S54 Header Design talk

[George Hill]

I can't recall for sure, but this one. I'm not sure I am going to use it when I put the KKs back on, maybe something custom and valved, IDK yet.

[Casa de Mesa]

I would expect if that is your metric for this project then its probably not going to be something you switch too. There are so many variables in regards to exhaust and what they deliver that just a total hp curve vs V1s or whatever is going to be hard to blow them out of the water.

I'm installing the Kromers for no other reason than I like the way they sound. They absolutely lose torque over stock manifolds and don't pick up a ton over V1s (if any). They don't fit well and are very custom. So my interest is very much different than yours and that is ok, no shade as the kids say.

I'm really enjoying this project as a thought exercise and seeing these guys smarter than myself work through it. If a commercial product materializes that works for my setup, even better. If not, I hope they achieve their goals and maybe that knowledge will be something I can recall back upon in the future.

This is the camp I'm in. I'm also closely following this in hopes of spectacular aural results. We all have different perspectives on S54 exhaust, but almost unanimously agree the factory left some room on the table.

Beyond that and probably even more so, I'm enthralled with the idea of "grassroots" efforts in this community continuing to progress this chassis some 30+ years after BMW first penned the design. It's kinda hard to call it grassroots with the folks working on it...

Board certified layman question here: Are there any durability concerns regarding DMLS / 3D-printed metal holding up to heat cycles and vibrations of an S54? Or any other motor for that matter?

I know there have been plenty of applications that have seen 3D printed exhaust components work (Koenigsegg, Inconel headers in F1/motorsports, Kline Innovation and 1016 Industries on cars like the Ferrari SF90 and 911 GT3, etc.) but I'm also drawing the through-line that these are some deep-pocket applications.

[Cronenberged]

This is the camp I'm in. I'm also closely following this in hopes of spectacular aural results. We all have different perspectives on S54 exhaust, but almost unanimously agree the factory left some room on the table.

Beyond that and probably even more so, I'm enthralled with the idea of "grassroots" efforts in this community continuing to progress this chassis some 30+ years after BMW first penned the design. It's kinda hard to call it grassroots with the folks working on it...

Board certified layman question here: Are there any durability concerns regarding DMLS / 3D-printed metal holding up to heat cycles and vibrations of an S54? Or any other motor for that matter?

I know there have been plenty of applications that have seen 3D printed exhaust components work (Koenigsegg, Inconel headers in F1/motorsports, Kline Innovation and 1016 Industries on cars like the Ferrari SF90 and 911 GT3, etc.) but I'm also drawing the through-line that these are some deep-pocket applications.

you forgot Czinger 🙃

[3staxontheradio]

This is the camp I'm in. I'm also closely following this in hopes of spectacular aural results. We all have different perspectives on S54 exhaust, but almost unanimously agree the factory left some room on the table.

Beyond that and probably even more so, I'm enthralled with the idea of "grassroots" efforts in this community continuing to progress this chassis some 30+ years after BMW first penned the design. It's kinda hard to call it grassroots with the folks working on it...

Board certified layman question here: Are there any durability concerns regarding DMLS / 3D-printed metal holding up to heat cycles and vibrations of an S54? Or any other motor for that matter?

I know there have been plenty of applications that have seen 3D printed exhaust components work (Koenigsegg, Inconel headers in F1/motorsports, Kline Innovation and 1016 Industries on cars like the Ferrari SF90 and 911 GT3, etc.) but I'm also drawing the through-line that these are some deep-pocket applications.

I have wondered that too. I think it's a reasonable question. Someone smarter than me ( Chrisyphus ) can probably look through data sheets and get a good sense of that or run some FEA. I'm planning to use 1.65 mm wall/16 gauge. People frequently go thinner for traditionally fabbed headers so I don't feel like it's an unrealistic starting place. If there are areas that see higher stresses we can always add wall thickness or support ribs in those specific places since it doesn't need to be a uniform wall thickness. We can also add webbing between primaries to connect them into one structural piece where possible. I Have seen other 16 gauge 316L DMLS exhaust parts but not many people doing headers. The zpipe and lots of y pipes and merges have been surviving just fine. They won't see the super high temps that headers can get to though.

[Bry5on]

I’ve never done any exhausts but I’ve been part of 3D printed inconel for turbomachinery and actual combustion chambers. Generally the printed part is similarly strong and dense after a heat treat/HIP cycle. I think stainless gets stronger but more brittle if it’s not heat treated, so we could see cracking under high thermal strain locations. I understand titanium parts have very similar strengths/ductility to normally produced Ti materials, just to throw something more wild into the mix.

[digger]

fabricated parts normally crack at the welds as that is the weak point , with 3D printing you can in theory eliminate welds completely or in the very least eliminate them from the highest stress (complex geometry) areas, so as long as any re-entrant internal corners have a nice radius and thickness doesn't drastically change within a short distance (to minimize temp gradients) i wouldn't be too concerned for a typical NA engine.

[Casa de Mesa]

fabricated parts normally crack at the welds as that is the weak point , with 3D printing you can in theory eliminate welds completely or in the very least eliminate them from the highest stress (complex geometry) areas, so as long as any re-entrant internal corners have a nice radius and thickness doesn't drastically change within a short distance (to minimize temp gradients) i wouldn't be too concerned for a typical NA engine.

Appreciate this insight.

I've cracked more than my fair share of short and long tube headers in NA applications (not on a BMW), hence the reason I asked. And maybe not-so-ironically, they all cracked at weld points.

[digger]

Appreciate this insight.

I've cracked more than my fair share of short and long tube headers in NA applications (not on a BMW), hence the reason I asked. And maybe not-so-ironically, they all cracked at weld points.

yeah well that isn't surprising as there isn't anything good about welds other than being cheap and relatively easy for mass production. Some of the reasons they are bad for durability:

- usually introduce sharp notches
- introduces high residual tensile stresses
- introduces micro and macroscopic flaws into the material
- changes the microstructure of the material
- makes it more susceptible to stress corrosion cracking

[ApexPothole]

And maybe not-so-ironically, they all cracked at weld points.

With stainless steels part of the alloy composition that makes the steel stainless is chromium. When stainless is welded if conditions aren't optimal the chromium can react with carbon in either the base metal or contamination like oils or soot to form chromium carbides, which rob the weld area of its ability to be stainless and cause it to be more brittle than the base metal.

If stainless isn't back purged with an inert gas it'll get sugaring, which I understand to be chromium oxide, which also reduces ductility, corrosion resistance, and flow.

I've never cut apart really nice headers but the ones I have cut up didn't have full penetration welds, probably cause they're harder to make pretty. From what I can tell it looks like the various eBay and my E30 schmiedmann headers I've had apart appear to be just welded autogenously with no filler rod, and then remelted using a pulse mode to get the stacked dime look. I believe this is kind of bad for strength because iirc filler has a lower carbon content compared to the base 300 series alloy, so adding filler decreases the carbon content in the area of the weld which offsets some of that carbide formation, and somewhat cools the weld to help with the same. Furthermore, remelting a welded joint increases carbide formation as well.

I am not a metallurgist though, so my understanding of this is likely to be flawed.

I have never welded anything DMLS printed (though I hope to change that soon) so I can't weigh in on how they perform, however I have been in the loop with people studying the process at a post graduate level and from what I have learned from them I am confident that the machines that perform this process will have a better handle on weld quality than I ever will melting them together with a Tig torch.

[freebird]

Came here to check on things. Great thread! My current setup is Megan headers, to a custom exhaust. I'm currently fighting a knock problem between 6300-7300 rpm when I have a lot of valve overlap (VANOS tuning), and am pretty sure the best fix is a different header and/or exhaust design to change the scavanging. So, this effort peaks my interest!

[EthanolTurbo]

Came here to check on things. Great thread! My current setup is Megan headers, to a custom exhaust. I'm currently fighting a knock problem between 6300-7300 rpm when I have a lot of valve overlap (VANOS tuning), and am pretty sure the best fix is a different header and/or exhaust design to change the scavanging. So, this effort peaks my interest!

How do you know it's not an octane issue? An S54 on low quality pump gas like 91 will pull a ton of timing and have way more knock than one with appropriate octane.

[freebird]

How do you know it's not an octane issue? An S54 on low quality pump gas like 91 will pull a ton of timing and have way more knock than one with appropriate octane.

Fair question. 98 octane, or significantly reduced timing, made some but not much difference. The only thing that helped was to remove a ton of valve overlap. FYI, I don't have EGTs, would love to see that data (This is all on a MaxxECU standalone.)

[George Hill]

Fair question. 98 octane, or significantly reduced timing, made some but not much difference. The only thing that helped was to remove a ton of valve overlap. FYI, I don't have EGTs, would love to see that data (This is all on a MaxxECU standalone.)

Who is doing your tuning? I am considering swapping to a Maxx as well and curious to see what you find. There are a bunch of S54s running Maxx and likely a similar exhaust so I'm a bit surprised to hear of this issue.

[freebird]

I had Kassel do my initial tuning. I've done a lot since. To be clear, I don't think the Megan header is the core issue. I have a 2-1 merge located too far back, around 44" where the OEM xover is. Great for midrange I think, but my math and the valve overlap and knock symptoms points to exhaust reversion in the rpm zone where I'm getting knock.

I'm pretty happy with the Maxx. The new Race Gen2 looks like a real nice product, all around better than the Gen 1 Race which is what I have. There are a few nit picks, happy to go into the details if you want.

[George Hill]

Sent you a PM