S54 Header Design talk

[3staxontheradio]

That's super interesting. In that sim at least, its not a huge difference between taper and step but the general appearance seems to show the tapered runner having a stronger wave signal but less spread out than the steps. Is that how you read it too? Your runner is shorter than the pipemax calced one so not surprised its looking a little more top end heavy? Could you try with a 480mm runner and also can you show the torque across RPM for a full idle to maybe 8500rpm sweep?

Also can you help me understand that valve pressure chart? Is it at a specific RPM? Seeing whats happening before EVO is cool

On the power trace the 38mm straight runner just outperforms the taper. The stepped runner is at least holding more power on the far left but its only one point on the sim and maybe not enough to show a trend.

[digger]

That's super interesting. In that sim at least, its not a huge difference between taper and step but the general appearance seems to show the tapered runner having a stronger wave signal but less spread out than the steps. Is that how you read it too? Your runner is shorter than the pipemax calced one so not surprised its looking a little more top end heavy? Could you try with a 480mm runner and also can you show the torque across RPM for a full idle to maybe 8500rpm sweep?

Also can you help me understand that valve pressure chart? Is it at a specific RPM? Seeing whats happening before EVO is cool

On the power trace the 38mm straight runner just outperforms the taper. The stepped runner is at least holding more power on the far left but its only one point on the sim and maybe not enough to show a trend.

i just used 450 mm as a round number so each step was 150mm, i can change the length. you can see that the changing the pipe size also needs a length change to ensure the scavenge wave arrives at the same "time"

Its probably not much value to do full rpm sweep with open primaries (no collector and not realistic) and haven't set up variable cam timing. i was just using the model to see the impact on the wave at an rpm when the pipe is in tune

Also it appears 35 mm is likely too small overall for this hp level, 38 mm ID is probably better as a minimum so i might add 44 mm (ID) and see if a step 38-41-44 is "better" than a single pipe size once the size is more appropriate

The taper does produce a slightly better suction wave with less pumping loss. I don't understand why the step header having the higher peak pressure, maybe it is something to do with a difference in residual pressures at EVO

It is also worth noting that there is approx less than a 1% change in VE, so the power differences for the most part are pumping loss differences between BDC and TDC as they are all scavenging well enough during overlap (0.5-0.6 bar "suction" or about 200" H20 ). Adding a collector and other piping is likely to degrade the scavenging at this rpm

The model is also rough if you can tell me the following to double check i could improve it

Inlet port
- min CSA and distance of the min CSA from valve (if not the throat)
- port entry area / diameter equivalent

Exhaust port
- min CSA and distance of the min CSA from valve (if not the throat)
- port exit area / diameter equivalent

Inlet manifold
- area/diameter at the head
- overall length of the cast throttle body section
- overall length from head flange edge of trumpet
- trumpet inlet ID

[3staxontheradio]

Cool. I can get all of those measurements to you soon

[digger]

changing to 480mm and 38, 41 and 44mm



pressure


480 vs 450 41mm

[George Hill]

Hey digger what is this software you are using? These simulations are still an open primary with no collector?

Can you explain the pressure graph? Is this what the back of the valve is seeing at specific crank angles? How do we apply this to real world applications?

Would you be able to run my Kromer Kraft's through to compare? 3staxontheradio initial pipe max simulations output a similar spec and curious to see how it compares in your software.

#6 primary has 2 steps - 44mm to 47.5mm to 49.5mm od (roughly), lengths are 178mm, 229 and 203mm for 610mm total (roughly).

[digger]

Hey digger what is this software you are using? These simulations are still an open primary with no collector?

Can you explain the pressure graph? Is this what the back of the valve is seeing at specific crank angles? How do we apply this to real world applications?

Would you be able to run my Kromer Kraft's through to compare? 3staxontheradio initial pipe max simulations output a similar spec and curious to see how it compares in your software.

#6 primary has 2 steps - 44mm to 47.5mm to 49.5mm od (roughly), lengths are 178mm, 229 and 203mm for 610mm total (roughly).

this is ENGMOD4T 1-D simulation software.

That is the pressure in the exhaust port about 1" downstream of valve at each crank angle.

This is for the open header as the simplest kind of exhaust. The only usefulness from this simple example is understanding some basics about what is going on, and that you can see that the exhaust length is fairly close to optimal for 8000rpm as shown by the negative exhaust port pressure prior to IVO and post EVC which helps with scavenging. Also the gain or loss between diameter sizes at 8000rpm is mainly due to pumping losses (the larger pipe reduces the in cylinder pressure faster) rather than scavenging differences but this is not always the case.

The size written in the graphs is the ID add 3mm for the pipe size

There are endless options to simulate and arrive at concepts for testing

[George Hill]

digger appreciate the explanation. In the 2nd graph what is Pcyl(1) vs Pex(1)?

[3staxontheradio]

I have another cylinder head scan in progress. Hopefully I get cleaner ports this time and then I'll get those measurements. Does anyone know the stock or CLS box trumpet length and the depth of the S54 ITBs? I'm about to order a karbonius box which I'll also scan and upload with this stuff.

[digger]

digger appreciate the explanation. In the 2nd graph what is Pcyl(1) vs Pex(1)?

pressure in the cylinder vs exhaust port. The (1) is the location being cylinder number 1 and port number 1

there is also inlet pressure that needs to be considered but i tried to avoid cluttering

[Bry5on]

I have another cylinder head scan in progress. Hopefully I get cleaner ports this time and then I'll get those measurements. Does anyone know the stock or CLS box trumpet length and the depth of the S54 ITBs? I'm about to order a karbonius box which I'll also scan and upload with this stuff.

I had the trumpet number approximated at some point - from my measurements, the CSL trumpets appear to be identical to the non-CSL trumpets. I never did scan them though so #s could be off

[3staxontheradio]

Finally finished the rescan of the head with a focus on ports. This still isn't as clean as I would like but i have way too much time into it already. Roughly 1600 photos. I also had my workstation down for 3 days with hardware issues + corrupted windows system files but It seems to be all good.

The current pipemax config is around stock cams with the common bolts on (Full exhaust, CSL box, good tune, etc) I'm ball parking it at 365 crank horsepower. This directly impacts runner diameters. Anyone have thoughts on that for an estimate of crank horsepower?

I also want to do another configuration with a bigger cam set. What are most people running? I think either the Schrick 288/280 12.5mm lift set or the catcams 288/280 12.5/12.75 set make sense. Either of these sets probably looking at adding 20whp? Again I need to get the power output dialed in to get the runner sized.

digger I'll try to get all of those measurements today.


Also my CSL box shows up today so I'll get that scanned over the weekend.

[digger]

Another way to characterise the ports is silicone moulding, then slice it up at various locations for direct measurement or using grid paper and you can plot cross section as a function of position which is also useful if you happen to want to do some porting to target the restricted areas / complex shape areas but probably not within the scope

The CSL was 360bhp so more aggressive cams, exhaust and what not i would think 380hp (unopened bottom end and unported head) as an upper limit on a healthy engine.

[JakeM346]

Another way to characterise the ports is silicone moulding, then slice it up at various locations for direct measurement or using grid paper and you can plot cross section as a function of position which is also useful if you happen to want to do some porting to target the restricted areas / complex shape areas but probably not within the scope

The CSL was 360bhp so more aggressive cams, exhaust and what not i would think 380hp (unopened bottom end and unported head) as an upper limit on a healthy engine.

Exhaust Vales were lighter. otherwise exhaust is the same except than lighter tubing. Practically all of the power comes from airbox and cams

[m3racer]

Great thread! I have been looking at designing and building some S54 headers for my race car and this is all good info. My use case is different: a fully built high compression with high duration/lift cams, which can hit ~400whp, but only looking for power in the 6000-8500 RPM range. I have found that the engine is limited to around 400whp due to the headers, and have tried a few different versions: BMW Euro catless, ARH 6-1, eBay, Epic and SuperSprint. So far the Epics were the best, but Randy doesn’t make them anymore. I would really like to try BMW’s P54 race headers, but have not found any for a reasonable cost. Anyone have any?


I measured the SuperSprints if you still need that data. It was their 044001 part number. I think they are too small for my application.

I worked with Burns Stainless on a design based on their vast empirical testing and design knowledge, and just need to make them. It was a 6-2-1 design with 24in long primaries (3 stepped 8in long sections, 1.750, 1.875, 2.000 OD, 16ga pipe), then secondaries 2-1/4" as short as possible, and then a collector to 3.5 OD exhaust.

My PipeMax output came out a little different that this design, and I might try something closer to it:

--- 3-Step Tri-Y Headers --- ( possibly higher RPM HP than 2-Step Headers )
Primary Y Segment 1st Step Diameter = 1.909 Length= 7.424 to 7.659 inches
Primary Y Segment 2nd Step Diameter = 2.034 Length= 7.424 to 7.659 inches
Secondary Y Segment 3rd Step Diameter = 2.284 Length= 7.424 to 7.659 inches
Primary + Secondary Segment's Total Best Length = 22.273 -to- 22.980 inches

I was just going to have a fabricator make them, but if DMLS cost is reasonable now I can try that. This Celeritech company looks good too.

[3staxontheradio]

Can we compare and swap pipemax configs? Also do you ahve the measurements you took from the supersprint V1s? I have some euro CSL headers being loaned to me to scan/measure as another reference point.

Also very curious why burns had the secondaries as short as possible.