive turbine units. There are problems associated with the use of conventional pulsejet engines in model airplanes however – not the least of which is that of providing a constant fuel supply under varying G-forces. The most effective methods used to achieve this essential steady flow of fuel are either the use of a pressurized bladder tank or the addition of a small electric fuel pump and the use of direct-injection. See the chapter on fuel systems for more information on bladder tanks.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 44

Another component finding great favor with aeromodelers is a device known as the Cline regulator. This is a “demand regulator” that only allows fuel to flow when there’s suction applied to the outlet connection. The advantage of this for pulsejet flyers is that it avoids an undesirable situation which could see an engine stop running unexpectedly but continue to be flooded with fuel from a pressurized system. Boats Yes, believe it or not, pulsejets have been used to power both model and full-sized speedboats. Back in the late 1950s an enterprising guy by the name of Bill Pearson fitted two 105-lbsthrust pulsejet engines to a small speedboat and had some fun. According to an article published in Popular Mechanics, the boat had a top speed of more than 45 miles per hour but consumed an enormous six gallons of gasoline per mile. The other problem with this jet-powered boat was the fact that it could be heard from miles away. Not exactly the type of boat you’d want for a quiet afternoon’s fishing and I pity anyone who might have tried to use it as a ski-boat.

Here’s another boat powered by a huge Lockwood valveless pulsejet – watch for it on your TV screens sometime in 2004. No, I didn’t build it or design the hull but I did design the engine. Unfortunately it was not built very well and failed to produce anything like the power levels it should have but you can still see that it does a good job of chopping up the water behind it.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 45

Helicopters Back in the 1950s a number of experimental helicopters were designed that used tip-mounted pulsejets to power the rotors. None of these ever saw commercial production and they all suffered from a number of problems. The arrival of the gas turbine turboshaft engine effectively killed off such ideas but there are a number of places that will sell you plans for such suicide machines if you have a death wish. The NACA even published a research paper on the effects that the powerful centrifugal forces experienced by a tip-rotor pulsejet would do to its combustion efficiency and reliability. Civil Aircraft Although I’m not aware of any civilian light airplane that has used a pulsejet as its primary source of power, valveless pulsejets were once used (mainly in Europe) to power highperformance sailplanes. Because these craft are so aerodynamically efficient, only a very small amount of thrust is required to keep them airborne meaning that pulsejet power was almost practical despite the high fuel-consumption.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 46

Schmidt’s Contribution Earlier in this book I used Tharratt’s formulas, mainly because of all the designers, his work is the most readable and easily applied by those without math and physics degrees. However, Tharratt wasn’t the only researcher who spent time working with pulsejets and many others have come up with all manner of findings in relation to these engines. Perhaps the most notable of these other pulsejet designers was a German by the name of Paul Schmidt who was responsible for much of the Argus V1 engine. Schmidt laid a lot of the groundwork from which others such as Tharratt made their observations. In fact I’d wager that there has been no other pulsejet engine in history that has undergone as much prodding, poking and analysis as the Argus. In the years after WW2 the US NACA invested a huge amount of time and effort investigating the potential of pulsejet engines and much of their work centered around copies of the Argus. The US Air Force even commissioned Ford to make a run of Argus engines to power a planned 1,000 “flying bombs” of their own carrying the designation JB-2. Even the US Navy had its own version of the JB-2 which was more commonly known as the “Loon”. Many of the NACA reports into the design and operation of the Argus have been released into the public domain and have been included in the appendix of this book or can be downloaded from the NACA archives on the Internet. Like Tharrat, Schmidt was very much enamoured of engines designed around straight or slightly divergent pipes. One such engine, designated the SR500, was little more than an almost straight tube. I say “almost straight” because it was actually a cone that diverged by 2 degrees. With a total length of around 11 feet, this divergence meant that the, 18 inch diameter at the front became 22 inches at the rear, Despite it’s simple shape, the static performance of this engine appears to exceed that of the Argus eventually used to power the V1. The SR500 was reported to develop a static thrust of some 1,500 lbs which would result in a performance of some 3.2 lbs thrust per square inch of mean cross-sectional area (some 50% better than Tharrat’s constant). This would make the SR500 very efficient for a conventional engine. I should not that I’ve not had first-hand access to these reports but have no reason to doubt their veracity. Schmidt observed an interesting phenomenon during his pulsejet development work – a rapid mode of combustion that appeared to be much faster than regular deflagration. He went on to spend much effort in trying to achieve true detonation but was ultimately unsuccessful.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 47

It is possible that this rapid combustion phenomenon goes some way to explaining the seemingly extraordinary performance of the SR500.

My own experiments into promoting the simultaneous combustion of the entire air/fuel charge within a pulsejet have also yielded very promising results. It looks as if Mr Schmidt really knew how many beans makes two. There is a lot more about Schmidt’s work along with the history of the V1 and Argus engine in the other historical files on this disk.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 48

Ignition Systems

One of the great things about pulsejet engines is that, unlike the engine in your car, once started, they don’t require any regular source of timed ignition to keep running. However, if there’s one part of pulsejet starting that leaves many people scratching their heads it’s how to create the spark needed to initially ignite the air-fuel mixture during that start-up period. I fit spark plugs to all my engines and have built a simple electronic circuit to create a suitable spark – but if you feel you don’t have the necessary skills to do this there are alternatives. One method I’ve used before is still electrical but just uses parts you can get from any autowrecker or auto-electrical store. All you need is an auto ignition coil, an auto flasher unit and a 12-volt battery. Wire the three items together as in this diagram:

The component labeled C1 is a capacitor (or condenser) that stops arcing across the contacts inside the flasher unit. Any auto parts store or wrecker’s yard should be able to supply this but if not, you can use a 0.47 to 1.0 mfd capacitor rated to about 250V from Radio Shack or some other electronic components store. It’s pretty easy to see that the signal light flasher turns off and on at regular intervals and this on/off current flow causes the coil to create a nice bright spark. There’s not much to go wrong here and there aren’t many mistakes you can make wiring it up so have a go. It might be worthwhile adding an off/on switch and a fuse in the lead that goes from the battery to the coil + connection.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 49

However, if you’re comfortable working from electronic circuit diagrams you might want to build the circuit below.

This simple circuit will drive a regular auto coil and produce a continuous stream of about 5-8 sparks per second. By comparison, the previous circuit may only produce one spark every second or so. Another alternative is to dispense with a spark plug all together. Some people start their pulsejets by simply placing a lighted fireworks sparkler in the tailpipe. I can’t say I’ve actually tried this myself but I don’t see why it wouldn’t work. Apparently, once the engine starts, the sparkler wire is blown out the back so doesn’t interfere with the engine’s operation. Of course if you’re working on a new design or have an engine that refuses to start, you could waste a lot of money on sparklers using this technique so having an electrically operated ignition system is still a good idea. One point that is seldom discussed is exactly where do you put the sparkplug? Once again – it’s compromise time!

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 50

If you put the sparkplug towards the back of the combustion chamber then the engine will probably start more easily because, by the time the air/fuel mixture reaches this point it will be more thoroughly mixed and thus more explosive. However, the the further back you place it, the greater the chance that the protruding plug will interfere with the passage of the fast-moving exhaust gases. It’s worth noting that the original Dynajets had their sparkplug right at the back of the combustion chamber just where it starts to taper down to join with the tailpipe. Later versions saw the plug moved much further forward. I tend to stick mine half way along the chamber and have no problems starting, nor do I see any significantly adverse effect on power output. And while on the subject of sparkplugs, what type should you use? Well there are special miniature plugs available from companies like Champion, but I’ve not been able to find a source of such plugs at anything like a reasonable price. So, I tend to use commonly available plugs that are originally designed for use on chainsaws, weed-wackers and the like. These plugs are much smaller than regular automobile units, but still use the standard 14mm x 1.25mm thread. On a very small engine they may look somewhat out of place but on anything larger than an 8lb-thrust unit they’re just fine. Note that you will want to open up the gap to about twice the factory setting before use. This will ensure easy starting. I’m sometimes asked whether it’s possible to use a model airplane engine glow-plug instead of a sparkplug. To be honest, I haven’t tried this but I’m pretty sure it would only work if you were using methanol as a fuel and even then you’d probably find that it was pretty marginal.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 51

How to start a pulsejet Pulsejet engines have a sometimes well-deserved reputation for being very difficult and problematic to start. Fortunately this doesn't always have to be the case. In order to start a pulsejet you need three things: 1. fuel 2. air 3. an ignition source Not only must you have all three -- but they must also be provided at the correct time and in the right proportions. Fuel Pulsejets can run on a wide range of fuels ranging from LPG/propane through to diesel or kerosene. For the purposes of small pulsejet engines however, the most common fuel is gasoline of some kind. This can be white-gasoline or low-octane gas from the local pumps. You can use highoctane gasoline but you'll simply be wasting money and possibly getting a little less power at the same time. In cold conditions where the air temperature is less than 60 degrees F or around 17 degrees C you may find that gasoline isn't sufficiently volatile to ignite reliably when starting an engine. If this is the case then it is recommended that you add some ether -- up to 25 percent. This will significantly increase the ease with which the fuel can be ignited. Cold can also affect engines that are using propane/LPG because the pressure available from a tank of this gas reduces quite significantly as the temperature drops. It's worth mentioning at this stage that there are two methods of delivering the fuel to the engine: aspiration This is when the fuel is drawn into the engine through an atomizer by the air which enters through the intake. This has the advantage that it is very simple and requires no fuel pump or other ancilliary equipment. injection This involves spraying fuel directly into the combustion chamber where it mixes with air that has already passed through the valves. This has the advantage that you can throttle the engine simply by varying the amount of fuel injected -- but it does require the use of a pressurized fuel system such as a bladder or pump. The Air Supply

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 52

Just supplying the engine with fuel is not enough -- you need to force some air into the intake so as to create an explosive mixture inside the engine. In the case of an injected engine, an air source such as a leaf-blower or vacuum-cleaner with a blow attachment will likely do the job. You don't need a lot of pressure but you do need a reasonable volume of air. With an aspirated engine you'll need less volume but more pressure. This is because the air has to draw the fuel up the fuel-line and atomize it into a fine spray before it passes through the valves into the engine. An air-gun driven by a compressor is perfect for the job. However, if you don't have such luxuries available to you a great substitute is to inflate a car tube or tire without a valve in it -using a length of flexible plastic tubing slipped over the valve stem to deliver the air to the engine's intake. The flow of this air can be controlled by kinking the tubing. Slipping a thinner piece of pipe into the open end of the tubing will give you a narrower and more easily controlled air-jet to spray into the engine. Between starting attempts you can replace the valve and pump the tire up to 40-60 psi using a foot-pump or whatever. This is a great low-cost way to start your engine at the flying field or away from other sources of compressed air. The Ignition Source The best ignition source is a spark plug mounted in the combustion zone section of the engine – but there are alternatives... The simplest but least effective ignition source is a naked flame situated at the end of the tailpipe. This could be a gas-torch or a spirit burner but you'll find starting an engine using this method to be more difficult than with a sparkplug. One other technique sometimes used when a sparkplug isn't mounted in the engine itself is a spark-wand. This consists of two wires, separated by an insulator with a spark-gap at the end. It is inserted and energized when starting the engine and quickly withdrawn once it is running. A third option is to simply place a lighted fireworks sparkler into the engine. This will burin for about 20 seconds and provide a nice hot ignition source. Try to place the sparker as far towards the front of the engine as possible. It will be blown out when the engine starts. Check the chapter in this book on ignition systems for more information and the plans for two electrical ignition systems. Putting It All Together Here's the sequence for starting an aspirated pulsejet: 1.Connect the fuel line and tank. Make sure that the fuel level is no more than an inch or so (20mm) from the engine's fuel jet (the hole where the fuel comes out into the engine). 2.Turn on the spark or light your ignition source

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 53

3.Direct a jet of compressed air over the fuel jet so that it creates a fine spray of fuel droplets that are then blown towards the engine's valves. Note that it will take some experimentation, practice, and coordination to get this right and you'll probably find that it helps to move the jet of air back and forth a little so as to vary the spray a somewhat. If your engine has a “flowjector” setup like the Dynajet then you won’t be able to vary the angle or position of your air source very much. At this stage the engine should at least pop, bang or burb a little, even if it doesn't immediately burst into life. If you don't get much activity, try richening the mixture a little (if your engine has this capability) by losening the locknut and opening the mixture screw by a quarter of a turn. Repeat this process until things improve. If they don't improve, return the screw to its original position and try closing it a quarter turn at a time in case the mixture is already too rich. The procedure is even simpler for an engine with direct LPG/propane injection. Thanks to the fact that the fuel is already in the engine's combustion chamber, all we need do is turn on the spark and blow some air into the intake. If the engine doesn't fire immedately then the gas should be turned up or down until the engine starts. If you still have trouble, try varying the amount of air being blown into the intake. Once you get the hang of starting a LPG-injected pulsejet its exceptionally easy to do

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 54

Valveless Pulsejets While most people marvel at the simplicity of the humble pulsejet with its single moving part, there are designs that are even simpler. I refer of course, to the incredible valveless pulsejet engine. A device that has absolutely no moving parts at all! The design of valveless engines is even more of a black art than the design of regular pulsejets, as witnessed by the amazing diversity of shape and form to be seen amongst the various valveless designs. If you’ve performed the jam-jar experiment detailed elsewhere in this book then you’ve already seen a valveless pulsejet engine in action, otherwise here’s a brief description of how they work. Regardless of the shape, form and size of a valveless pulsejet engine, they are all basically a combustion chamber with a pipe or pipes attached. Inevitably, one or more of the pipes is much shorter than the other(s). The short pipe is often referred to as the intake and the longer is referred to as the exhaust. When the air/fuel in the engine explodes, hot gases rush out both pipes (yes, even the intake) at great speed -- just as they rush out of a conventional pulsejet’s tailpipe. When Mr Kadenacy pokes his head in to see what’s happening, the rapidly exiting hot gases create a partial vacuum inside the engine. At this stage, the gas-flows reverse and the engine starts sucking back through its exhaust and intake pipes. Now, since the intake pipe is much shorter than the tailpipe, fresh air is able to be drawn into the combustion chamber. It is then mixed with fuel (usually directly injected) and we have a combustible mixture all ready to fire. Meanwhile, gases continue to flow into the other end of the combustion chamber from the exhaust pipe. Since the exhaust pipe is much longer than the intake pipe, these incoming gases contain no fresh air – just hot reminants of the previous combustion cycle. As soon as these hot (still burning) gases hit the fresh air/fuel mixture – boom – the whole cycle starts again. Advantages It’s pretty easy to see that the biggest advantage of a valveless pulsejet engine is the fact that it doesn’t have any delicate valves to wear out or break. All else being equal, the operating life of a valveless engine is limited only by the amount of fuel available and the durability of the tubing from which it’s made.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 55

Without the need for valves, such an engine is also, at least in theory, much easier to make. No holes to drill, no spring steel to etch – just pipes and cones to weld together. Disadvantages Oh boy, didn’t you just know this wasn’t going to be a free lunch? That’s right, there’s always a price to pay in the pulsejet world and in return for its improved simplicity of construction and operation, the valveless engine suffers from lower levels of performance and efficiency. Why is this? Well you’ll recall that a little earlier on I mentioned something called “post ignition confinement” (PIC) and mentioned that it sometimes had a beneficial effect on an engine’s efficiency. In the case of a traditional (valved) pulsejet engine, there’s always a reasonable degree of PIC because before the burning gases can expand, they’ve got to push all that gas in the tailpipe out of the way. This slug of cold tailpipe gas provides a form of “inertial confinement” that ensures more efficient combustion occurs in the front part of the engine. However, when we look at a valveless engine – there are now two (sometimes more) gaping holes in the engine – and in the case of the relatively short intake pipe, there’s not a lot of cold gas there to provide any reasonable level of inertial confinement. So, when the air/fuel mixture in a valveless engine ignites, the gases are able to escape before they can create the same pressure levels as you’d find in a valved engine with its much smaller escape route and the larger inertial mass in its tailpipe. It is for this reason, plus the problems associated with making a valveless engine into a sensible shape, that engines like the Dynajet continue to dominate the small pulsejet arena. Perhaps the only valvless pulsejet that has ever enjoyed any success is the oddly-shaped Lockwood Hiller design. By bending the engine in half, Lockwood was able to have the intake and exhaust pipes facing in the same direction so that they both contributed to the total thrust. Unfortunately this engine has never lived up to the claims that were made for it. Most notably it’s claim to provide an amazingly low (for a pulsejet) fuel consumption figure have been independently verified despite numerous engines having been built exactly to Lockwood’s own plans. Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 56

Just to prove to myself that this weird engine doesn’t live up to the claims made for it, I built one that was supposed to produce 75lbs of thrust. No matter how hard I tried, I could only coax between 50 and 55 lbs of thrust out of this beast and it used nearly four times the amount of fuel claimed by Lockwood. Several other people have verified this lack of thrust and thirst for fuel so I think the results are quite conclusive. Of course having built such an engine, I had to do something with it, so I stuck it on my gokart just for fun. The only problem is that a full BBQ tank of propane only lasts about 5-6 minutes. Here are just a few of the various valveless pulsejet engine designs that have been created over the years: The Reynst This design was never intended to create propulsive thrust and was intended to simply be a more efficient way of burning fuel to create heat. It’s included here mainly because it’s a great example of the operating principles behind pulsejet engines and because its inventor (Reynst) was such an important part of pulsejet history. The Marconnet Designed almost 100 years ago, this engine, named after its inventor Frenchman Georges Marconnet, is one of the oldest valveless designs. It seems to have relied on having a smaller diameter intake tube than exhaust tube so that while some of the combustion gases escaped out the front, even more rushed out the rear, producing a thrust imbalance that pushed the engine forwards. [diagram] The Escopette This was a logical evolution of the Marconnet and a similar engine known as the Resojet, the major difference being that the Escopette’s intake tube is bent through 180 degrees to point Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 57

towards the rear. By doing this, the gases that are ejected through this tube contribute to the total thrust. [diagram] The Lockwood This “bend it in half” theme was taken further by Lockwood except that he bent the exhaust tube on his design which meant that the overall length of the engine was reduced by half. This compact size and the ready availability of plans has meant that Lockwood’s engine is perhaps the most popular valveless design amongst amateur pulsejet builders. The full patent application for the Lockwood engine is included in an appendix at the end of this book for those who are interested in the exact theory of its operation. The Chinese I don’t know how this engine got its name but it’s a simple design that follows the traditional valveless design of having a long tailpipe and a short intake tube, both of which face towards the rear of the engine.

From all reports, this engine is a rather poor performer, producing barely half the power of a Dynajet despite being almost twice the size. Plans for a version of this engine are included in the appendix of this book.

The Thermojet Although it looks rather cool, the Thermojet was reportedly something of a disappointment in terms of the amount of power it generated. As you can see, it’s not too much different to the Chinese design except that it has two intakes that are parallel to the tailpipe. Note also that the tailpipe has a fairly large flared cone. Some variants of the thermojet were built with as many as four intake tubes.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 58

The Calgary University Design This is actually just a variation on the basic Lockwood design. However, instead of bending the exhaust or a single intake tube through 180 degrees, this engine uses four intake tubes (much as I did with my first valveless engine) and relies on the output of these tubes being fed into “flow rectifiers”. These flow-rectifiers are actually just U-bends (not shown in the diagram below) that are spaced slightly away from the intake tubes (so that intake air can enter through the gap).

J. Kentifield, who played a major part in the design of this engine, claims that this design produces more power per unit of length than a conventional valveless engine and a better TSFC (fuel consumption) figure. Details of the Calgary University engine can be found in a paper published by the AIAA with the reference number AIAA 98-3879. This paper can be purchased from the AIAA’s website at www.aiaa.org

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 59

The Design of Valveless Pulsejets A valveless pulsejet is a deceptively complex device. While it may appear that it’s simply a combustion chamber with a short tube at one end and a longer tube at the other, there are a lot of different factors which affect the way (or whether) the engine runs. Regardless of the exact design of the valveless pulsejet, there are always three basic components that work together to create a pulsating combustion cycle. If we examine the function of these three basic parts it becomes possible to get an idea of how they interact.

The Combustion Chamber This is where the air/fuel mixture is burnt. It is this combustion that creates the heat, and therefore the pressure that results in the rapid flow of exhaust gases through both the intake and exhaust tube. Obviously the combustion chamber can be almost any shape, from long and skinny to short and fat, and indeed there have been valveless engines built with just about every conceivable type of chamber. From an efficiency perspective, it’s important that the combustion chamber be of a shape and volume that the rest of the engine is able fill (but not over-fill) it with a fresh charge of air and fuel during each operating cycle. If the combustion chamber is too small then some of the intake charge will be drawn down into the tailpipe where it may not be burnt as efficiently. If the chamber is too large then it will likely contain too much contaminating residue from the previous combustion cycle and that can also reduce the efficiency. A too-large combustion chamber will also reduce the magnitude of the vacuum which draws in the fresh charge of air and that will mean a further power loss. There are also some gains to be had by employing a combustion chamber shape that provides a fair measure of containment during the combustion phase. In the case of most valveless engines, this is obtained by using a chamber that is significantly larger in diameter than either the intake or exhaust tube. With such a configuration, the area where the combustion chamber narrows to join up with the intake and exhaust causes the combustion gasses to be choked, which raises the pressure in the chamber during combustion. Higher combustion pressures have the potential to provide more rapid combustion and better fuel-efficiency but like most things in the pulsejet world, there are trade-offs.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 60

If there is too much choking then the thrust will be reduced. In fact, one of the interesting things about trying to squeeze gas through a hole at high pressure is that you’re limited by the speed of sound. Once the gas is pushed so hard that it reaches the speed of sound, no amount of extra pressure will convince it to go any faster – so this will mean our engine will not produce as much thrust as it otherwise would. Now I could go into the rather complex math required to calculate the exact size of hole required to achieve the ideal speed of gasflow through the intake and exhaust tubes, but there are a whole heap of variables that are often difficult to accurately predict. The reality is that using a few rules of thumb and a bit of cut and try is sometimes a more practical way to get things going. The Intake Tube The intake tube is the short one. It’s shorter than the exhaust tube because it needs to be able to pass enough air to pretty-much fill the combustion chamber during the engine’s intake phase. If the intake tube were too long, the engine would simply end up sucking back the exhaust gasses that filled it during the last combustion phase. However, the intake tube also needs to be long enough that it can hold a large enough slug of cold, dense air to help contain the combustion gasses during the early phase of the combustion cycle. If the intake tube is too short, then it will contain only a limited mass that of cold dense air when combustion occurs. This means the exhaust gases will rush out too quickly because there’ll be little resistance to their flow. As a result, very little of the combustion gases will actually be forced into the exhaust pipe and, as you’ll see in a minute, this will make it unlikely that the engine will run. The intake tube also has another subtle design feature – it’s often slightly tapered, being wider at the combustion chamber and narrower at the opening through which air is drawn in. This is done for two reasons. Firstly, this taper produces an additional choking effect during the combustion phase – further ensuring that there’s plenty of pressure built up in the combustion chamber and that this pressure forces a goodly proportion of the gases out the tailpipe. Secondly, as the fresh charge is drawn into the engine, this shallow taper acts as a diffuser and produces an increase in pressure within the engine. In order to get maximum combustion efficiency we want the air/fuel mixture to be at as high a pressure as possible before ignition. The taper of the intake tube has a small but useful effect in achieving this.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 61

Another feature of the intake tube is its flared mouth. On smaller engines, this flare can be simply a curved lip, on larger engines it’s another cone. The purpose of this flare or cone is to make it easier for air to enter the intake during the intake phase. Without the flare, a large amount of turbulence forms right at the intake opening and this can choke the flow of fresh air entering the engine – adversely affecting power. The Tailpipe This is the long, usually conical tube attached to the other end of the combustion chamber. The tailpipe serves two important roles. Firstly, it is this tube that accounts for around 60% of the thrust produced by an unaugmented valveless engine. Hot gasses from the combustion chamber exiting through this tube produce a reaction that creates the thrust. Secondly, this is the pump that drives the engine. For this reason, its dimensions are critical. As with a regular valved pulsejet, it is the inertia of the hot gases travelling down the tailpipe that ultimately produces the partial vacuum in the combustion chamber responsible for drawing in the fresh charge of air via the intake tube. There are two critical factors in the tailpipe design: 1. the volume of gas it contains 2. the length of the pipe. In order to provide the necessary pumping action to draw in the next charge of fresh air, the gases in the tailpipe must contain sufficient energy to create a partial vacuum in the combustion chamber. The energy or momentum (signified by the symbol p) of those gases is determined by their mass and their velocity using the formula: p = mV (momentum = mass times velocity) To obtain a given amount of momentum, need to establish a suitable combination of mass and velocity. The velocity of the gas is determined by the diameter of the tailpipe but limited to a maximum of mach 1 (the speed of sound) at the narrowest point. This narrowest point is usually where the tailpipe joins to the combustion chamber. If we design an engine to achieve a peak gas velocity of mach 1 at this point then we can calculate the mass needed to create an adequate vacuum to provide proper breathing.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 62

The mass of the gas in the tailpipe is determined by two things: the volume of gas and its density (I told you there were many variables). At this stage it would be very easy to simply use a straight pipe, equal in diameter to the hole in the combustion chamber and long enough to hold sufficient mass of exhaust gas – and indeed this will work – but not very well. Not only would this tube be incredibly long (probably 30-40 times its diameter) but it would also be very inefficient, thanks to the friction of the gases against the sides of this very long tube and, more importantly for another reason. Before we get onto that other reason, let’s look at how we can increase the volume of the tailpipe without making it unreasonably long… The obvious way is to increase its diameter – since the volume of a cylinder is a function of its length and its diameter. However, we can’t simply connect a big tube to the little tube coming out of the combustion chamber as this would result in massive turbulence and huge inefficiencies due to that turbulence. The most logical way therefore, is to gradually transition from the small opening in the combustion chamber to a larger diameter that produces a greater volume – hence the coneshape so characteristic of a valveless engine’s tailpipe. But back to that second important reason for using a cone… When the air/fuel in the engine ignites, the gases enter the tailpipe at great speed – producing a pressure wave that travels the full length of that tailpipe before escaping in to the surrounding air. Pressure waves are just another name for sound waves and, as anyone who’s heard an echo knows, they tend to bounce around under certain conditions. What isn’t so obvious however, is that when a pressure wave travels down a tube and reaches the open end, some of its energy is reflected back down the tube – not as a pressure wave but as a rarefaction wave (a kind of negative pressure wave). This rarefaction wave is a narrow band of low pressure that can be used to some advantage in a pulsejet engine. When the tailpipe is the correct length, the rarefaction wave produced by the escaping pressure wave will assist in raising the strength of the vacuum formed in the combustion chamber during the intake phase. What’s more, the conical shape of the tailpipe actually serves to strengthen this rarefaction wave as it travels towards the combustion chamber.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 63

If the tailpipe is too long, this wave will take too long to be reflected back to the combustion chamber and its energy will be lost to the rapidly expanding pressure wave from the next combustion cycle. If the tailpipe is too short, this wave will arrive too soon and not interact optimally with the partial vacuum caused by the momentum of the exhaust gases. Once again, this will result in the energy being wasted rather than used to help draw in the next charge. The optimal time for the arrival of the rarefaction wave is dependent on a raft of different factors, including the volume of the combustion chamber, the flow-rate of the intake tube, the shape of the combustion chamber etc. So, with all this in mind, it’s obvious that the tailpipe length is important if we want to take advantage of the rarefaction wave. If we can calculate the optimum length for the tailpipe then it also becomes possible to calculate the angle of the cone – so as to create the ideal volume (aka mass) of gas in the tailpipe. It is very interesting to note that valveless engines with very long tailpipes (such as the original Ecrevisse) have a very shallow taper. Looking at other (more modern) engines shows that the shorter the tailpipe becomes for a given combustion chamber volume, the more acute the angle of taper (so as to hold a similar volume/mass of gas). The Lockwood engine is a good example of a medium-taper tailpipe and the Calgary one is an even more extreme example. I hope you can now see that while it’s not too difficult to design a valveless pulsejet that will run, it’s a whole lot more difficult to design one that will perform efficiently and which is optimum in all respects. For that reason, it’s probably a good idea to use a tried and tested design for your first attempts. It’s also worth reiterating that, in the world of pulsejets, bigger is better. Don’t be tempted to build a tiny little engine or you’ll be almost guaranteed to fail. A good size for a valveless engine is one with a combustion chamber of around four inches (100mm) in diameter. You’ll find the plans for a simple valveless engine of this size in the plans section of this disc or you can scale the Calgary design to suit. Flow Rectifiers The final element in designing a valveless pulsejet engine is making sure that the gases from the intake and tailpipe both exit the engine travelling in the same direction. In the case of an engine such as the Lockwood, this is achieved by simply placing a U-bend in the tailpipe. Unfortunately, the result is an engine which then has a really awkard and, for some applications, decidedly non-aerodynamic shape.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 64

An alternative to this is to place a 180 degree bend in the intake tube but this is also not without its problems – namely the fabrication of a U-bend with a constantly tapering diameter. Placing the u-bend in the intake also produces an engine that is significantly longer than the Lockwood configuration. Another alternative is something like the Chinese design or the Thermojet, where the intakes are attached to the combustion chamber so that they already point in the same direction as the tailpipe. Unfortunately (wouldn’t you know it), this also involves some compromises that tend to reduce the power and efficiency of such engines. One good idea is to use an external flow-rectifier which also doubles as an augmentor. Using this concept, you can kill two birds with one stone. This approach does still require the fabrication of a tapered u-bend and results in an even longer design than any of the other methods. Fuel Systems As with all pulsejet engines, the manner in which fuel is mixed with the incoming air charge is critical to performance. Ideally we want to create a very turbulent swirling mixture of air and fuel mixed at the ideal stoichiometric (combustible) ratio. It would also be extremely good if we could time the fuel flow so as not to waste any by injecting it when the engine is in its exhaust or combustion phases. The first criterion is fairly easy to achieve, especially considering that the gases inside the combustion chamber of an engine are going to be forced into turbulent flow patterns by the rapid inrush of the cold charge. It’s still a good idea however, to try and get the maximum mixing of air and fuel without simply relying on combustion chamber turbulence to do the job. Obtaining a good level of air/fuel mixing is pretty easy with a volatile fuel such as propane (see the method commonly used for the Lockwood design) but quite a bit harder if you plan on using a fuel such as diesel, kerosene or gasoline. These less volatile fuels usually require a good atomizing nozzle and a relatively high-pressure fuel pump in order to achieve the levels of mixing needed to provide efficient combustion. The second objective of timing the flow of fuel is a little more difficult and, unless obtaining the maximum fuel efficiency is important, generally not worth worrying about. It should also be remembered that one of a pulsejet’s few virtues is its simplicity. If you start adding complex timed injection systems you may gain some efficiency but at the cost of much increased complexity. Design Starting-points The best way to design your own valveless pulsejet is to start with the basic dimensions of a tried and true engine. This could be the SNECMA/Lockwood or the Calgary University’s shorter four-intake design and I’ve presented the basic dimensions for all these here. Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 65

Note that I’ve specified the dimensions in terms of ratios related to the diameter of the combustion chamber (D). These ratios remain relatively constant for most sizes of engine, except for very small engines (where D is less than 3 inches or 75mm) in which case Reynolds numbers become so low as to cause problems.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 66

Improving Pulsejet Performance Designing a pulsejet that produces more power and uses less fuel was the Holy Grail of engineers during the period from 1940-1960. However, as the turbojet, and later the turbofan, continued to provide increasing levels of power and performance, no serious effort was applied to improving the humble pulsejet. This doesn’t mean that the pulsejet is a lost cause though, and in recent years there has been something of a renaissance in the design and construction of these engines, driven partly by a global community of enthusiasts. It’s worth noting that as long ago as 1947, engineers working with pulsejets suggested that there were significant performance gains yet to be made with this technology. Fuel Consumption The single biggest problem with pulsejet engines is their horrendous fuel consumption. Most conventional pulsejets use somewhere between 3.0 and 5.5 pounds of fuel per hour for every pound of thrust generated. That means a that even a pulsejet operating at the more economical end of this scale and generating 100lbs of thrust would consume 300lbs of fuel per hour. That’s some 50 gallons of regular gasoline. By comparison, a modern turbofan engine uses as little as 0.34 pounds of fuel per hour for every pound of thrust produced. This means a 100lbs-thrust turbofan would use less than six gallons (just 34lbs) of fuel to accomplish the same task. I hope you can see now why there are no commercially made pulsejet-powered aircraft around. Not only would such a craft struggle to get into the air under the weight of its enormous fuel burden, but it would also be incredibly expensive to fly. So how can we go about reducing the fuel consumption of a pulsejet engine? The single largest problem is that pulsejets have a very low pressure ratio. In plain-speak, this simply means that it uses fuel very inefficiently because there’s a bloody great hole out the back that makes it hard for the engine to build up any significant amount of pressure inside. When fuel is ignited in your car’s engine it has already been compressed to about 1/10th its normal volume and 10 times atmospheric pressure. By compressing the air/fuel before it’s ignited, more energy is obtained from that combustion and greater pressures are generated inside the engine’s combustion chamber. A car engine uses a solid aluminum piston to compress the air/fuel mixture and confine it once combustion has occurred.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 67

Unfortunately, pulsejets don’t have nice solid pistons to compress and confine the air/fuel mixture prior to, or immediately after ignition. The only thing compressing and confining our air/fuel mixture is the light, relatively ineffective column of gas in the engine’s tailpipe. Clearly this column of hot gas is going to act like much of a piston. This is confirmed by measurements made inside a running pulsejet engine which indicate that the fresh air/fuel charge is only compressed by a tiny amount (around 2psi) before it is ignited. Other measurements indicate that the peak pressure generated inside the engine as the hot burning gases push against the gas column in the tailpipe is a very modest 12-15psi. Clearly design change that allows an engine to increase the operating pressures that are generated inside it will have a positive effect on both power output and fuel consumption. Some suggestions for obtaining higher operating pressures have included: 1. multi-point ignition to reduce the time taken to ignite the entire air/fuel load 2. the use of detonation waves to provide pre-ignition compression and much faster ignition of the total air/fuel load 3. increasing the level of turbulence inside the engine so as to promote faster ignition of the total air/fuel load All of these are valid suggestions but none of them are easily achieved without compromising some other aspect of the engine’s operation. Multi-point ignition The faster you can burn the entire air/fuel charge inside the engine, the more power will be produced and the more efficiency will be obtained. In a perfect engine we’d be trying to obtain what’s known as “constant volume” combustion where the air/fuel charge isn’t allowed to expand at all until it’s been entirely consumed. In such an engine, the internal pressure generated will often be more than ten times that which was present before ignition occurred. If we have a pulsejet that achieves an internal pre-ignition pressure of just 2psi The reason for this is simple – the only thing stopping the burning air/fuel from escaping out the rear of a pulsejet engine is the relatively light column of gas in the tailpipe. Being light, this gas actually does a very poor job of stopping the air/fuel from The most promising of these has been the use of a detonation wave to ignite the air/fuel load. It is this concept that has spawned a whole new arm of pulsejet research and the development of the pulse detonation engine (PDE). Unfortunately, PDEs have yet to make it out of the laboratory and into practical application anywhere. Although they do offer the promise of massive improvements in efficiency, they Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 68

are complex beasts that require a truckload of support gear in the form of pumps, valve actuators, oxidizer tanks and much more. Augmentors The simplest way to improve both the static/low-speed thrust and fuel-efficiency of a pulsejet is to add an augmentor to it.

Augmentors increase an engine’s thrust through two basic mechanisms: 1. Increasing the total mass-flow. Since the amount of thrust generated by an engine is equal to the amount of mass it expels times the amount of acceleration imparted to that mass, there are obvious gains to be had if we can increase the total mass of air that gets ejected out the back (providing we don’t slow down the speed at which it’s ejected. An augmentor draws cold air into the exhaust through the gap between the engine tailpipe and the augmentor intake – thus adding mass and increasing thrust. Because the hot exhaust gases heat this cold air, it expands and therefore helps maintain the velocity of the exhaust gases. 2. The Bernoulli effect. The front of an augmentor tube has a curved lip which serves two purposes. The first is to make it easier for cold air to enter the gap between the tailpipe and the augmentor cone, the second is to produce “lift” thanks to the Bernoulli effect. (see the chapter later in this book for an explanation of what the Bernoulli effect is). As the air passes around this curved lip, an area of low-pressure is formed and which means that atmospheric pressure on the back of that lip effectively applies a forward force on the augmentor – adding to the total thrust produced. When I added an augmentor to the intake tube of my standard Lockwood engine, the total thrust produced jumped from 57lbs up to around 80lbs – a very worthwhile improvement indeed.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 69

Accidents and Failures I’m often asked why I indulge in such dangerous activities as running large pulsejet engines and dodging balls of flame.

Well the answer is that, providing sensible safety precautions are taken, running pulsejet engines are no more dangerous many other activities. In my five years of intensive experimentation, I’ve only suffered one accident worthy of comment, and that was simply a case of tripping over a propane tank and knocking myself senseless on a concrete driveway. Not Explosions but Implosions “What happens if that thing explodes” is another question I frequently have to answer. The truth is that pulsejets are far more likely to implode than explode. The peak pressures encountered during combustion in a conventional pulsejet engine are less than 50psi and average out to just a few psi across the entire operating cycle.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 70

Because the metals most often used for pulsejet construction (steel, stainless steel, etc) are extremely strong under tension, they are very unlikely to burst, although I have had a few engines that split after many hours of hard running. Fortunately however, if a pulsejet engine develops a crack or a split it will usually stop instantly. This is because pressure variations necessary for proper operation are disrupted by the leakage that such a crack produces. A good example of this can be seen in one of the videos on this disc where the Lockwood engine on my jetkart stops abruptly because the sparkplug falls out. Even though the area of resulting hole is tiny when compared to the areas of the intake and exhaust tubes, the leakage that it produces immediately causes the engine to stop running completely. The real problem, especially with valveless engines, is that during the intake phase, the pressure inside the engine drops to very low levels. This effect of this low pressure is further exacerbated by the presence of a powerful rarefaction wave that travels back down the tailpipe towards the combustion chamber. When the pressure inside the engine is low, the material from which it is made is placed under a different kind of stress – namely compression. Whereas this metal is very strong under tension, it tends to be very weak when placed under compressive loads. This means that, if the proper construction techniques and materials aren’t used, there’s a very real risk of the tailpipe cone collapsing under the effects of atmospheric pressure, as in the picture on the right. The first sign of the failure was a gradual ovalling of the tailpipe cross-section – but when it failed it was spectacularly quick and happened in the blink of an eye. One second the engine was running nicely, the next it was looking like someone had run over it with a steamroller. Even the smaller Lockwood engine I used on the gokart had the same problem as seen in this picture In this case I’d made the tailpipe from 0.65 mm (0.026”) stainless steel and, although it did run fine for quite a while, it eventually succumbed to the effects of atmospheric pressure.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 71

Metal Fatigue Next to implosions, metal fatigue is likely to be the most common mode of engine failure. The metal from which a pulsejet is made takes an incredible amount of punishment when that engine is running. If you think about it – the combustion phase of a pulsejet’s operation accounts for less than 50% of the operating cycle. That means, in order to produce (say) 50lbs of thrust, the engine actually as to create a power-pulse of over 100lbs. When running, that 100lbs of force is transferred directly to the engine itself – and if it’s made from thin rolled metal, that places an awfully large amount of stress on certain areas. Not only is a pulsejet exposed to high levels of vibration and constantly changing tensile and compressive stresses but it also undergoes a huge and constant amount of thermal cycling. The end result of this is that it becomes very difficult to design and build large pulsejet engines that will hang together for any significant amount of time. In the case of designs such as the Lockwood, the whole problem of metal fatigue is further aggravated by the difficulties associated with mounting such an oddly shaped engine. This often means that parts of the engine are left largely unsupported and thus suffer from even greater levels of vibration/fatigue than might be encountered in other designs. The critical point for fatigue failure with a valveless pulsejet is the area between 1/3 and 2/3 the length of the tailpipe cone. If you place a weld in this area (which is often unavoidable) then you really need to keep a close eye on it. It is very likely that, no matter how carefully you make that weld, cracks will appear around it, especially if you’re using stainless steel. I’ve tried numerous methods for reducing the tendency to crack in this area with varying levels of success but have come to the conclusion that using thicker material than you might think necessary is about the only way to reduce the problem. I tried butt-welding, overlap welding and flanged welds – but they all eventually developed cracks when using a thickness of material that, in all other respects, appears more than adequate. Mild steel is far less prone to cracking than stainless steel – but it is also much weaker and therefore needs to be much heavier/thicker if you’re to avoid implosions. It also suffers from rapid oxidation when used in a pulsejet, which means your engine will turn to dust/rust in a fairly short space of time if plain steel is used. I have found however, that the use of high-temperature paint of the kind used on automotive exhaust systems (VHT is just one brandname) work reasonably well on the cooler sections of a mild-steel engine. This paint can reduce the tendency for the outside of the engine to rust but it will burn off the really hot sections.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 72

Beware of Thermal Expansion When I built my first Lockwood pulsejet engine I got badly caught out by the relatively high rate at which stainless steel expands when heated. This rate of expansion is known as the coefficient of thermal expansion (COTE). In simple terms, it’s a measure of how much a material grows when heated. As we all know, most materials (including metals) shrink when cooled and expand when heated. Most of the time we tend to ignore the effects of this heat-induced expansion because it is generally quite small. However, if you’re dealing with very long pieces of metal (like railway lines) or very high temperatures (such as found in pulsejet engines) then the amount of expansion can be quite extraordinary. If you’ve ever seen railway lines that have buckled due to the extreme heat of a summer’s day then you’ll understand just how much expansion can be produced by a combination of temperature and length. While most pulsejets are far shorter than a length of railway line, they are subjected to extremely high temperatures, and this can cause very significant expansion. To work out just how much a pulsejet of a given size is likely to expand when running we can use the following simple formula: Length increase = original length x temperature increase x COTE So, if we build a decent-sized pulsejet that is (say) 5 feet long and we assume that it will operate at a temperature of 850 degrees Celsius then we can fill in the figures as follows: Length increase = 60inches x (850-20) degC x 1.8x10-5 Note that we’ve subtracted the starting (ambient) temperature of the engine (20 degC) from the operating temperature (850 deg C) and that the COTE is 0.000018 which has been expressed in scientific notation. If we do the math, we find that the length will increase by almost an entire inch (0.8964 inches to be precise). Now imagine what would happen if you rigidly mounted the front and rear of this engine to a frame that remained cool and therefore didn’t expand… Yes, that’s right – your engine would be unable to expand so would crumple as it heated up – and that’s exactly what happened to my poor Lockwood. Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 73

So how do we get around this problem? One option is to mount the engine only at one point – but with a larger (and longer) engine, this becomes somewhat problematic, since it will likely introduce all sorts of other stresses due to the effects of gravity and vibration. A better way is to firmly anchor one end (the front in the case of a conventional pulsejet and the center of the U-bend in the case of a Lockwood) and allow the other end to slide freely on its mounting or have a sprung mount that allows sufficient movement to occur. The mounting I used on my Lockwood-powered gokart consists of a solid mounting tab welded to the inside center of the U-bend and a sliding mount that supports the other end of the engine at the intake tube and tailpipe. As you can see in this picture, the movement produced by thermal expansion is quite obvious and turns out to be almost exactly the figure I calculated for this length of engine running at 850 degrees C. Making it difficult is not the only problem created by the issue of thermal expansion. An even more critical effect is the way that such expansion causes fatigue and the eventual failure of bad welds in certain parts of the engine. Here’s a diagram of two weld beads. Note that weld A has quite a thick bead while weld B shows no difference in thickness. You might think that weld A is stronger and therefore better – but you’d be wrong. Once you start adding heat to these two welded join you’ll see that something horrible happens to weld A. Because the two pieces of metal being joined are relatively thin, they will heat up (and expand) very rapidly. The thick weld bead however will take longer to heat up (due to its thermal mass). This means that it will not expand at the same rate as the metal around it and that will produce a shearing stress to be created along the very edge of the weld bead. Now stainless steel is tough stuff, so it won’t break or crack immediately. However, stainless also suffers from a phenomenon known as “work hardening” which means that the more you stress it, the harder (and more brittle) it becomes. You can observe this behavior yourself by cutting a thin strip of stainless sheet (say ½ inch wide and four inches long). Start bending this strip back and forth with your hands. You’ll notice that very quickly, it stops bending in the middle and starts to bend away from that initial flexing point. You’ll also notice that the area where it was bending has become much stiffer and resistant to bending than the areas yet to be bent.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 74

That’s work-hardening in action. Once stainless becomes work-hardened, continued flexing will cause it to crack and break. Going back to our welds, after being heated and cooled a number of times, weld A will quickly develop a work-hardened area immediately adjacent to the weld bead. Weld B won’t be affected by this problem because all the metal in the area of the bead will expand and contract at the same rate – because it will be heated evenly. The effects of this problem will be dependent on several factors: 1. The amount of extra material in the weld bead 2. The change in temperature involved 3. The number of heating/cooling cycles experienced It’s worth noting that in a pulsejet engine, there will be two different mechanisms causing the temperature changes that create this weld-line stress. The first, and most obvious, is the massive temperature change that occurs when a cold engine is brought up to running temperature, and again when a running engine is turned off and cools to ambient. Although the absolute magnitude of the stresses created by these heat cycles is very large, they are relatively infrequent since most engines are run no more than a few dozens or hundred times in its life. A more insidious source of this heat cycling is found in the way a pulsejet operates. Some sections of the engine (most notably the tailpipe (or exhaust cone in a Lockwood) is subjected to quite significant temperature changes each time the engine completes an operating cycle. One moment the pipe is being heated by a blast of hot combustion gasses, the next it is being cooled by an inrush of cold air. Although the magnitude of these temperature changes is far lower than those produced by the startup/shutdown of an engine, they are also far more frequent (occurring tens or even hundreds of times per second). Thus, even a brand-new engine with bad welds will fail within a matter of minutes even if it’s only run once. With this in mind, it becomes apparent that good welding is essential if you want to build an engine that is long-lasting and reliable. If your welds are lumpy or suffer from a thick bead (as commonly occurs when using a MIG welder), then grind them down so that they’re flush with the surrounding material. That will significantly reduce the amount of stress produced by the heat cycling associated with pulsejet engine operation.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 75

Here are the COTEs for some common materials which might be used to build pulsejet engines (expressed in inches per degree Celsius). Austinetic (300 series) stainless steel: Ferritic (400 series) stainless steel: Martensitic (400 series) stainless steel: Mild steel (non-carbon):

1.8x10-5 1.1x10-5 1.2x10-5 1.5x10-5

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 76

A Simple Guide To Anodizing What is Anodizing? Anodizing a process that creates a very hard, protective, and sometimes decorative finish to aluminum. Aluminum is a wonderful material and is often used for the valve seat and intake section of pulsejet engines. Its excellent thermal conductivity and ability to cushion the impact of reed valves as they slam shut make it a good choice for these components – but its lack of hardness also means that it’s easily damaged. In this picture you can see the effect of hardened reed valves repeatedly slamming into a plain aluminum valve plate. You can see how the repeated impacts have damaged the surface of the valve seat and this can affect the ability of the valves to seal properly. By anodizing the valve plate, a thin protective layer of aluminum oxide can be created which will significantly reduce the amount of damage that occurs. This protective layer is extremely hard – almost as hard as diamond in fact – but it is so thin that it doesn’t adversely affect the valve seat’s ability to absorb the heavy impacts of the valves – thus you get the best of both worlds. Choosing an Aluminum Alloy Note that only some aluminum alloys can be anodized successfully. Some grades, such as the copper-alloys of the 2000 series (2024, etc) which are often considered “high-strength” are not suitable. The easiest aluminum to anodize are the “pure” alloys of the 1000 series (1050, 1100, etc) – but these are very weak and not suitable for such things as valve-plates where there are significant physical stresses involved. On the PJ8 and PJ15 engines, I use 6061 grade aluminum tempered to a T-5 level. This alloy is a considered a medium to strong structural alloy and anodizes very well indeed.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 77

How To Anodize Aluminum The anodizing process is deceptively simple and uses readily available chemicals. Providing you exercise reasonable care, it’s also a very safe process with no adverse environmental impact. Materials: To anodize aluminum you’ll need the following materials: 1. Fresh sulfuric acid. This can be obtained from your local automotive battery supplier and is very cheap. Don’t be tempted to use acid decanted from an old battery – it will be contaminated and produce very poor results. 2. Distilled water. If you don’t have a distiller, this can also be obtained from your local battery supplier. You’ll need to dilute the acid with 2 parts of water for each part of acid. 3. A plastic tub to hold the acid solution and the items to be anodized. I used a container designed for kitchen use – this had the added advantage that it came with a secure clip-on lid so I can store the acid solution in this container between anodizing sessions. 4. A lead or pure aluminum plate to act as the cathode. The purity of this plate is fairly important – if it’s alloyed with some other metal then there could be leaching into the acid solution with resultant contamination. Use 1050 or 1100 grade aluminum or lead from a known source (although some people have reported good results using lead flashing torn from old roofing iron). 5. Some aluminum wire or thin rod. A good source of this is aluminum welding wire as used in MIG welders. Alternatively you can cut a thin strip of aluminum from a sheet of the metal and use that. This will be used to hang the item in the solution and provide an electrical connection. Remember that this piece of metal will also be anodized and if there isn’t a firm, watertight connection to the workpiece, the oxide layer created may break the circuit and stop the current from reaching it. Remember that at the point where the wire connects to the workpiece, anodizing will not occur. Make sure that the connection point is not in a position where this lack of anodizing will cause an unsightly blemish if looks are important. You’ll also need a source of 12-24V DC electrical power. Since the anodizing process can draw several amps of current, you’ll need something more powerful than a simple wall-wart type of supply. I use a car battery (or two) that I recharge between sessions. To connect everything up you’ll want to use several clip-leads (lengths of wire with a crocodile clip on each end) and be sure there’s a fuse in the circuit close to the power supply or battery terminals. If the item being anodized accidentally touches the lead/aluminum plate then a short-circuit will occur and, without a fuse, this could cause a fire or even an explosion. Setting up Your Anodizing Tank Because anodizing uses an acid solution and creates large amounts of potentially explosive hydrogen gas you should set up your equipment in a safe, well-ventilated location where it’s not going to be knocked or in the vicinity of any type of naked flame.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 78

First, cut and bend your cathode plate (aluminum or lead sheet) so that it runs around the inside edge of your plastic container. To get an even and “all-over” layer of anodizing, you need to have your item surrounded by the cathode plate. Place the item to be anodized into the container and make sure that plenty of space (at least a couple of inches) remains between it and the cathode plate. This is to make sure that you don’t accidentally touch the two together and produce a dangerous shortcircuit. The amount of acid solution you need to mix up depends pretty much on the size of the object you’re planning to anodize and the volume of your plastic container. To avoid having a “bald-patch” on the bottom surface, you need to suspend the item to be anodized in the acid solution so that it’s not touching the bottom of the container so allow extra liquid for this. Don’t’ forget however, that when you lower your item into the solution, the level will rise. If you are planning to anodize a larger item, the level may rise significantly so factor this in so as to ensure you don’t overflow the container. Workpiece Preparation It can’t be emphasized just how important it is that the workpiece is scrupulously clean prior to anodizing. Any grease, oil or fingerprints will be magnified by the process and produce unsightly blemishes in the final finish. To this end, the following steps are useful in ensuring the cleanliness of the workpiece: 1. wash the workpiece in a weak solution of detergent in hot water 2. rinse thoroughly with hot water as any detergent residue will also cause problems. 3. Dip the workpiece into a warm (120 deg F) bath of 5% sodium hydroxide (lye) solution for about one minute. This will remove any remaining oil or grease and provide a very consistent surface. 4. Rinse thoroughly with hot water as in step 2 (above) I find that wearing a pair of cheap latex or nitride gloves when performing the above steps allows the workpiece to be held without leaving further fingerprints on the cleaned surface. Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 79

It’s also advisable in light of the fact that the sodium hydroxide solution is extremely corrosive to your skin. Performing The Anodizing Having set up your anodizing tank and prepared your workpiece, the item to be anodized is lowered into the acid so that it is completely immersed. It is then connected to the positive side of your battery or power supply. The aluminum cathode plate is connected to the negative side. Within a minute or so you should see a steady stream of bubbles rising from the workpiece – if not, you have a problem. The most common reason that the workpiece fails to bubble is because the electrical connection to it is inadequate. Once the anodizing process is underway you can go and read a book for 20-60 minutes, depending on the size of your item and the depth of anodizing required. You can briefly remove the workpiece from the tank from time to time to see how the process is going. Depending on the alloy involved, it will turn a dull gray or very light yellow color as the oxide layer forms on the surface. Using the anodizing setup described above, most workpieces will have a health layer of oxide on them within 60 minutes of processing – at which stage they can be removed and rinsed thoroughly in COLD water. It is important that you continue to avoid direct contact with the surface or contamination with any form of dirt, oil or grease. Dying the Workpiece One of the great things about anodizing is that it not only creates a tough, hard layer that protects the metal from wear but it also allows you to impart a color to it. The reason you can dye a piece of anodized aluminum is because the oxide layer formed during the process is quite porous. The millions of tiny oxide crystals act like a sponge and soak up any dye that has a fine particle structure. Of course the dying step is optional and if you’d like to simply retain the gray color of plain anodized aluminum you can move right on to the “Fixing” stage. The choice of dye is critical to successfully imparting a nice rich color to the anodized layer. Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 80

I’ve had great results from the Dylon range of cold water dyes. These are the little plastic pots with an aluminum lid that can be bought from most fabric stores. They’re dirt cheap and one little container produces enough dye to last a very long time. So far, I’ve tried the gold, red and blue colors with good results but others have reported that the greens also work well. To dye your workpiece you should remove it from the anodizing tank, rinse well in COLD water (it is important that the water is not warm or hot) and then imerse it in the dye solution. I mix one little pot of Dylon to almost a pint of water (500ml) and this concentration works fine. Leave it in the dye for about 10 minutes or so to allow the color to get deep into the crystaline structure of the oxide layer. Fixing the Anodizing This is a critical step that physically and chemically changes the structure of the oxide layer produced by the anodizing process. If you have dyed your workpiece, it effectively locks the dye in place so it won’t leach out. To perform the fixing you need to hold the anodized (and optionally dyed) workpiece over a source of steam for at least 10 minutes. I find that a pot of vigorously boiling water on the stove works fine. Rotate the work so that the steam comes into contact with all the anodized surfaces. It’s normal for some of the dye to leach out at this stage and color turn the boiling water – but most of it will stay in place. After 10 minutes of steaming, the workpiece can be immersed in the boiling water to finish the job. Leave it boiling for another 10 minutes. Once you’ve removed the item from the boiling water the anodizing is complete. Adding Luster You will notice that your newly anodized (and optionally dyed) piece of aluminum probably looks quite dull when it dries. This is normal. The anodizing itself produces a rough surface that imparts a matte finish rather than a shiny one.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 81

To make the surface shine and add luster all you have to do is either rub in some oil or polish of some kind. With some polishes you may get a small amount of the dye coming off on the polishing cloth but it’s nothing to worry about. The result should be a nice shiny workpiece with a deep luster. Hard Anodizing The anodizing process described here is often referred to as “decorative anodizing” because the thickness of the layer produced is quite low (about 0.0001 to 0.0005 inches) and therefore only provides a limited amount of protection to the underlying metal. Another process known as “hard anodizing” can produce layers of up to 0.005 inches which are extremely hard and protective. Performing hard anodizing is a little more complex than the above process however and involves the use of a chilled acid solution (not necessarily sulfuric) and much higher voltages/currents. While it is possible to perform hard anodizing in a home workshop, there are safety issues involved since the amount of explosive gas produced and the risk of electrical shock are far higher. For most purposes, the simple anodizing process described here will provide more than enough protection.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 82

Making Reed Valves Using Electro-chemical Etching The most common form of reed valve used on small pulsejet engines is the petal-valve. They are given this name because they look like the petals of a flower. Cutting such a shape from thin (usually 0.006”) spring steel can be very problematic. While scissors and shears work fine for cutting straight lines, it is almost impossible to cut the curves and thin slits needed to create a petal valve. Even if one were careful and lucky enough to produce a petal valve using such tools, the resulting valve would almost certainly have areas where the metal was bent or where there were overcuts that would encourage the rapid formation of cracks and premature failure. So just how can you create the relatively complex shape of a petal valve using commonly available equipment? The answer is to use electrochemical etching. By painting both sides of the reed valve material and then scratching the shape of the valve into that paint, it becomes possible to etch the exposed metal so that the valve virtually falls out of the sheet from which it has been made. Acid Etching At first glance it might seem that we could simply drop a suitably painted and scratched sheet of valve material into a container of acid and the exposed metal would be eaten away to do the job – but that’s not a particularly good idea for several reasons: 1. Acid is corrosive – and few of us have a pint or two of sulfuric, nitric or hydrochloric acid laying about the workshop. 2. Acid etching has the undesirable effect of rapidly under-cutting any exposed edge. This means that as well as eating directly into the exposed metal, the acid would start etching away under the paint. As a result, our finished reed valve would probably end up with very thin and ragged edges that would be prone to burning and splitting. Electrochemical Etching By comparison, the electrochemical etching method requires nothing more than a battery, some wire, a bowl and common household salt in solution with water. What’s more, electrochemical etching suffers far less from the under-cutting tendencies associated with acid etching. Here are the steps involved in etching a reed valve using this technique:

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 83

Preparation Let me say right now that preparation is everything. If you skip on this step then you will end up regretting it later. The metal from which your reed valve will be etched must be absolutely clean with no traces of rust or grease as these will cause the paint to lift and allow etching to occur in all the wrong places. It’s also important that the metal is slightly roughened so that the paint can adhere properly. A perfectly polished service will give the paint nothing to hold on to and it will come off in big flakes. For absolutely the best results you should scrub the reed valve material with a soapimpregnated wire-wool pad. This will remove all traces of grease and (if your arms are up to it) any rust spots. Now rinse in very hot water, taking care to hold the metal only by the edges – oily fingerprints will ruin your hard work. You can now give the metal an acid-etch if you have some dilute sulfuric acid available. This is done by dipping the bare metal into a very dilute solution (battery acid can be diluted by 4 parts of water). Place the metal in the solution and you should see small bubbles start forming. Lift it out at regular intervals and when it’s turned a dull gray color you can rinse it under hot running water again. This acid-etch will provide the best surface for paint to adhere to – but if you don’t have any sulfuric acid then don’t worry – you can give both sides of the metal a light sanding with 1200 grade wet-and-dry sandpaper. This will provide a similar surface roughness to help paint adhesion. Painting The type of paint and the manner in which it’s applied will also be a critical factor in the success of the etching operation. Don’t use a cheap spray-can enamel – it probably won’t stick well enough, even if you’ve roughened the surface. What’s needed is an automotive undercoat. These paints are designed to be sprayed directly onto bare metal and can be obtained in spray-can form from your local paint store or auto accessories outlet. If you can, use white undercoat – it makes the job of getting a nice even coat much easier than when using the traditional gray. Make sure you get an even and thorough coating of paint on the metal. I actually find it easier to lay the metal on a sheet of newspaper and Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 84

spray it while it’s flat. This avoids creating paint runs if you over-do it slightly. Of course it means that you have to wait for one side to dry before you can turn it over and do the other – but automotive primers tend to be very fast-drying anyway. Once the first coat is dry, give it another thorough all-over coat. Don’t be tempted to make do with a single coat. Even though it may look as if you’ve covered all the metal, experience has shown that there will probably be some very tiny pinholes that will cause similar holes to appear in your reed valve. A second coat is good insurance against this type of thing happening. Another reason you want at least two coats is because as you near the end of the etching process, it’s only the paint that will hold everything together. Insufficient paint means that the valve will start breaking away from the rest of the metal prematurely and this can cause the paint to rip away from a surface you don’t want etched – with disastrous results. Let the paint dry at least overnight if you can. Even though these paints are fast-drying, the very bottom layer tends to remain slightly plastic for several hours and this can cause the lines you scribe later to close-over. Marking Out Now that you have a nice piece of reed valve material, totally covered in a good solid layer of paint, you need to scribe the lines that represent the outline of your reed valve. NOTE: You only scribe one side! The undisturbed layer of paint on the back-side of the plate will hold everything together as the metal under the scribed lines is etched away. You can draw and scribe the pattern of your reed valve directly onto the piece of metal you’ve prepared – but it’s a better idea to make a template that you can trace around. The reasons for this are obvious: If you make a mistake while drawing the pattern onto your painted material then your work to date will have been wasted. I’ve made a template from 1mm (0.040”) stainless steel and I simply press this against the prepared reed valve material and scribe around the edges with a sharp modeling knife. Don’t forget to also scribe the hole in the middle! If you have an existing reed valve in good condition then you can use that as the template for scribing your pattern.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 85

When you’ve finished scribing, the shiny steel underneath the paint should be visible at the bottom of the scribe lines. Check to make sure that all your lines join where they should – a line that doesn’t quite join up will leave a bridge of metal that will make complicate removal of the valve from the sheet of prepared metal. Etching You’ll need a plastic or glass bowl or container that is large enough to fully submerse your valve material while it’s stood on edge. As you can see here, I’ve used an old yogurt container but you can grab your mother or wife’s Tupperware if she’s not looking and it would do just as well. Now mix up a solution of common table salt and water. About a tablespoon per pint will do the trick – the strength of the solution isn’t that critical. You’ll need enough to fill your container to the desired level. Hint: the salt will dissolve more easily if you use warm water. Now find yourself a piece of stainless (preferred) or regular steel that will act as a cathode plate in the solution. It should be about the same area as your blank sheet of reed valve material – although, once again, this isn’t too critical. Next, you’ll need a source of 6-12V DC. This can be a lead acid car or motorcycle battery or, if you have one, a variable voltage/current power supply. Connect that plate to the NEGATIVE terminal of your battery or power supply and place it on once side of your container, immersed in the salt solution. Place your painted and scribed piece of reed valve material in the salt solution on the other side of your container – making sure that the scribed side faces the cathode plate. Make absolutely sure that the two pieces of metal can not accidentally touch together if they move. One good way to do this is to place a sponge in the middle. This will absorb the salt solution and allow the current to flow but stops the two plates from meeting. Connect the reed valve material to the POSITIVE terminal of your power supply. You may want to include a resistor (8 ohms is about optimum) in this lead to limit the current flow if Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 86

you’re using a small container as I have here. If you don’t have an 10 watt, 8 ohm resistor, you can use a 10W-20W 12V light bulb instead. Once you connect things up, you should see bubbles begin to rise from the cathode plate as in the picture. At this stage the salt solution will still be clear. Depending on a number of factors, it may take between 10 minutes and an hour to etch your valve. Once the process gets underway, a rather awful looking green or brown sludge will begin to form on top of the solution. This is the iron that has been removed from the scribed lines. Things will go more quickly and the results will be better if you give the reed valve plate a bit of a shake now and then. This dislodges the crud that forms on the scribed lines so that a fresh salt solution can reach the bare metal and continue etching. If you remove the plate from the solution you’ll see that the formerly shiny metal under the scribed lines has turned black. Eventually the scribed lines will etch right through and when you remove the plate from the solution you’ll see the paint on the back surface exposed. If you hold the plate up to a lamp at this stage you can see exactly where the etching is complete because the light will shine right through as in this picture. Post-etching Steps Once you get to this stage you can disconnect all the wires and carefully push the reed valve out of the plate. There will probably still be some areas where the etching is not quite complete but the metal at these points will now be so thin that it will break away very easily. Don’t worry if the edges seem a little ragged – this is normal. Now wash off the paint with suitable thinners. Another benefit of the automotive primer is that it washes off very easily with lacquer thinners.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 87

Now you’ll have a new valve – but chances are that the edges will still be ragged as mentioned above. In order to avoid premature cracking of the valve it pays to file or sand those ragged edges to make them smooth(er). In theory, you can further reduce the risk of cracking by putting the valve in your oven at about 200 deg F for an hour or so. What this will do is bake out any hydrogen that might have entered the structure of the reed valve material as a result of the etching. Hydrogen in the molecular structure causes what is known as “hydrogen embrittlement” and that makes steel far more prone to cracking. For what it’s worth – I don’t bother. Let’s face it – the valves are going to get hot enough anyway once you fire up your engine.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 88

Newton’s Third Law Good old Sir Isaac Newton was a smart guy and his observations formed the basis of what we’ve come to know as Newtonian physics. One of his cornerstone laws of physics (Newton’s Third Law) says that: for every action there is an equal and opposite reaction It is this law that explains why a pulsejet (or any jet engine for that matter) creates thrust. In short, our pulsejet only provides a forward push because it is also pushing hot gas out the tailpipe with equal force. You can experience the effect of this law by sitting on one of those swivel-type office chars and rapidly twisting your body to the left. Notice how the seat of the char twists to the right? That’s Newton’s third law in action. The force you used to twist your upper body one way created an equal and opposite force that twisted the seat (and your bottom) the other. You can also try sitting on the chair with a heavy weight in your hands. Throw that weight across the room and you’ll find that you and the chair move off in the opposite direction, courtesy of Mr Newton. But what goes on inside a pulsejet and how does this equal and opposite force cause our jet engine to move forwards?

To see just how this works, look at the diagram above which represents two containers filled with pressurized gas. The first box is sealed so the pressure inside is perfectly balanced and there’s an equal force pushing on all interior surfaces. The result of this is that everything is balanced and no thrust is created. If, as in the second diagram, we suddenly remove one side of the box then suddenly the pressure has nothing to push against. Consider this to be the open tailpipe of a pulsejet. Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 89

However, there is still pressure acting on the other interior surfaces of the box but now we have an imbalance. The pressure pushing the box upwards is still balanced by the pressure pushing it downwards but the pressure pushing the box to the left is no longer balanced by the pressure pushing it to the right. As a result, the box in the second diagram will be pushed to the left by a force equal (but opposite in direction) to the force with which the compressed gas blows out through the large hole. We can actually calculate the thrust that will be produced by summing the force vectors. If we assume that the box is a one-foot cube and that the pressure inside is 10psi then we can see that if one side was suddenly removed there’d be 1 square foot (144 square inches) of unbalanced force pushing on the left hand side of the box. That would produce a thrust equal to 1,440 lbs. Of course this simple calculation assumes that the pressure inside the box will remain at 10psi even when the side is removed. If, instead of knocking the whole right-hand side off the box, we simply cut a 1 square inch hole then just 10lbs of thrust would be generated. The other 1,430 lbs would be pressing against the part of the side that was still intact and balancing the force pushing on the left-hand side. This simple calculation validates Tharrat’s 2.2lbs of thrust per sq in of cross-sectional area constant we mentioned in an earlier chapter. Tests conducted on the Argus V1 engine showed an average pressure at the end of the tailpipe of around 2.2-2.6 lbs per square inch. This figure represents the average value of a pressure wave that fluctuated from a maximum peak of +6.2 psi achieved during the combustion phase, to a minimum of –4.46 psi produced during the intake phase.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 90

The Reynolds Effect Way back in the 1880s, a gentleman by the name of Osborne Reynolds spent some time analyzing the way in which gas flows. Amongst other things, he discovered that “size matters” – well at least when it comes to pushing gases down pipes or around obstacles it does. Although I could again bore you with pages of formulas, I won’t. I’ll condense his lifetime’s work into a simple observation: As you make things smaller, the air appears thicker (more viscous). What this means is that it’s disproportionately harder to suck or blow air through a thin straw than it is through a thick one – and that’s quite important in the world of pulsejets. To all intents and purposes, the air that passes through a small engine like a Dynajet behaves a whole lot differently to the air that passes through a big pulsejet such as the Argus V1 engine. This also explains why it’s not possible to simply scale down a large pulsejet engine and still expect it to run properly, if at all. To draw an analogy – at Dynajet sizes, air behaves more like maple syrup. It flows but is reluctant to squeeze through holes and tends to stick to the sides of any pipe you pour it through. At Argus V1 sizes, air behaves more like water – flowing quickly with far less tendency to stick to the sides of a pipe. This explains why the Dynajet needs to have a valved area which is a full 50 percent of its tailpipe area whereas larger engines often perform very well with as little as 20 percent. It also explains why small pulsejets generally require a larger L/D ratio to run properly. In order to overcome the effective viscosity of the air at these smaller sizes, you need a greater (longer) mass of air in the tailpipe to create sufficient Kadenacy effect for proper breathing. Mr Reynolds has a lot to answer for!

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 91

The Bernoulli Effect This might sound a bit crazy but a smart … by the name of Bernoulli discovered that the faster a gas moves, the lower its pressure becomes. Here’s a simple demonstration to show this effect. Take a glass of water and stand a regular drinking straw in it. Note that the water inside the straw is at the same level as the water in the glass. This means that the air-pressure inside the straw is the same as the air-pressure outside. Now blow briskly across the top of the straw. The water will rise up inside the straw and, if you’ve got really good lungs, it might even reach the top and make a mess all over the table as exits as a fine spray of droplets. So why did this happen? Well, as Mr Bernoulli predicted, by increasing the speed of the air directly above the straw, its pressure was reduced. Since the pressure at the top (and inside) the straw was then less than the normal atmospheric pressure acting on the water outside the glass, the water rose up. So what use is this effect in the wonderful world of pulsejets? Well an understanding of this effect allows us to control the pressure inside an engine by controlling the speed at which gas passes through it. If want more pressure is desired, we just slow down the gases and if we want less pressure we speed them up. So why do we want to change the pressures inside a pulsejet? Well the atomizer that creates a fine stream of fuel droplets is one example of why this effect is useful You’ll remember that in our drinking straw experiment, we were able to convert the liquid water in the glass into a spray of droplets simply by blowing across the top and creating an area of low pressure. To atomize the fuel in our pulsejet we need only do the same thing. Engines such as the Dynajet use just such a system for atomizing their fuel. By forcing the incoming air through a narrow opening called a venturi, an area of low-pressure is created that

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 92

sucks the fuel through a ring of small holes. When the fuel hits this fast-moving airflow it breaks up into droplets. Without the Bernoulli effect a Dynajet wouldn’t run because it couldn’t suck any fuel into the engine. There’s another area where some believe that the Bernoulli effect plays an important role in pulsejet operation. Take a look at the shape of the average petal-valved engine again. Note that it has a larger diameter section at the front and a smaller diameter tailpipe. These two sections are joined by a cone. Now imagine what happens when the hot exhaust gases are drawn back into the front section by the partial vacuum left after combustion. While in the tailpipe, those gases will be travelling relatively quickly – so they’ll have a low pressure. As they pass through the cone, into the larger diameter front section of the engine they will slow down and in doing so, their pressure will (thanks to Mr Bernoulli) increase. One of the things we want for efficient combustion is as much compression of the air/fuel charge as possible. Personally I don’t believe that this particular effect has any real bearing on a pulsejet’s operation – an assertion supported by research that shows pulsejets with a completely straight pipe are capable of producing just as much thrust as those which have an enlarged front section. Others have pointed out that the Argus engine has a diffuser section immediately after the valve-grid. The claim is that this diffuser is a device that acts like a divergent cone and therefore increases the pressure of the incoming air/fuel mixture by slowing it down. Perhaps it does have this effect, but my own experiments with intake diffusers indicates that their main benefit is that of forcing the air/fuel mixture through a narrow section so that mixing is more thorough and, as a result, combustion is more rapid. I also believe that the Argus diffuser also reduces the valve grid’s exposure to hot combustion gases by effectively choking their flow towards the front of the engine. One only has to look at how the Dynajet (which has absolutely no internal diffuser) produces almost an identical power/volume output as the Argus (with its internal diffuser) to realise that the effect of such a device on an engine’s power output is minimal at best.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 93

The Coanda Effect Some time during the 1930s, a Romanian by the name of Henri-Marie Coanda observed something very interesting. He spotted that when air (or any other gas) was flowing along a surface that curved away from the flow, the gas didn’t just carry on straight ahead but followed the curvature of the surface. Or, to put it another way, a stream of fluid or gas will tend to hug a convex contour when directed at a tangent to that surface. You can see the effect illustrated in this diagram. Instead of blowing straight past the circular surface, the air will stick to its surface and bend through 90 degrees. You can check this out for yourself by turning on a tap, so that there's a steady but gentle continuous stream of water flowing. Now bring the back of a spoon into slight contact with the stream and you'll find that the water will no longer fall straight down but actually stick to the curve of the spoon. My own experiments indicate that a single curved surface will deflect a flow of roomtemperature air by a maximum of about 90 degrees -- after which the flow detaches itself from the curve and once again travels at a tangent. Although the Coanda effect doesn’t play a large roll in traditional pulsejet engines, an understanding that it exists and how it affects the flow of gas is something that all budding pulsejet designers should have.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 94

Plans This chapter of the book includes a number of plans for various engines, including the 55lbs –thrust enhanced Lockwood Valveless engine that currently resides on my kart and which featured in the popular TV series Scrapheap Challenge. Although based on similar designs, this particular version has a flared intake cone and slightly different combustion chamber dimensions both of which help to assist in throttling and easy starting. This engine should be built from material that is at least 0.7mm in thickness and, if weight isn’t too much of an issue, you might find it worthwhile going up to 1.2 or even 1.5mm material. Although thicker material will cost and weigh more, and is harder to form, the result will be a more durable engine that is less likely to suffer from cracking.

This plan is included as a separate PDF file (55lbslh.pdf) and contains all the dimensions you need to build your own. One detail often missing in other Lockwood plans you might find on the internet is how the fuel-injection system works.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 95

I’ve experimented with two methods that work very well. Both methods involve passing a tube cross-ways through the engine. This tube is cross-drilled with a number of small (1.5-2mm) holes through which fuel sprays into the engine. In the first method, the tube is placed in the intake tube, as close as possible to the cone that joins that tube to the combustion chamber. If you’re using a tube here it can be made from plain steel rather than stainless because it gets plenty of cooling from the incoming airflow. The second method involves running a much longer tube through the engine so that it passes through the cone connecting the intake tube to the combustion chamber. This tube is positioned as close as practical to the combustion chamber end of the cone and will be quite a bit longer than the one required for the first method. In both cases, the tubes should be around 8mm in diameter and the crossdrilled holes should point towards the sides of the engine, not to the front or rear. The fuel-tube, once installed, is fed with liquid propane which sprays out through the row of small cross-drilled holes and vaporizes into propane gas. If you want to run this engine on a liquid fuel such as gasoline, methanol, jet A1 or diesel then you can install a second fuel tube running at right angles with the propane tube – so that the two tubes form an overlapping cross when viewed down the intake tube. Because these liquid fuels are not as volatile as propane, you will need to start the engine using propane and then start the pump that delivers liquid fuel to the second fuel tube. The propane can then be turned off and the engine will run entirely on liquid fuel. Amongst the other files on this disk you’ll find many plans of valved and valveless pulsejet engines, most of which are readily available on the Net from websites such as www.pulsejets.com and others. I’ve included them on this disk only for your convenience. Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 96

Building a no-weld pulsejet If you don’t have a lathe, welding equipment or other specialist metalworking equipment then you’ll be pleased to know that this disk also contains a full-length video that shows how you can build your own pulsejet using nothing more than regular hand-tools. There’s also another file which accompanies this video and that provides additional material in text and pictures.

This simple engine won’t produce a whole lot of thrust, but it’s a great way to get your hands on a real working engine for probably less than $20 in materials.

Other plans will be posted to my website as they become available. You should check the subscriber section regularly for updates using the ID and password that will have been supplied to you when you purchased this disk.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 97

Wacky Ideas I regularly get email from people who want to know if their idea will work when applied to pulsejets. Here are some of the most common: 1. Water Injection Full-sized turbojets can gain significant amounts of power by the use of water-injection. This works in two ways. If water is injected into the intake of the engine, its evaporation cools the incoming air, this increasing its density. Since the power of almost all engines depends on the difference between the minimum and maximum temperatures involved, cooling the incoming air will raise power. This method of water-injection was actually used in some WW2 aircraft piston-engines where a mixture of water and methanol was used. The second way to use water injection is to squirt water into the combustion-chamber where the extremely high temperatures will cause that water to flash into steam – increasing the pressure in the chamber. One modern jet aircraft that uses this technique is the Hawker Harrier AV8 “jump-jet”. The use of water-injection is essential when this plane needs to perform a vertical take-off or landing while heavily laden with weapons or other external stores. Unfortunately, since water itself is very heavy, these planes only have sufficient water on board to allow a minute or two of vertical flight. But back to pulsejets. In theory, water injection should provide some additional power when used in a pulsejet but I have yet to try this in practice. A big problem however, is the additional weight and complexity that the water, plumbing and pump would add to an engine that has simplicity as one of its few virtues. In reality, it might be more practical to use a pulsejet as a flash-steam boiler by wrapping it in copper pipe through which water is pumped. That water would be flashed into steam by the intense heat of combustion then the steam could drive a turbine or piston. 2. Adding a turbine to the pulsejet Although this is an idea that was first contemplated many years ago, my own experiments indicate that it doesn’t work very well at all. The problem is that pulsejets suck and blow through the same pipe. That means our turbine will be subjected to alternate forces that first try to rotate it in one direction then in another. Since efficient turbines tend to be rather delicate devices, pounding them hundreds of times per minute with pulses of hot gas is not likely to create a very reliable machine.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 98

When I placed a small fan in the exhaust of a pulsejet I was surprised at just how little inclination it had to spin – even though the engine was putting out good power. Bringing it closer to the tailpipe actually reduced the spinning as the effect of the “suck” phase became more pronounced. 3. Running a pulsejet on hydrogen They keep telling us that hydrogen is the fuel of the future, mainly because when it burns all you get is heat and water (usually in the form of steam). While that’s true, hydrogen does not make a good fuel for pulsejets for several reasons: Firstly, hydrogen is an incredible gas insomuch as it has an extraordinarily wide stoichiometric range. In simple terms, that means it will burn equally well when there’s just a little oxygen as when there’s a whole lot of oxygen. If we compare hydrogen to regular gasoline we see that in order for gasoline to burn it must represent no less than 1% and no more than 7.8% of an air/fuel mixture. Any less and there’s not enough fuel to form a flammable vapor, any more and there’s not enough oxygen to support combustion. Hydrogen however, will burn happily when it constitutes anywhere from 4% to 75% of an air/fuel mixture. This produces an undesirable effect when used in a conventional pulsejet engine. As you’ve probably already figured out – a pulsejet does not burn its fuel continuously but in short bursts. The fuel won’t ignite unless just the right amount of air is mixed with it. During the operating cycle of a conventional pulsejet there are times when we have air, fuel and flames in close proximity to each other – but the air/fuel mixture seems to ignite at just the right time. One of the reasons for this is that the comparatively narrow stoichiometric range of most hydrocarbon fuels suppresses premature ignition – the fuel won’t ignite until there’s just enough air in the engine. In the USA, NASA attempted to run a popular model airplane pulsejet (the Dynajet) using hydrogen as a fuel and they failed to get it to pulse. The hydrogen seemed to burn continuously because of its wide stoichiometric range. There are numerous other problems that preclude hydrogen from being a viable pulsejet (or any engine) fuel for now. It is extremely difficult to store hydrogen safely. In order to store enough of it to be useful it has to be either compressed to incredibly high pressures, or chemically bound in a matrix of highly reactive materials such as metal hydrides. In the former case there are obvious risks should a pressure vessel fail and the weight of such containers detracts from the overall efficiency of a hydrogen-powered engine package. Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 99

In the latter case, the chemical storage of hydrogen involves the use of very expensive and dangerous chemicals that react violently with common substances such as water. 4. Fitting an afterburner Sorry, conventional afterburners won’t work on pulsejets for several reasons.

Firstly, they are designed for turbojets where the gas-flow from the engine’s exhaust is a constant flow. Unfortunately pulsejets create a flow that actually reverses briefly during the intake phase of the engine’s operation. However, always keen to try the impossible, I added an afterburner to my jetkart. Check out the next chapter on pulsejet afterburners to see how that turned out. 5. Building a really small pulsejet It’s sad but true that the smaller you make a pulsejet, the more difficult it will be to start and the less power it will produce. And don’t think that by halving the size of a pulsejet you halve the power because that’s not so. Halving the size of a pulsejet will actually reduce the power by a factor of four. Yes, that’s right, all else being equal, half the size means just a quarter the power. Another factor that adversely affects small pulsejets is their very low Reynolds numbers (see the chapter on the Reynolds Effect). As pulsejets become smaller, the air appears to become thicker – until it becomes just too thick to be sucked and pushed through the engine given the small amount of power available to do so. The smallest practical pulsejet I’ve been able to get running on regular fuels was about 14 inches in length and ¾ inches in diameter.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 100

6. Pulsejet-powered helicopters I shudder everytime I hear of someone who is thinking of building a helicopter powered by a pulsejet. Trust me when I say that this is not a good idea if you value your life. There have only been a very few such craft ever built and, without exception, they were never flown for any length of time or made into commercially viable products. Valved pulsejets are simply not reliable enough to bet your life on and valveless engines are difficult to design in such a way that they are compact and streamlined enough to fit to a helicopter’s rotor-tips. Lockwood-Hiller did experiment with special valveless engines designed to be built into the blades of a helicopter but this project was dropped – ask yourself why.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 101

Afterburning Augmentors

In theory, you can’t fit a working afterburner to a pulsejet for several reasons: 1. An afterburner is really just a specialized ramjet engine, and ramjets only work when the air flowing into them exceeds a speed of around 350mph (560kph). Although the gasses ejected from the tailpipe of a pulsejet regularly exceed this velocity by a factor of two or more, they also slow to zero and actually reverse direction, being sucked back into the tailpipe, once per engine-cycle. This would mean that most of the time, a pulsejet afterburner would be doing nothing but getting in the way. 2. An afterburner is an air-breathing engine and requires that the gases that enter its intake contain enough oxygen to support the combustion of the fuel injected into it. In the case of a gas-turbine engine, this oxygen is provided by the large percentage of air which passes straight through the engine without being involved in the combustion process. In the case of a pulsejet however, there is absolutely no oxygen left in the exhaust gases so there’s nothing to support the combustion of additional fuel injected into an afterburner So it would appear that there’s simply no way to add an afterburner to a pulsejet, right? It could also be argued that there’s no point in adding a fuel-guzzling afterburner to a pulsejet because an augmentor will provide the same effect (a thrust increase) far more easily and with the added bonus of not increasing fuel consumption But what if we injected fuel into the augmentor and tried to produce a device that performed both roles? Since an augmentor works by drawing in fresh, cold, oxygen-rich air and mixing it with the hot exhaust gases, it should be possible to inject fuel and burn it in the augmentor cone itself. Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 102

To test this out, I added an afterburning augmentor to the 55-lbs-thrust Lockwood valveless pulsejet on my gokart. As you’ll see by the videos on this disk, this afterburner certainly produces some impressive visual effects, creating a long tail of yellow flame out the back of the engine – but does it increase the thrust at all? Well the first clue is in the color of that flame. A yellow flame indicates poor combustion, generally resulting from inadequate oxygen being available for complete combustion. An obvious solution would be to reduce the amount of fuel being injected so as to lean-out the mixture. Unfortunately, due to the very inconsistent airflow through the augmentor, it is not possible to reduce the flow much before the fire goes out completely. I tried a range of different flame-holders (devices designed to protect the flame from being blown out) but found that the wild variations in the speed of the gases through the augmentor made it impossible to get a consistent lean-burn. Undeterred, I measured the thrust being developed as follows: plain (unaugmented) Lockwood engine augmented engine augmented with afterburner

25.0Kg 35.5Kg 36.0Kg

57lbs 80lbs 81lbs

Clearly in this case, adding fuel to the augmentor makes no practical difference in terms of thrust generated. Although I could claim these results prove my claim that you can’t add an afterburner to a pulsejet, it must be acknowledged that the design of a good augmentor is far from the design of a good ramjet. Experimentation with the shape and design of an afterburning augmentor may result in some improvement, and possible a small amount of thrust Increase. However, wouldn’t it just be a whole lot easier to build a slightly larger engine in the first place?

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 103

A Little History and a Few Important People The pulsejet has a rather erratic history, having been worked on by a number of very bright individuals, each of which has contributed a little more knowledge and innovation. One of the key players in the evolution of the pulsejet is a Dutchman by the name of Francois Henri Reynst who was born in 1909 and whose collective works were published by Pergamon Press in 1961 under the title Pulsating Combustion. Like all good researchers, Reynst drew heavily on the work done by others and the book detailing his work and findings is a must-have for anyone who is seriously interested in the theory of pulsejets. If you can get a copy of the book titled “Pulsating Combustion, the collected works of FH Reynst” then you’ll find it a fascinating read and well worth a place on your bookshelf. Reynst’s primary focus and interest was his valveless combustion pot which, unlike virtually every other valveless pulsejet, has only one orifice which acts as both an intake and exhaust tube. This pulsejet was designed primarily as an efficient combustor for heating purposes and not as a thrust-generator. He explored many interesting ideas, such as using the pulsejet as a gas-source for driving a turbine and the ganging of two pulsejets together in to produce a more steady exhaust flow. The fact that this book, despite its age, contains so many gems of information that are still totally relevant today speaks volumes for the dearth of research that has been done into pulsejet technology since the end of WW2.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 104

Tharrat Another of the leading lights in pulsejet design is Tharrat and his short but extremely enlightening paper “The Propulsive Duct”. In just a few short pages, Tharratt concisely lays out the basic principles of pulsejet operation and offers some simple formulas for calculating elementary “propulsive ducts” of almost any size. Kentfield Through is work at the University of Calgary, Kentfield has offered some interesting insights in to variations on the basic valveless engine design. His four-intake abreviated Lockwood derivative claims superior fuel efficiency and a more compact form factor than the engine that sprung from the Hiller workshops in the 1960s. I know of several versions of this design that have been built and appear to work as intended. Schmidt I’ve already dedicated an entire chapter of this book to the work performed by Schmidt since it really was pivotal in the single most obvious use of the propulsive pulsejet engine – the V1, but this chapter would not be complete without a mention of his work. Lockwood To be honest, Ray Lockwood didn’t really come up with much in the way of new ideas during his time at Hiller in the 1960s. However, he did put much existing theory to good use in the design and refinement of his well-known horse-shoe-shaped valveless engine. His designs are perhaps the easiest to build an operate of all the valveless engine designs and for that reason they deserve recognition. Lockwood also put quite a bit of effort into the design and analysis of augmentors, which became a key element of his “lift engines”. The lift engines were heavily augmented and designed primarily to create high levels of static thrust suitable for vertical take-off vehicles and such things as lifting military equipment into a hover so that it could be easily moved. Unfortunately none of this turned out to be practical.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 105

The future of pulsejet technology Although it enjoyed most popularity in the mid 20th century, the pulsejet is still far from dead. Pulsejet thrust engines are still used occasionally on target drones where their poor fuelefficiency and short life is far less important than their low cost. The future of pulsejet engine technology is almost certainly going to be in the pulse detonation engine which, researchers hope, will provide a useful way of propelling manned and unmanned craft to speeds well in excess of Mach 3 and with vastly greater efficiencies than existing engine technologies offer. As for the plain old pulsejet, it’s still around in various forms but you’re more likely to see one operating as an insecticide fogger than affixed to an airframe (model or full-sized). It would appear that with only a few exceptions (myself included), research and development of the humble pulsejet has slowed to a crawl as propulsion engineers focus on bigger, brighter opportunities. I doubt however, that the pulsejet as we know, will ever disappear since there will always be enthusiasts who enjoy building and running one of the world’s simplest jet engines.

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet

Page 106

Updates to this book and CD It is my plan to publish regular updates and additions to this book and CD which all purchasers will be entitled to download for free. You should have received a login ID and password either with this disk or by email. Using these, you can log onto the webpage at: http://aardvark.co.nz/pjet/subscriber and access this information as it becomes available. This part of the website will also carry any corrections that may be necessary. If anyone has any specific problems or questions, they’re always welcome to contact me using the form at http://aardvark.co.nz/contact

Copyright 2004 Bruce Simpson. http://aardvark.co.nz/pjet