The diameter size thing is an issue of exhaust gas speed, or velocity. The idea is that there is a range of speeds that will help extract spent gasses from the chamber, with out going so fast as to create more friction than the inertia of the gas can overcome. It's strictly an issue of the volume of gas versus the area of the pipe. More displacement requires a larger pipe, and more RPM from the same displacement also requires bigger primaries. The length of the pipe controls the the time or RPM that the wave returning from the end of the pipe, reaches the exhaust valve. Longer pipes require more time for the wave to travel back and forth, and therefore work better at lower RPM's.

   There's a number of Formulii for this but most of us use the empirical results from other guys pipes. (we watch to see what works) and then, of course you have the issue of fitting the headers into the car. It's easier to find room for smaller diameter headers.  

Frank/Rebop

Bristol, In ( by Elkhart) 

Thanks for the concise answer.  The concept of inertia of the gas flow makes sense for what I see termed "scavenging" in these types of discussions.  It never occured to me that a wave travels back to the valves and might interfer with gas exiting the exhaust port in the head.  This must be what I see referred to as "regression" in some discussions.  I work a lot on water wells and associate with engineers who do "surge analysis" of water transmission lines, analysing the same type of back-and-forth wave action in a pipeline when a valve opens or closes.  Their goal is to make sure the "water hammer" doen't break something in the pipeline.

You have added a new variable (to me), namely, primary pipe length.  Now the combinations and permutations get staggering.  Presumably, too large a pipe results in loss of inertia and scavenging, too small a pipe results in excessive friction and restriction of exhaust gas flow, and the wrong length of pipe results in wave interference at the gas port.  It appears that to check this out emphirically, one would need to run a number of experiments on a dyno using a number of pre-determined header designs.  The costs would likely be enormous.

Is this why addition of a large-diameter collector is important?  Something has to generate the regressive wave.  If you exhaust directly to the atmosphere at the end of the primary pipe, the timing of the wave should be controlled mostly by the primary length, as you already pointed out.  This must be the basis for "tuned" exhausts.  Evidently, the next best thing to exhausting to the atmosphere is exhausting into a large-diameter pipe, i.e, a collector.  This would be because you are now trying to cram exhaust from two primary pipes at a time (or four at high rpm) through the collector or exhaust pipe; therefore, it seems to me that you simply want to increase the header and exhaust pipe diameter to flow the same volume of exhaust as is coming in from the primaries, and make that flow velocity the same as the velocity in the primaries.  Eventually friction in the collector and exhaust pipe will make this impossible, but by then the distance to where the regressive wave is generated should be so far from the exhaust port in the head that the regressive wave is very weak.  In effect, the collector should have the same effect as making the primary tubes longer.  Does this make any sense?

If the above is conceptually correct, I still don't have a clue how to use it in a practical sense except trial and error experimentation.  I would be curious about the equations, but working with vapor, which is compressible, is a lot tougher than analyzing surge in water, which is essentially incompressible.  My guess is that the equations will have some inherent limitations due to the latter problem and the unexpected results of localized turbulence in the exhaust system which will likely be unique to each and every system.

I am still at least a year out on putting my project together (still building a house first) so would be curious to hear the results of your experiment with the 1-3/4 inch headers.  Considering all the above, it makes sense that engines using a supercharger will need larger primaries simply because they will flow more air than an engine without the supercharger.  However, the loss of torque at the low or mid rpm range may still be an issue since the largest flows through the heads will not occur until the higher end of the rpm range.  Like a lot of things, this looks like an issue of timing and compromise.  The question is, when are we starting to split hairs over very small factors, such as a small loss of exhaust flow at lower rpm, versus causing a large loss of low and mid-rpm performance trying to squeeze that last bit of performance out at the high end?  Life would be simpler if we could run at wide open throttle.

Let me know how your headers work.

Mike K

        I didn't explain some of this very well. The object of varying the length of the primary pipe is to return the negative portion of the wave just as the exhaust valve is closing. Theoretically this aids "extraction" of the chamber. The wave is returned from the end of the pipe by bouncing off  atmospheric pressure. Or in the case of the collecter, from there. Obviously that will only happen over a limited rpm, or in multiples of that number. That's why street headers are longer than race versions. Actually, on the street, the Tri-Y  or 4 into 2 into one thing seems to work. It all changes when you run mufflers and tailpipes too!!

      Collector size seems to also be trial and error, with smaller/longer helping low end. I don't think there's and exact correlation between the total primary area and the collector area. Don't forget, the exhaust gas cools as it travels down the primary, and less heat equals less appearant volume. So the collector is smaller than the total of the primaries.

Edelbrock published some interesting formulas about 15 years ago, when they got into the header business. There's also  a couple of very good books on the subject. One by Phillip Smith, called "the scientific design of exhaust and intakes. ca 1960. Chrysler did a lot of work and published some of it around the same period. (SAE)  And of course the motorcycle guys have done LOTS of this stuff, starting back in the '30's. They make more power per cubic inch without blowers than we do with them!    

Frank/Rebop

Bristol, In ( by Elkhart) 

I think I'm getting a headache.

At least two things are becoming clear to me based on these discussions.  One is that there are so many variables involved, I can see why trial and error has been the most practical approach for those of us without the resources of GM or Ford.  Second, it appears that any specific design is likely to work best within a fairly narrow range of rpm, where the timing of the waves traveling back and forth in the exhaust work best, or in other words, for a specific application, say low end torque versus wide-open throttle performance.  That is where the information obtained by trial and error becomes valuable.  Does that information compare favorably with any general guidelines resulting from the research in the various books?  I will try to find the references you mention.

I'm not sure that after the initlal improvement over stock exhaust systems and some choices about what range of rpm you want the most improvement, we are not getting to a point of diminishing returns worrying about this.  Obviously, getting the headers to fit the car may be a bigger constraint.  After going out and looking at the Fairlane 500 I am using for my project (I haven't worked on one of these since about 1968), my headache is getting worse looking at all the stuff in the way on the driver side.  It makes one appreciate what Jerry Christenson has been dealing with, developing some good headers.  I have a 57 convertible at my farm about 200 miles from here that my brother put Hooker headers on back in about 70 or 71.  I will look at it next time I am up there and see what they look like.

Thanks for all the input.  It gives me a lot to think about.  A final question that I am now just starting to frame, based on all this, is if a lot of the performance of the headers is based on the timing of waves in the exhaust gas, the factors determining the frequency of the waves are the controlling factors.  If this is the case, those factors might not be sensitive to primary header pipe diameter, i.e., the same timing of the waves might be obtained over a range of header pipe diameters.  Then we would be back to the issue of how large we can make the header pipe diameter without loosing inertia and effectiveness of the complementary waves forming in the exhaust system.  Presumably, a Y-block with a supercharger would then accomodate a larger diameter primary tube than a normally asperated Y-block, which is consisent with Ted Eaton's observations in an earlier reponse to my questions that engines with superchargers seem to follow a different set of rules.  Accordingly, an engine with a blower might use larger diameter primaries and still have low and mid-end torque, if the lengths of the pipes and collectors are properly sized.

This gets us back to gas flow and volume.  The arithmetic to determine the volume displaced by an engine at a given rpm is fairly straight forward.  However, the true volume will depend on the pressure of the gas passing by the intake valve, since air is compressible.  How much does that volume change when you increase the pressure to more than atmospheric pressure with a blower?  I think I also have that math, but it must be taken into account.  Well, I am starting to spin my wheels here.  I appreciate the insights and think I will go off and contemplate something simple for a while.

If I can find an affordable dyno in my area after I build my engine, I will probably try to get some different headers and do some experimentation.  For example, I could buy some pipe and build some headers with different diameters and lengths, not made to fit the car, but simply to experiment with some of this stuff on the dynol  I would need a dyno shop that would charge on a daily basis for multiple runs instead of a high price each for a few runs.  Seems like a lot of trouble for probably less than 10% gain in horsepower, but maybe worth it if the overall torque and horsepower curves are improved to eliminate some weak areas.  Mostly, the whole subject had piqued my interest, but I need to find out more so I don't spend my time duplicating something that someone did 25 years ago.

Mike K

Found the book you mentioned on Amazon plus a lot of other stuff on exhaust systems and supercharging.  Some hints that exhaust design for superchargers does not follow same rules as without supercharger.  I ordered five books so will keep myself entertained for a while.  Probably could have bought half a set of headers for the same price.  I don't know why I get off on these technical tangents.  I grew up on a farm where if we couldn't fix it with a hammer, we figured it was probably an electrical problem.  Seriously, I have wanted to study this type of stuff for years, but too busy working and raising a family.  Hopefully, studying some of this will make me more practical, but informed practical.

Mike K

Lorena, Texas (South of Waco)

      I think the diameter of the pipe correlates with the volume of gas. With the supercharger adding to that the increase would be directly proportional to the increase of air fuel in the cylinder. All kinds of variables eneter the picture there. How much boost, how good are the heads (what do they flow) cam timing, what rpm range do you want to use as a basis?

          I'm taking a conservative approach at 1 3/4" because I have a relatively mild engine. Pure stock actually, and I won't change that so it would still be legal with manifolds on. I do know both Ted and John Mummert have discovered the Y's like big primaries. Possibly because the exhaust ports aren't great in the iron heads. The hp numbers I'm making would dictate an 1 3/4 or bigger, (1 7/8") primary. I don't want to chop the car up for fender wells, so I'll build a shorty at 1 3/4 and see what happens. No math, just a guess and a  welder.

Frank/Rebop

Bristol, In ( by Elkhart) 

Charlie,

         This is strictly a guess, but I'd imagine for clearance more than any other reason. It couldn't hurt the flow much either. They actually work pretty good for a 50 year old cat iron manifold. And, Hedman did follow that pattern with their early designs, but the pipes were too small. I had a set on at Columbus this year.   

Frank/Rebop

Bristol, In ( by Elkhart) 

Thanks for the pointer.  I was giving that some thought already.  I notice that the tri-Y headers I have reviewed do not seem to attempt to match up the cylinder firing order.  My understanding is that the tri-Y setup doen't work that well over 3500 rpm anyway.

You are the first guy to mention that you have done some dyno work on your headers.  Do you have any data that would suggest the idea that primaries bigger than 1-5/8 will cost me low end torque on a 331-cid Y-block with 10-12 inches boost at the blower and about 5-6 in the manifold?  Or is this a carry-over idea from engines without blowers?

My heads are ported and I am using big valves (see previous message for details).  In other words this is a Y-block with as good of air flow as possible by porting out the cast iron heads.  I would be curious what your dyno data for the 2-inch pipes indicated at low and mid range.  Of course the cams will probably be different, etc., but if nothing else in your engine and cam design should limit low end torque, I wonder how the 2-inch headers performed.

You will probably get this done before I get started so would like to hear the results.  Are you checking this on a dyno?

If I can work out a deal on an affordable dyno, as I mentioned yesterday, I would like to start out with some long header pipes, making several diameters, and experiment by cutting back the lengths until I see if things are getting better or worse and where the best resuts occur.  Then I would experiment with adding collectors of different lengths.  I would not be trying to make these to fit a car, only to experiment with to see how headers work on my particular engine.  Like you said, some pipe and a welder.

I am in central Montana, so need to do some research on the closest engine dyno facility that might work with me.

Mike K