Heres a answer racer1320 gave on another board. I would say his answer is pretty close.
>
>
>
Most everybody chooses headers that are too big right along with cams and carbs.

Smaller is better!

I also don't agree with the notion of sizing header primary tubes to cubic inches.

Primary tube size should be sized based on the HP a motor is making which in turn dictates where the powerband is/peak torque is achieved.

When you increase primary tube diameter you raise the RPM at which peak torque occurs!

500HP is 500HP regardless of whether it's made by a V6 or V8, 350 or 454 because it takes X amount of hot air to produce 500HP!

So do the math. My motor is making about 650HP at the flywheel and peak torque occurs about 4700 RPM.

Established trends tell us that for every 1/8" increase in primary tube size, we raise the torque peak 500 to 1000 RPM. Same trends reveal that for every 1/2" increase in collector size we raise the torque peak 250 to 500 RPM. So as you can see primary tube size is everything.

Length also comes into play but not to the extent of primary pipe diameter.


Here's my guidelines:

500HP and < use 1 5/8"
525 - 625HP use 1 3/4"
625 - 700HP use 1 7/8"
700 - 775HP use 2"
775 - 850HP use 2 1/8"
850 - 925HP use 2 1/4"
925 - 1000HP use 2 3/8"

 

You didn't say what motor, I would assume you are referring to a 350 SBC. Good rule of thumb: match the header diameter to the exhaust valve diameter for normal street motors. Strip motors and motors with power adders (N2O) can use more header size. Don't oversize. I run 1-3/4" primaries w/ 3" collector on a a fairly healthy 408BBC.

 

Here's a link to an interesting article in PHR about header design by Jack Burns of Burns Stainless. One of the most interesting statements I saw was:

Jack's Tri-Y design pipes are favored by road racers and NASCAR Winston Cup teams, in addition to many of the top NHTA and IHRA Pro Stock engine builders. One of the largest errors street enthusiasts make is in running too large of a primary pipe, and too large of a collector. For example, NHRA and IHRA Stock teams using Burns' headers in H/SA in traditional small-block powered musclecars are running 1.5-inch diameter primary pipes- much smaller than many street machines are trying to use.

There is a little sidebar story about a lady racer who went 2 tenths quicker with a set of these small headers.

http://www.popularhotrodding.com/tech/0310phr_burns/

 

These are some other guidelines I have collected..Yes they do overlap.
IMPORTANT NOTE: "IF MAXIMIZING BOTTOM END-MIDRANGE TORQUE AND TOP END POWER IS NOT OF IMPORTANCE.. USE THE NEXT SMALLER TUBE SIZE"

V-8 H.P.output.. Tube size
140-180********1 1/4" OD
168-220********1 3/8" OD
208-308********1 1/2" OD
300-400********1 5/8" OD
380-500********1 3/4" OD
480-580********1 7/8" OD
560-660********2.00 " OD
620-800********2 1/8" OD
780-1050*******2 1/4" OD
1000-1200******2 3/8" OD

PRIMARY LENGTH AND COLLECTOR SIZING ALSO ARE IMPORTANT IN HEADER SELECTION!

 

Called Header by ED once and was on the phone with him for at least an hour. Anyway, his main complaint about header selections is that most everyone was running too large a header and most racers picked up with smaller tubes.

 

I'm not too sure about these guys that say they need to be at least as big as the exhaust valve either. Its not a constant flow like water,you can run smaller tubes and get the performance you're looking for and/or want Keep the torque down low where our heavy chevy's need it.

 

I am at this crossroad as well...I cant decide if I should go 1 5/8" or 1 3/4" on my 408sbc. It seems like everyone has a different opinion on this one. So from what I gathered it would make my peak torque about 1000 or so rpms lower in the band?? So do you think with a 3500 stall the larger one would be better so that you dont flash stall over your peak torque??

 

Originally posted by hoffbug:
These are some other guidelines I have collected..Yes they do overlap.
IMPORTANT NOTE: "IF MAXIMIZING BOTTOM END-MIDRANGE TORQUE AND TOP END POWER IS NOT OF IMPORTANCE.. USE THE NEXT SMALLER TUBE SIZE"

V-8 H.P.output.. Tube size
140-180********1 1/4" OD
168-220********1 3/8" OD
208-308********1 1/2" OD
300-400********1 5/8" OD
380-500********1 3/4" OD
480-580********1 7/8" OD
560-660********2.00 " OD
620-800********2 1/8" OD
780-1050*******2 1/4" OD
1000-1200******2 3/8" OD

PRIMARY LENGTH AND COLLECTOR SIZING ALSO ARE IMPORTANT IN HEADER SELECTION!

I agree

I think Ed has a good point on lots of people oversizing their headers, but other factors come into play besides HP (intended use, car weight, gearing, stall, etc). Ed's guidelines are right on IMO if you are running a 3900 lb Chevelle --- but on the conservative side if your car is lighter. You cannot cookie cutter parts and assume every car and engine combo is the same (especially when talking V6 vs. V8 --- a 300 HP V6 will have the same primary pipe size requirements as a 400 HP V8, both make 50 HP per cylinder)

Crank stroke comes into play as well because a long stroke SBC (like a 383 or 406) naturally will make more TQ down low and is less sensitive to larger primary pipe sizes. Shorter rods will have the same effect (more TQ down low).

Primary pipe length and collector length is also very important. RPM range will in general be lower with longer primary pipes and collectors. Higher RPM bands with shorter pipes and collectors. An engine may have the same RPM band with a 1-5/8" shortie header and a 1-3/4" long tube header.

JMO

 

Originally posted by Eric68:
An engine may have the same RPM band with a 1-5/8" shortie header and a 1-3/4" long tube header.JMO [/QB]

Eric how much a different will it make if you use a 1 5/8 shorty header and a 1 5/8 long tube header. That the delimma that I have with my exhaust system. How much tq/hp will I lose over the rpm range up to 6000?

 

EA3.0 predicts that from 1500 - 6000 rpm you will lose an average of 41#ft of torque and 29 hp when swapping from a set of 1-5/8" x 40" primaries with a 20" collector to a set of 1-5/8" x 10" shorty headers. The peak loss is 80#ft at 2500 rpm and 43hp at 4500 rpm.

 

There is a formula for finding where a certain pipe diameter contributes the most to an engine.

Peak Torque RPM = (88,200 X pipe area in sq in.")/ 'displacement of single cylinder'.

Since we haven't got the exact inside dia. of the particular headers---I'll use 1.5" for the 1 5/8" headers and 1.625" for the inside of the 1 3/4" headers.

355: 88,200 X 2.355(sq.in.)/44.375(cid 1 cyl.) = 4680rpm for the 1 5/8"
88,200 X 2.55/44.375 = 5068rpm for the 1 3/4"

408: 88,200 X 2.355/51 = 4072rpm for the 1 5/8
88,200 X 2.55/51 = 4412rpm for the 1 3/4

 

What does 'contribute the most' mean? How does length figure into that formula? It seems to me that the length of the tuned pipe would have a much greater effect than the diameter of the pipe. I've used 1-5/8" headers on a variety of 327's, and have come out with different torque peaks on every one.
If I increase the size of the shorty headers to 1-3/4" it makes virtually no difference to EA3.0.

 

we have done this before so with out starting any great debates i will tell my story 66 chevelle drag car lighten up dont know exact weight406 sb car ran 720s and teens in summer with 1 5/8 headers 21/2 pipes no mufflers changed to super comps 1 3/4 headers 21/2 pipes car slowed down 2 tenths in the 1/8th mile no other changes .went back to 1 5/8 car ran 7.20 and teens again .now in the 1/4 mile it may have been alittle different story but 2 tenths would be hard to make up .now my nova i did the same change but when i change i went to wheel well 1 3/4 headers and 3 inch pipes and the times didnt change enough to make a difference the only difference in the motors was the heads the 66 had sportsman and the nova had iron eagles both 200cc runners

 

On my old 383 SBC that ran 7.30's at 3600 lbs and dynoed at 389 RWHP, I saw no difference at the track in switching from 1-5/8" Hedman headers to 1-3/4" Hedman headers.

Granted I did make the header swap when the car was running 7.50's-7.60's.

I did see a 2-3 MPH gain in the 1/8th going from 2-1/2" exhaust w/3-chamber Flowmasters to 3" exhaust with Flowmasters. The 2-1/2" system did have an H-pipe. The 3" did not.

I later added an H-pipe to the 3" system but it showed to be worth little to nothing at the track. It did however quiet the car down quite a bit at idle. It also seemed to smooth the idle out a little.

 

"Contibute the most". The area of a pipe cross section as used in the formula is the most important part of "where" the pipe contibutes the most to the engine. Pipe lenght rocks the curve, but doesn't change the RPM the peak occurs at unless we've miss-matched in the intake manifold and/or cam. By rocking the curve, I'm referring to raising the average above the peak vs below the peak. We do the same with cam centerline. Intake manifolds are also matched to the planned power curve with "port area". Again, lenght will rock the curve without effecting where the peak occurs if we are in the ballpark with the other parts.

The formula came to me via Jim McFarland.

 

I'm still kinda confused here. Which came first - the torque peak or the header diameter? If I swap from 1-5/8" headers to identical 1-3/4" headers should I expect the torque peak rpm to increase by 20%? What is the physics behind the calculation? I have a reasonable understanding of reflected waves and scavenging pulses, but this linear diameter effect isn't clear at this point. It basically says that every engine of a specific displacement has the same torque peak if it has the same diameter headers. If the cam timing, cylinder head design, intake manifold design, etc. have some effect on the torque peak also, how is that entered in the equation?

Peak Torque RPM = (88,200 X pipe area in sq in.")/ 'displacement of single cylinder'.


My numbers seem to come out a bit different than yours. For a 1.5" ID tube I get an area of 1.77 sq.in. and a 1.65" ID tube comes up to 2.14 sq.in.
For the 1-3/4" tube the peak calculates out to be 4253 rpm for the 355 and 3700 rpm for the 408. For a 1-5/8" header the peaks would be at 3498 rpm and 3043 rpm for the 355 and 408 respectively.

If I put a set of 1-1/2" headers on my 11.6 second H/SA Camaro with a 410 hp 355, would I expect it to pick up a tenth and a half over the more traditional headers I have been running? 1-1/2" headers would have a torque peak of 2972 rpm in a motor that never sees the low side of 5500 rpm.

 

Guys, for inside diameter of a primary header just take the OD and subtract 2x the wall thickness.

Most headers have primary pipes made from either 16 or 18 gauge DOM mild steel tubing. 16 gauge is .0598" thick where 18 gauge is .0478" thick. So the ID of a 1-5/8" 18 gauge header tube is expressed as follows:

1.625" - (.0478 x 2) =
1.625 - .0956 = 1.5294" or 1.53"

Sorta half-way in between 1-1/2 and 1-5/8.

DEEBO - I have no idea . . . BUT its safe to assume the shortie header will be a bit short on TQ down low when compared to the long tube header and will MAYBE make a few more HP on the top end.

Primary tube length WILL shift an engine's RPM band up or down BUT its impossible to simplify that enough to say "going from 1-5/8 to 1-3/4 shifts every engines power band up 1000 RPM" or "going from shortie to long tube will shift it down 1000 RPM." It's much more complicated than that and other engine design factors play a significant role - intake, cam profile, compression ratio, cylinder head flow, etc.

If you want to look at this more closely try going to

www.headerdesign.com and play around with their calculator.

PS. Exhaust pipe size DOES NOT affect the engine's RPM band the same as primary tube and collector design. Basically, as long as there is no restriction, any size exhaust pipe will have the same RPM band. In other words, larger or smaller pipes will not shift the engines RPM band up or down only only restrict and kill power if too small or be too loud if too big.

 

I do not think swapping in 1/3-4" headers from 1'5/8" headers will increase the peak torque rpm of the engine by 20%. The formula stated above predicts that the larger header will tend to favor peak torque at a higher rpm.

However, the entire engine combination, and each component (heads, cam intake, etc.) that influence the shape of the overall torque curve must be considered. Let's say that cam, heads, etc would produce peak torque in a given engine at 4,500 rpm, but the above formula predicts the headers will favor a torque peak at 3,200 rpm. The headers will not shift the peak torque of the engine as measured at the flywheel down to 3,200 rpm. I imagine the difference between peak torque produced when using 1-5/8" headers vs 1-3/4" headers would not vary by much more than a couple hundred rpm, if that.

I would imagine the smaller header would contribute to increased overall torque and throttle response below the peak torque rpm compared to the larger header. The larger header would favor slightly higher HP above the peak torque rpm.

Would a 1-1/2" header increase performance of an engine that rarely sees under 5,500 rpm? Probably not. However, when choosing headers one should attempt to maximize torque throughut the anticipated operating rpm range of a given engine combination, and not just peak torque rpm. In a street application, headers selected to produce their torque component 1,000 - 1,500 rpm below the peak torque rpm of the engine will fatten the torque curve below peak torque without sacrificing significant performance 1,500 rpm or so above the engines peak torque rpm.

 

Here's the Camaro with the small headers.

http://www.popularhotrodding.com/tech/0310phr_burns/

 

All this theory is fine and dandy, but what about the physical aspects of the headers and the heads exhaust ports? For example, on my 388, I am using pro-topline 200cc iron heads. I fully intended on using the 1 5/8" full length headers that came with the car, but comparing a header gasket that was port matched to the header to the exhaust port of the head showed that the round port gasket had a quite substantial amount of overhang into the larger, more square shaped exhaust ports. Is the loss of flow around the edges of the blocked exhaust port negligable, especially compared to the torque gains from using the smaller headers? I would much rather have used the smaller tube headers simply for clearance reasons but just didn't get a warm fuzzy feeling about the 1 5/8" headers with such an obvious port mismatch.

Of all the header discussions on here and the camaro site, I have never seen this issue addressed. Is it even an issue? I certainly looks like it would be.

 

You got it travis. Port matching is one thing I have never seen anyone talk about. I had exactly your situation when I installed my WP Sportsman heads. And, it makes a huge difference. With the 1-5/8" headers my car was slow to respond and would not spin the tires under a 2500rpm launch. I installed the 1-3/4" headers and retuned, then magically I could light them up at about 1800rpm. My low end torque increased with a bigger header. Where the 1-5/8" headers adequate by my mathematical standards? Yes they where. In reality did they work? No, they didn't. This is why picking headers is so much more complicated than just saying "OK, I have 'X' hp, I need a 'Y' diameter header."

 

When you buy headers, buy ones with the right flanges. They match the ports already.

When you look at header comparisons, seldom do you see two sets of headers compared that are as alike as two carburetors. Most header swaps going from 1-5/8" to 1-3/4" involve using the same 3" collector diameter. The relationship of diameter of the collector vs. the primary plays a very important role in the operation of the headers. If you went from a set of x" diameter headers tuned to a specific engine and rpm band and replaced it with an identical y" diameter set tuned to the same engine and rpm you would get a much more meaningful test. Ed Henniman has a few thoughts about magazine header testing...

http://www.headersbyed.com/info1.htm

 

So is it even possible to get 1 5/8" headers that would match my exhaust ports, or would they have to be custom built? Any suggestions on what would be a better performing header for my application? Or better yet, a better fitting header I have headman 1 3/4" headers with 3" collectors...spark plug access on several cylinders is horrible, as is clearance around the steering shaft. How about just adding something like the 4-2-1 collectors offered by flowmaster (and others)? I don't remember what the tube length on my headers was but it was like 29", with 13" collectors if I remember correctly.

 

Originally posted by travis g:
Or better yet, a better fitting header I have headman 1 3/4" headers with 3" collectors...spark plug access on several cylinders is horrible, as is clearance around the steering shaft.

So why the heck did you buy them dang Hedmans for anyway? Geez, who's the idiot that recommended them to you in the first place . . . LOL

Seriously though, some headers have a D port, some have a square port, others have a round port. Just buy headers with flanges that match your heads. The header catalogs will tell you what shape flange they have.

My 1-3/4" Hedmans fit great and have a square port. My TFS heads have a D port. The square header flange is bigger than the exhaust port on the top and bottom of the opening except a couple small spots on the sides that I fixed with my Dremell. If you are carefull you CAN port your header flange just a little but only if you cut on the header flange - if you cut on the tube you are done.

Having a header tube that is bigger than the exhaust port outlet is fine to a point. A "step down" in theory would help prevent reversion while a "step up" would cause a restriction.

JMO.

 

Sometimes you have to look past the street type headers to get the type of flange you want. Expect to pay a little more for a race type header. True custom headers will run upwards of $700.