Who has made their own headers?
 

I've seen a couple of the HP Junkies with very impressive home-made header pics. Who has done this and what process have you used? Just jack the car up and start to fab it up as you go?

Suggestions, tips, pics, guages etc are most welcome. Show off your work!


Also, where can one get the materials? (proper flanges, pipe etc)

I made some SS shorty headers for my 930 motor. I tig welded them out of 321 SS which I bought at Burn Stainless. Mark Hargett made some nice ones also but he got his materials some were else. And his radiuses are smaller. I made the flanges myself out of steel. Steel is much easier to work with than SS. Here is a pichttp://forums.pelicanparts.com/uploa...1065109164.jpg I had the motor on a stand when I built mine. Pay no attention to the tail in the pic;)

Nice work Dean...I was thinking of your project when posting the thread.

BTW, any advise on a buying a used TIG welder? I've been keeping a casual tab on EBAY for used Miller TIGs. However, although the price is quite attractive on some, there are quite a few models with some weighing a ton. I'm not educated on Arc welders with TIG conversion kits (?), AC or DC units etc.....I'm kind of in the dark here. Any suggestions?

Here are a couple of shots of the ones I built. I too made them out of 321 SS. If your application isn't turbo'd you could use 304SS or mild steel and hotcoat it. Doing it with the motor in the car would be a challenge to say the least. The cuts have to be very accurate indeed, there is no gap welding with TIG! If you've never TIG'd before I suggest you get a lot of practice in b4 you start on the headers. TIG is beautiful welding but takes some time to get used to. As far as welders, I would highly suggest the ThermalArc 185. It's the one I bought. It is very compact compared to Miller's. You won't beleive that such a small welder can be good but with today's electronics, bigger is not better. The TA185 is highly rated in the industry and I have to say I am more than pleased with it. It also only draws 35 amps so you don't have to install a special circuit breaker and dedicated line. Feel free to ask any questions.

http://forums.pelicanparts.com/uploa...1065125249.jpg
http://forums.pelicanparts.com/uploa...1065125290.jpg

MArk, very cool.

When designing the headers do you attempt to make equal length runs?
Would that be ideal? BTW, the flanges are looking real nice.

Also, I've been keeping an eye on the older Millers because they go for relatively low $. Some decent looking units go from $400-$600 (less shipping). I'll live with the older technology if it's tried-n-true and inexpensive. Does the Thermal Arc fall into this category?

Below is a set of headers I fabricated for a 910 last year. And yeah, I spent a fair amount of time getting all the primaries to be equal length.

Agreed, tig is the way to go. I might humbly recommend spending time to fit up your pieces nicely before welding. Definitely makes for a much nicer header that doesn't move around too much during welding. Also, getting the radii to line up tangentially really helps flow at the higher revs. A good belt sander is invaluable.

For components, I like Burns Stainless and Specialty Products Design in Rancho Cordova, California. Both of these guys have really nice mandrel bends. Lately, I have been having flanges laser cut or water jet cut. However, I'm told that Stahl has flanges in their catalog as a stock item; they might be worth a call.

Stainless is beautiful stuff to weld, however it's spendy and hard on cutting and abrasive equipment. Mild steel is the way to go for your first time out. Steel also offers you the option of getting steel parts Jet Hot coated (or some equivalent) when you're done.

Good luck and have fun!

Cheers, Chris

PS. Check out the Metal Fabricator's Handbook by Ron Fournier. It's a great resource.



http://forums.pelicanparts.com/uploa...1065132061.jpg

Wah! Chris, that is some beautiful craftsmanship.

Dumb question....in order to attain equal length would you purposely make a primary longer than it would need to be otherwise?

Rick, I fantasize about fabricating headers. My dream when approaching semi-retirement is opening a dyno shop!!!! :)

My amateur thoughts:
1) You can buy a TIG welder that will do only stainless and mild steel. It won't do aluminum. Those TIG welders are cheaper (around $700-900).

2) This is time when you really could use a horizontal bandsaw. The Harbor Freight, Commie-made, Chinese POS bandsaw I use has paid for itself many times over. I regret buying Chinese, but I don't regret buying the machine. :)

3) I think welding is like painting cars: practice, practice, practice.

4) I think building headers can only get you into more trouble. What about the complimentary hardware to go with your masterpiece? EFI, cool turbos, trick Porsche hardware. It is like a disease, except the only cure is to spend more money... :)

5) There are formulas for primary lengths, and lengths are suprisingly long. For normally aspirated, you want the equal length, but some guys forego equal length when you go turbo. There is so many ways to make power, and more than one way to succeed. Challenging stuff, for sure.

Good luck,
Jürgen

We are just in the process of fabricating headers for our project car.

I chose SS2333 stainless (non acid-resistant as it has slightly higher thermal limits than SS2353 which is acid-resistant).

I'm a "calculating" guy so i CAD-ed everything long before i went out to shop (expensive) tubing. That way i know how it fits, how it should be sawed, how many pieces are needed etc.

That being said, SS bends a lot when TIG-ed, so you can never cut all pieces and expect them to fit afterwards...so CAD is just an aid and cannot compensate for skill.

This is the model:
http://project911.luminasweden.com/l...s/MK_III_2.JPG

Pics of headers (actually GT40 collector pipe which is critical and thus welded first) are comming soon!

Looks cool Beep, what program do you design with for 3D? I have used ProDesktop Express for a while and love it. I've recently been playing with AutoDesk Inventer. Just curious cause I love doing this stuff in computer first like you said.

Wow...some hot ***** stuff here! Beep, your CAD illustration is neat. At first I ws going to guess Pro E, but I don't thinks so.

Looking more closely, I saw an interesting detail. At the engine head flange of the nearest tube, it looks as though the diameter necks down, and flattens out a bit as it connects to the head flange. And, to some degree, the tube itself looks as though it is not a constant diameter. Rather it appears to expand and contract as a two cycle pipe might. Is that the case? If so, fill me in on the theory and details!

May I ask an elementary question - is it practical to make your own bends, or is it necessary to buy the "elbows" (dunno what you call them)? I vaguely recall being in a P-car mechanic shop and being told they bent their own headers there.

It is definately NOT practical to bend your own bends. You really need to use mandrel bent tubing. It's the only way to maintain the roundness of the tube throughout the full radius of the bend. If they aren't mandrel bent they will have wrinkles on the inside of the bend which leads to unwanted turbulance and restriction. Mandrel bends are not as spendy as they used to be. You can get mild steel for @15 a pop. The SS gets very expensive. The 321SS ones will cost you 30-50 a piece.

Good welding. But is it still good to oxy-acetylene weld exhaust systems?

It might look as headers are made of custom-bent elbows but everything is assembled with off-the-shelf SS 44.5mm 90-degree bends that are cut in half at certain places. But your observation is correct, short pipe-stubs that are welded into head flange are actually oval on one side and circular on the other. It's done that way beacuse exhaust ports are oval (it's 4v/cyl head). Actually, those pieces are nothing else but 40mm long pieces of tubing squashed on one side with vice-grip, stuck into each port and welded (circumference of flange port is equal to pipe-circumference so it's perfect fit after it's squashed). GT40 collector/merge pipe is made from two straight pipes angled 20-degrees from vertical flowing into each other. Theory is that gases should accelerate before they go into turbo so it should spin better. Two stubs stuck in turbo-flange are somewhat special though as they are squashed "with offset" so they clear each other.

Here are some wireframes that might help you understand better:
http://forums.pelicanparts.com/uploa...1065168764.jpg
http://forums.pelicanparts.com/uploa...1065168788.jpg

Initially, i was planning to make real pulse-split pipe with ports 1-4 and 2-3 going into each turbo-port (it's twin-scroll turbo) but came to conclusion that it's not worth it. It would be too hard to make a good turbo-collector and keep weight/dimensions down with such design so i went for current one.


All that is needed to make that headers is this (turbo-flange not visible):

http://forums.pelicanparts.com/uploa...1065169602.jpg

Whan it comes to software, i used MORAY as (CSG) modeller and POV-Ray as renderer. I find Solidworks too cluttered for such mundane tasks. All i needed is to se how it fits and measure some dimensions.

Again, great stuff! Any more advice on an inexpensive used TIG welder.


Also, If I have a 14" cut-off saw why would I need a horizontal bandsaw? They seem to serve a similar purpose.

Please keep the home-made header pics coming.

In the top-down wireframe view of the header, the two center primaries have a dogleg. How do you make this bend? I know you must cut u-bends and use 2 separate bends, but how do you ascertain exactly where to make the cuts? Is this just a case of cut a little at a time until it fits perfectly?

I can draw a pic if my description is confusing.

jürgen

I thought pro e was the only program to make 3d models with. I tried it on auto cad before, and although successfull, it made pro e seem like a vacation.

And lateapex, i don't think beep was trying to make expansion chambers. They are designed on 2-cycle motors because they require something to create suction on high performance engines. Basically it uses the natural affect of hot gases to want to expand, and in the chamber it accelerates the exhaust for a short period of time to create that suction effect. If you ever watch racing 2 cycles (dirt bikes and such), the ones designed to with "shorty" pipes are designed to create peak suction at lower exhaust flow rates. The ones designed with elongated expansion chamber create peak suction at high flow rates. I think beep as stated just made it oval to fit his saab motor's ports.

Your observation is perfectly correct. Turbo-flange has dual cut-outs and two symetrically squashed pipes stuck there won't fit one to another. In reality, they will be offset-squashed and somewhat angled until they fit into holes and clear each other. CAD-drawing of those primaries is made by Bezier-patches. Altough it's fully possible to press them to look that way, it involves expensive equipment and end result will probably be cut-press-bend job.

It's important to plan headers in modular fashion. With other words, you must be able to compensate for SS warping with straight pipes at certain places. In this case, two straight merging pipes are fabricated first, untill they merge in perfect circle, then you just need to worry how to route them into turbo-flange. Same thing goes with secondary pipes. If primaries going to turbo are too wide spaced, no problems, you just weld little longer straight-pipe segments until it fits. General rule is to build collector as good as you can (it's most important piece, at least in turbocharged car), then weld oval-to-circular "adapters" into head flange and finally, fill in tubes in-between leaving straight sections last so you can compensate for warpage. Also, it's very important to bolt flanges to heads/turbos before you start welding or you will end up with totally warped piece that cannot be repared (and yes, we're talking 15mm+ thick SS pieces).

Here's exploded view of our header:

http://forums.pelicanparts.com/uploa...1065200033.jpg