A recent request request prompted me to write an article on hydroforming exhausts. "But that isn't a part of the frame" you say. That isn't always true - Eric Cheney once made a frame where the exhaust pipe formed the front downtube and top tube of the frame.

Have you ever wondered how those smoothly contoured expansion chambers or constantly tapering four-stroke headers are made? You've probably figured that the big manufacturers put equally big dollars into stamping dies, but how do the small specialty exhaust makers do it at an affordable price? You could make a die and hammer the sheetmetal into it (I've been told this is how the swoopy Magni MV Agusta pipes are done). Hydroforming is another way, and if you have access to a welding set and a water hose (and a couple of hundred dollars) you can do it too.

"Performance Bikes" magazine had an article a few years ago on an English expansion chamber maker (I think his name was Derek Huxley - Peter Gibson is known for hydroformed four-stroke exhausts). I read the article and decided I'd try my hand at hydroforming a curved megaphone for my 650 Alazzurra.

The technique is this: you figure out the shape the exhaust would take if you could smash it flat. A cylindrical section pipe would have, when flattened, a rectangular shape with the long axis of the rectangle the length of the cylindrical section, and the short axis being 1/2 the circumference of the cylinder. You can also envision the shape by thinking what it would look like if you cut the pipe in half down the long axis, and then flattened the two sides out. To build the pipe you reverse the procedure, cutting out two halves of the pipe, welding them together at the edges and blowing them back out into a rounded cross section.

One word of warning - it isn't immediately obvious what radius to use on curves, as the pipe seems to end up with just a bit more curve after it is expanded. We think that drawing the inner and outer curves with a common center point may help, (or at least reduce any tendency to buckle in the curves), but you may still need to allow a small percentage of extra bend into your pattern. I think after you've done 10 or so test pipes this will all come clear - I've only done a couple, and it is still a bit murky! But then, if it were easy everyone would be doing it.

I made the megaphone for the Alazzurra out of .035" mild steel sheet. A megaphone doesn't have much stiffness inherent in its shape the way an expansion chamber does with its complicated shape. If you are concerned about dents in the megaphone go up a thickness on the steel. The .035" sheet should be fine on an expansion chamber. The thicker steel is not too hard to form into a simple shape for a megaphone, but might be a problem in an expansion chamber.

One problem you'll immediately stumble over is that you can only make the blown pipe in a single plane. This may not be a problem with a megaphone as I made, but a single plane expansion chamber isn't likely to fit on your bike! What you do is take all the differently angled sections of the pipe as it would be in final form and then rotate the different sections into the same plane to make your pattern. After expanding the pipe you cut the sections apart, rotate them as needed, and weld them back together again.

Some commercial pipes look like they put the two flat sections of the pattern together and then welded the adjacent edges together. While this is expedient, I don't think it is the best way to go. What I do is bend the edged of the pipe halves over so that you have a butt weld between the two sections. This has two advantages: you don't have to allow extra metal in the pattern to be melted when forming the weld. and when the pipe is expanded under pressure the weld isn't being peeled apart. You also have a nicer looking product, without a raised seam.

When you make your patterns you have to be careful of the ends of the pipe. If you just made the ends square across the centerline of the pipe and then welded them together, that end of the pipe won't be able to expand properly. At the small diameter end of the pipe I add an additional 5 or 6 inches on the pattern, converging the sides down from the end of the exhaust to the diameter of the tube that the pump will attach to (I'll get to the pump after a bit). An expansion chamber isn't much of a problem at the other (outlet) end as you can add a section that tapers down to a point. On a megaphone what I do is add a straight section about 5 or 6 inches long and then give it a rounded end. This end will end up buckling and looking very ugly after expansion, but you'll just cut it off and throw it away when done.

After the two halves of the pipe have had the edges bent over and welded together, braze or weld a nut to the pipe on one of the ends that will be cut off and discarded. Drill a hole through the center of the nut and the underlying pipe, and screw a bolt into the nut. This bolt will let you bleed the air from the pipe. Air compresses, water won't, and you want your pump pressure to go into forcing the water to expand the sheet metal instead of compressing air inside the system. Removing the air also eliminates any possibility of explosive decompression of the pipe while under pressure (see the addenda re this subject at the end of the main article).

Now hook up the pump to the pipe, and the water supply to the inlet of the pump. My pump hooks up to a standard water hose, and the line pressure is enough to bulge the exhaust out a noticeable amount. Unless you were very lucky, you will probably notice a couple of minor leaks in the welded seam. Mark them with a felt tip marking pen, drain the water, and weld the leaks shut. Hook everything back up, and if no leaks are noticed loosen the bleed bolt you put in the exhaust and bleed the air out. When you get the air bleed, tighten the bolt and start pumping.

Pump the exhaust up to around 600 psi and take a look at how the pipe is expanding. If you have localized buckling, try to work it out with a small hammer while it is just starting. If you wait until the buckling is severe it is a lot harder to work out. After you've dressed the wrinkles pump the pipe up to around 800 psi and repeat the wrinkle removing step. If your pipe is of simple shape (like a megaphone) this may be enough pressure. If some sections aren't fully expanded go on up to 1000 psi. At the higher pressures you can often work a low spot out by hitting a high spot in.

When you are satisfied that you have shaped the exhaust as much as possible release the pump pressure and drain the water. Cut off the ends of the pipe that were just there to allow the pipe to expand properly. You can now do a little hammer/dolly dressing if you had some particularly recalcitrant wrinkles form. If you need to make the pipe into a multi-planar shape, cut the sections apart, rotate as needed, and weld them back together. You've now got a custom exhaust.

Time: I took me about 4 hours to lay out the megaphone, cut and trim the blanks, weld them together, expand them and do the final trimming. If you were making several identical pipes you could clamp all the blanks together when smoothing the edges, which would speed the process.

The pump: I looked through some industrial supply catalogs for a high pressure hand pump (I pump the exhaust up to 900-1000 psi) without finding anything. I stopped in at a big plumbing supply store and asked if they had any ideas on a pump. It turns out that plumbers use a hand pump to pressurize building plumbing systems to check them for leaks. I bought one of these pumps from the store. The pump is manufactured by Perfecto Products, Inc. and was about $180 dollars (which seems pretty high for a simple pump, but it was all I could find). If your local plumbing store can't locate a similar product I think Perfecto is located somewhere in east Los Angeles.

The pump needs a few modifications before being really useful. I'm adding a check valve to mine downstream of the pump outlet, followed by a radiator-style drain valve, and then the hose to the exhaust. The internal check valves in the pump won't hold the pressure in the exhaust while I work out the wrinkles, hence the additional check valve. The drain valve just makes it easy to depressurize the system. You may also want to add a 1200 psi (or thereabouts) gauge. I think my pump came with an 800 psi gauge, and I definitely had to exceed this pressure to get the pipe fully formed.

I received some posts from people concerned about the possibility of explosive decompression of the pipe when pressurized. I'll list one of these posts and my response, as well as some additional responses by several list members who are much more knowledgable on the subject of pressure vessels and their properties.

Subject: Re: Explosive decompression of blown pipes

> Good article Michael, I'm amazed at what you've done with fairly simple tools. The end result looked great.

> One thing though, I cringed when reading about you pounding out wrinkles on the meagaphone while it was pumped up at 800 PSIG. Weren't you afraid that the thing might burst and injure you? I don't know how much energy is really in one of the megaphones in that condition but the steam plant technician in me wouldn't be pounding on a pressurised vessel- especially one with sucha rupture potential (thin walls, brazed edges etc.). Have you ever had one let loose while you were working on it? I think I'd wear a face shield at the very least, probably gloves,also.

> Dave

Hello Dave,

I've gotten several posts about this, and I think you needn't be concerned. Water, for our purposes, isn't compressible. Therefore, if your pipe holds 1 gallon of water, that is all that it will ever have, and as soon as it ruptures the pressure drops precipitously. With a pinhole leak you just get a thin stream of water that drops off rapidly. Once the pressure is released the water is just sitting there. Remember too that you don't have a massive conduit full of high pressure water, just the little bit in the hose between the pump and pipe.

Now if we were using high pressure steam or compressed gas, which is compressible, I think you might have some problems (ref. exploding gas cylinders, etc).

I'm sure there is someone who can put this into the proper fluid dynamic terms, but as far as I can tell (or have read) it just isnt' a concern with the pressures/quantities of water we use in this process.




Hi Michael,

I've worked a fair bit with high pressures, designing equipment for operation up to 30,000 psi.

You are correct in your assessment of the risks using water vs. air. Compressed air is dangerous at 800 psig. Water is not.

Water is however compressible at higher pressures, and therefore there can be significant stored energy at 20,000 psi or so. Also, if the vessel itself stretches elastically and is large (i.e. composite pressure vessels otherwise known as rocket motor cases being proof tested!), the stored energy can be enough (even with water at lowish pressure) to make knowledgable people nervous. The term used for this type of stored energy is strain energy.

What you are doing seems fine to me as long as you use a pump with a maximum flowrate low enough that the pressure drops off fast with any leak at all. That way you don't cut holes in your skin with the leak.

Congratulations on a nice fabrication job!

Will Creedon


I deal with very high pressures and very significant volumes and temperatures in the course of my work as an industrial pipefitter/instrumentation technician/project estimator/foreman/superintendent/manager/shovel and broom operator. Gasses and liquids both come under the heading of fluids but behave very differently. As Michael Moore points out, liquids are non-compressable and very high pressures can be created in hydraulic brakes/jacks etc with relative safety. Imagine a 100% liquid filled system placed under 10,000 lbs/sq in pressure: The volume of liquid is now compressed to 99.9999% of its original volume. Any rupture or pinhole leak that allows the liquid to expand its volume by .0001% will result in the dissipation of all the energy absorbed by the compression of the fluid.

Real danger arises in hybrid gas/liquid or gas high pressure systems. Imagine the same system above, but with a cubic foot of gas trapped in with the liquid. Under 10,000 lbs/ sq in pressure, the cubic foot of air is compressed to approx. 1/600th of its free volume. Now a pin hole leak or rupture has a 10,000 lb spring behind it that will launch pieces or liquid at positively ferocious accelerations. An apprentice in my Local was paralysed for life when someone tried to test a plastic piping system with #100 air. The 6" plastic pipe ruptured/exploded and threw him across a room and crushed his spine. I am also aware of an individual who tried to seal a pinhole in a leaky auto wheel by firing up on it with a welder without deflating/demounting the tire. The pneumatic spring inside the wheel/tire threw a piece at least part way through him.

Even if you are using a fluid filled system, be very careful of creating gas inside the system. Welding on a closed water filled system can create steam and an instant bomb. Even if you are sure you know what you are doing, WATCH OUT!!

Joe Facer AFM #237

Back to the top frame page
Back to the home page
© 1996-2009 Michael Moore, last update for this page 07 December 2008

For more information contact us by using