S54 Header Design talk

[3staxontheradio]

Hi All,
I'm doing a clean sheet look at S54 header design. Eventually I will put everything up up in a github, including all scans and any cad files. There are many headers for the S54 but ther are are a few reasons I want to look at this again in 2025:
1) 3d scanning makes more complicated packaging much easier
2) Additive manufacturing. I have been using DMLS for parts and its reasonably affordable now. I think there could be some cool stuff possible with additive that wasn't necessarily feasible before.
3) Better informed header choices specific to people's cars and engine builds.

I have a first scan of an S54 head. I'll be re-scanning for a cleaner result before sharing it but its good enough for the port length and volume measurements I have been using. I have a stock cam S54 set up in Pipemax and I think its getting close to being configured correctly. It's putting out suggested header dimensions. I'll be building this to fit my S54 E36 M3 but I have an e46 chassis here that I can scan as well.




Questions for everyone here: I would really like to get my hands on some existing headers to measure and/or scan. Alternatively if anyone can carefully measure theirs that would be great. I'm interested in looking at what else has been done and seeing where that fits in with these header calculations. Specifically I'm interested in scanning or measuring a supersprint V1 and factory euro headers. CPI headers would also be good to get details on.

Important measurements are:
Runner lengths measured as well as is reasonable
Primary ID and (and OD too for completeness).
Collector/secondary ID dimensions


Pipemax as it is currently configured for stock cams:


And the header calculations for a stepped header:

Calculated Exhaust Port Flow @7 kPa= 76.7064 -to- 82.1417 L/S @ 9.956450 mm Lift (no Flow Pipe)
User's Exhaust Port Flow input @7 kPa= 94.3895 Liters/Second @ 12.000001 mm Lift (no Flow Pipe)

Engine Application = Race Engine • Mid-Range Peak TQ • Hi RPM Peak HP

Exhaust Header Type = --- Primary Tube Specs : Race Header • Multiple Steps ---
--- 2-Step Primary Pipe Specs --- ( Total Length= 555.329 to 646.081 mm )
1st Step Diameter mm= 40.556 Length= 277.664 to 323.041 mm
2nd Step Diameter mm= 43.731 Length= 277.664 to 323.041 mm

--- 2-Step Primary Pipe Specs --- ( possibly higher RPM HP )
1st Step Diameter mm= 40.556 Length= 138.832 to 161.520 mm
2nd Step Diameter mm= 43.731 Length= 416.497 to 484.561 mm

--- 3-Step Primary Pipe Specs --- ( possibly higher RPM HP than 2-Step Headers )
1st Step Diameter mm= 40.556 Length= 277.664 to 323.041 mm
2nd Step Diameter mm= 43.731 Length= 138.832 to 161.520 mm
3rd Step Diameter mm= 46.906 Length= 138.832 to 161.520 mm

--- 3-Step Primary Pipe Specs --- ( possibly higher RPM HP than 2-Step Headers )
1st Step Diameter mm= 40.556 Length= 138.832 to 161.520 mm
2nd Step Diameter mm= 43.731 Length= 138.832 to 161.520 mm
3rd Step Diameter mm= 46.906 Length= 277.664 to 323.041 mm

--- 4-Step Primary Pipe Specs --- ( possibly higher RPM HP than 3-Step Headers )
1st Step Diameter mm= 40.556 Length= 138.832 to 161.520 mm
2nd Step Diameter mm= 43.731 Length= 138.832 to 161.520 mm
3rd Step Diameter mm= 46.906 Length= 138.832 to 161.520 mm
4th Step Diameter mm= 50.081 Length= 138.832 to 161.520 mm

--- Primary Tube Harmonics --- ( One-End-Closed Tube = Odd Numbered Harmonics )
1st Harmonic = 1825.687 to 1916.439 mm long ... typically never used ( too long to fit any Vehicle )
3rd Harmonic = 555.329 to 646.081 mm long ... highly recommended , best Torque and HP Curve
5th Harmonic = 301.257 to 392.009 mm long ... shortest recommended ( Shorty or very Hi-RPM Header )
7th Harmonic = 192.370 to 283.122 mm long ... shortest recommended ( Shorty, Hugger, or Log Style )
9th Harmonic = 131.876 to 222.628 mm long ... shortest recommended ( Log Style or Cast-Manifold )
11th Harmonic = 93.381 to 184.133 mm long ... shortest recommended ( Log Style or Cast-Manifold )
------------------------------------------------------------------------------------------------------
--- Collector Specs : Merge Collector • Straight Tube ---
Peak TQ Diameter Range = 64.168 -to- 67.343 Best Length= 499.338 -or- 998.675 mm
Best Mid-Range Diameter = 67.343 Best Length= 499.338 -or- 998.675 mm
Peak HP Diameter Range = 67.343 -to- 70.518 Best Length= 499.338 -or- 249.669 mm

H-Pipe Location= 499.338 or 249.669 mm X-Pipe Location= 1997.351 or 998.675 mm
( both H-Pipe and X-Pipe locations are measured from Primary Tube ends inside the Collector )

Dual-Exhaust System Diameter = 64.168 to 70.518 mm Dual-Exhaust's each Muffler L/S = 187
Single-Exhaust System Diameter = 90.747 to 97.097 mm Single-Exhaust's one Muffler L/S = 374

--- Collector's Harmonics --- ( Both-Ends-Open Tube = Odd and Even Numbered Harmonics )
1st Harmonic = 3994.701 mm long ... longest recommended with Mufflers and TailPipes
2nd Harmonic = 1997.351 mm long ... longest recommended with Mufflers and TailPipes
3rd Harmonic = 998.675 mm long ... greater Low RPM Torque -to- Peak Torque RPM
4th Harmonic = 499.338 mm long ... highly recommended , best Torque and HP Curve combination
5th Harmonic = 249.669 mm long ... reduced Peak Torque , higher RPM HP gains possible
6th Harmonic = 124.834 mm long ... reduced Low RPM Torque , even though Tuned Length

Best TQ + HP Tuned Collector Lengths= 124.834, 249.669, 499.338, 998.675, 1997.351, 3994.701 mm long
Worst TQ + HP Loss Collector Lengths= 187.252, 374.503, 749.006, 1498.013, 2996.026, 5992.052 mm long

Collector definition: from the Primary Tube's ending inside the Collector -to- Atmospheric exit point
Note : all Tube Outside Diameters are based-off your Header Tube Thickness choice's value
------------------------------------------------------------------------------------------------------
Bore=87.00008 mm Stroke=91.00007 mm 3.24580304 Liters @ 7900 RPM Intake System= 100.00000 VE%
Complete Intake System Flow @7 kPa= 94.9698 -to- 101.6991 L/S @ 12.000001 mm Lift (5.00000 VE% Loss)
Cylinder Head Intake Port Flow @7 kPa= 122.7063 -to- 131.4010 L/S @ 12.000001 mm Lift (105.00000 VE%)
Cylinder Head's Exhaust Port CenterLine Length = 79.8500 millimeters
Target EGT= 1295.3 degrees F or 701.9 degrees C at end of 4 second 600 RPM/Sec Dyno accel. test
EGT Probe location = 19.05 to 25.40 mm from Header gasket flange at 12:00 O'Clock position
EGT Probe tip depth = 12.70 to 19.05 mm depth into Header Primary Tube
Speed of Sound = 626.0 Meters per Second at 25.40 mm distance into Header Primary Tube
Header Collector Spear Length = 0.0000 mm Exhaust/Intake Port Flow Ratio = 76.923 %
Exhaust System operating RPM Range from 5900 to 8400 RPM Hertz frequency = 65.8 Hz at 7900 RPM
------------------------------------------------------------------------------------------------------
Camshaft = OverHead • Finger - Follower Arm
600 RPM/Sec Dyno Test Level=7 Level=8 Level=9 Level=10 Piston MPM
Peak Power KiloWatts @ 7900 RPM 258.3 261.6 264.9 268.3 1437.80
Peak Torque Ntn-Meter @ 4900 RPM 338.9 346.0 350.4 354.8 891.80
------------------------------------------------------------------------------------------------------
Fuel BTU=18310.1 Air/Fuel Ratio=12.686790 BSFC=276.2444 Mixture Distribution= 90.0 Quality= 90.0
Dyno HP Weather Correction Equation = 8• SAE J607 (June1974) • STP • SuperFlow-FTQ • (Default)
HP Correction Factor= 1.000000000 Fuel Type= Ethanol • E10 • 10% Ethanol • 90% Gasoline
Station Barometer=1013.20748118 Air DegC=15.56 Vapor Pressure=0.000 Air Correction=1.00000000
Station Barometer NOAA= 1013.50748075 Pressure Altitude Meters= 0.4 Z•Elevation Meters= 0.0
Density Altitude Meters = 20.5 Relative Humidity % = 0.00 Dew Point DegC = -164.11
Virtual Temperature DegC = 15.56 Water Grains = 0.00 Wet Bulb DegC = 3.74​
​

My current pipemax config for anyone who wants to mess around with it

https://limewire.com/d/rpUDE#8t2YhD7kX9

[Bry5on]

I have much of what you seek, will share more when I’m home

[tlow98]

Let the party begin!

[bimmerfan08]

Would you use DMLS for the finished part or proof of concept?

[3staxontheradio]

If I use DMLS it would be for a finished part. Also it's very possible I would just do some parts of the header with DMLS and mix in traditional tube sections. It will just come down to a mix of cost and what I can get from additive than I can't from cutting and welding bends.

[Cubieman]

Man, when is someone going to make some worthwhile aftermarket followers? I fear those will disappear in the near future.

Not to take away from this topic, incredibly cool stuff!

[ZiMMie]

Great idea — I'll just leave this here as well.

There’s a company called Celeritech that produces 100% equal-length headers based on precise 3D Chassis Scans.
They currently offer what are widely considered the best S85 headers on the market right now.






Links:
Celeritech Website

https://youtu.be/4XHE-eXWz50

Price List (PDF)

Their prices are quite reasonable, and once you’ve finalized your simulated sizes, they could be a solid option for production.

It would also be great to include the E36 chassis in a scan for header design, as the differences are very minimal.
With this, the headers could potentially be made to fit both chassis.

[Obioban]

Good thread! Not being constrained by what can be done with bent tube should allow for some meaningful improvements over what's out there now. Port match flanges that transition well, exactly equal length, easy install, optimized steps at equal distances, cam specific should all be viable!

... if outside of my skill set

Just to get people excited for what can be printed for about the price of a set of high end headers these days:

[3staxontheradio]

Yep couchbuilt has excellent YouTube videos and his content was how I originally found out DMLS had become affordable. I did my shift carrier and linkage with DMLS aluminum and stainless and it was cheaper than off the shelf parts and a better result.

I think there are some pretty cool possibilities. Anti reversion geometry is high on my list of things to explore and integrate. I also want to get pretty far into how the factory exhaust ports function - they don't look ideal to me and that will have an impact on what the flange at the header looks like. A matched/blended shape is very likely not the ideal transition.

I really want to get a supersprint v1 to measure because at least from eyeballing it I think it's tuned on a different/less effective harmonic (as are all the off the shelf headers i have seen?) or they are targeting or a much higher peak torque rpm. My current peak torque target is at about 5k. That's inline with where the factory has the peak torque tuned to.

All the header math will get validated with a second source in some way, I don't want to trust this software blindly but I know drag racing people really like the output of the inputs are correct.

I'm also going to configure it all for different cams / engine VE. Runner diameter will increase as VE goes up.

[George Hill]

I saw your posts on FB (and offered a head to test), glad you are bringing this info to the forums too.

These are what I am messing with right now. Kromer Kraft headers, originally designed for an E36/S54 but I'm adapting them to the E46 chassis. Kromer was developing this system (and M54/S52 options) in mid 2000s. Time has marched on and there may be better options now but back then they were up there as something you could purchase for an all out car.

6-2-1

Measured #6 primary, it has 2 steps - 44mm to 47.5mm to 49.5mm od (roughly), lengths are 7", 9" and 8".

The 2-1 'Y' necks down to 60mm od











I had these on the car 4 yrs ago but pulled them off before fully developing the setup. This was the first dyno and tuning session with them compared to another car same day, same dyno.

Blue is my car - S54, GS6-53 trans, KK Headers, HTE Tune, Karbonious intake, Alpha N
Red is a clients car - (ZHP) S54, 420g, Euro headers and cats, Supersprint conversion section 2 & 3, stock intake, stock Euro tune





I'm getting them fixed up and are going back on the car this year.



*****Edit - I forgot to add I chatted with a guy awhile back who has these on (2) of his cars and both make 340+whp on stock long blocks as measured on a Dynojet. I can validate his numbers but comparing his graphs to mine the lower RPM follow mine and then his just keeps making power where mine start to fall off. He had a different tuner and the exhaust setup is a bit different after the single, but just something to compare to for me.

[3staxontheradio]

That's really good info. If I take my pipemax config and up the VE to get something like 370 crank (reasonable proxy for 340whp?) I get suggested header ODs with 3 steps of 42.7mm, 46mm, and 49.1mm. That's assuming 16 gauge tube which is most common. It wants a total primary length between 545mm and 636mm. These numbers are really close to that header. The range of primary lengths are also impacted by how I set the lobe separation and how that impacts overlap but that's the ballpark length. Also pipemax wants the 2-1 merge/x pipe/ collector length to be 980mm from the merge. I wonder what would happen if you moved it back.

[George Hill]

That's really good info. If I take my pipemax config and up the VE to get something like 370 crank (reasonable proxy for 340whp?) I get suggested header ODs with 3 steps of 42.7mm, 46mm, and 49.1mm. That's assuming 16 gauge tube which is most common. It wants a total primary length between 545mm and 636mm. These numbers are really close to that header. The range of primary lengths are also impacted by how I set the lobe separation and how that impacts overlap but that's the ballpark length.

That's really interesting! Cool to see some math backup old school knowledge, I assume these were designed when computing power wasn't as advanced and more "old school" knowledge was relied upon.

*Edit and to be fair I am more doing this for sound than anything. I really don't care *that* much about the loss of torque with this design as my car has an 8HP and 4.44 gears so once it gets going it never drops into the lower RPM where the torque doesn't exist. Driving on the street may be less than ideal, but meh I guess. lol

https://youtu.be/juk-elMFBZY

[Bry5on]

Yep couchbuilt has excellent YouTube videos and his content was how I originally found out DMLS had become affordable. I did my shift carrier and linkage with DMLS aluminum and stainless and it was cheaper than off the shelf parts and a better result.

I think there are some pretty cool possibilities. Anti reversion geometry is high on my list of things to explore and integrate. I also want to get pretty far into how the factory exhaust ports function - they don't look ideal to me and that will have an impact on what the flange at the header looks like. A matched/blended shape is very likely not the ideal transition.

I really want to get a supersprint v1 to measure because at least from eyeballing it I think it's tuned on a different/less effective harmonic (as are all the off the shelf headers i have seen?) or they are targeting or a much higher peak torque rpm. My current peak torque target is at about 5k. That's inline with where the factory has the peak torque tuned to.

All the header math will get validated with a second source in some way, I don't want to trust this software blindly but I know drag racing people really like the output of the inputs are correct.

I'm also going to configure it all for different cams / engine VE. Runner diameter will increase as VE goes up.

The super sprint V1 headers are unfortunately very much not equal length, so you may have trouble simulating their output. This is what I run in my car and has had me dissatisfied enough that I ventured down the path of designing my own 3D printed/mandrel bent stepped headers as well. But my 3D printed suspension knuckle project has pulled me away from the project

In any case, I have decent scans of an e46 engine bay and an s54 from George Hill's spare front clip and motor. I've also got scans of the US headers (important note that the banks switch sides L-R on US headers for packaging reasons), but haven't scanned euro headers yet.

The US headers (which I'd wager are similar in length/diameter to euro) runner length is approximately 246mm by my numbers after remodeling the header to fit, and I measured ~70mm for the exhaust port length, although my scans were not great for this. Your data looks way better. If you shoot me your email, I'll send you a link to my Fusion CAD file, which contains scans of the chassis, engine, and both headers, plus has bank two runners modeled to match the header.

The runner tube is 40mm OD, 1.5mm wall (37mm ID) and all bends in bank 2 are 40mm radius.

Hope that helps for now! If you're working in a CAD environment that allows simultaneous work, I'm happy to help there as having another person work this is providing some extra motivation to get back on the horse.

[Mike RT4]

This looks funky:

https://www.instagram.com/p/DQEKJC3EbnY/?utm_source=ig_web_copy_link

[3staxontheradio]

Another thought I want to look into: I think with longer runners we also get more latitude to place the merges in line with each other and can make the section 1/collector length equal without needing to resort to a funky bent section 1 design. its going to come down to packaging.

Also at least on the S52 those Kromer Kraft headers sound pretty awesome: https://www.youtube.com/watch?v=S4JdnVzNv0Y