S54 Header Design talk

[ZiMMie]

Keep in mind these KK's are E36 fitment, they do NOT fit an E46 without modification. Specifically the engine support arm, which is easy, but #4 primary is into the drug bin structure, I had to clearance the chassis right there. They will also not work with the factory A/C lines. When I reinstall them I am going to pull the whole front end off the car so I have access on (3) sides.



Why do all of this? Because they DO sound really good. I am going to test the original 'Y' pipe and and a more traditional dual pipe setup but if the dual kills the sound I'm going back to the 'Y' even at the detriment of the torque curve which is big.

https://youtu.be/juk-elMFBZY

It would be best to also get a scan of the E46 chassis so the two chassis models can be merged. That way, it’ll be much easier to create something that fits both.

[George Hill]

It would be best to also get a scan of the E46 chassis so the two chassis models can be merged. That way, it’ll be much easier to create something that fits both.

100% and I'm pretty sure Alec has all of that I more meant for people looking on that may not know.

[Cronenberged]

If anyone wants to mail them to me I would be happy to scan them.

Would be great if we knew someone who had them. I believe Epic’s headers were the closest models I’ve seen to them.

[3staxontheradio]

Yeah I'm planning to finally scan my e36 m3 tomorrow when it's warmer in the garage, all prepped and ready. I have an extra e46 chassis around I can also scan but it's an m54 car so I'm not sure header flange alignment will be quite right for an s54?

I do plan to try some original layouts for tubes, I'm just freeform laying them out with splines so it's all very loose. Longer primaries do on fact make it possible to get the merges in line. I think I can also get them lined up more neatly than this but I was starting with the kromer kraft locations.

It seems like getting the merges aligned needs to start with a bank 1 design that has as few bends as possible, trying to get the merge as far back as possible. Then place the merge for bank 2 and figure out how to reasonably route them.

I also did some pricing estimates on craftcloud for printing individual primaries that slip fit into merges and at this length it's looking like $180ish per primary including a flange. Thats for 1.6mm wall at about at about 550mm length.

[Bimma360]

I was really curious about how expensive 3d printed headers would be. I made a super fast sketch of one of the banks of the CSL header, made it without flanges (I figure flanges could be traditionally cut to save money), set it to 1.5mm wall thickness and uploaded it on craftcloud. If its made as a single piece its kinda pricey. $1200 for one bank or $2400 for a set. This is because its over the size of almost everyone's DMLS machines so only one company could quote it. Check out pricing when split into two pieces:


Longer runners, possibly slightly thicker wall thickness, etc is going to involve more material but this is way cheaper than I thought it would be. I think for people willing to weld them in peices and mix with traditionally fabbed flanges, vbands, etc. the cost may end up around $1000.

I've read through this thread twice already and something that occurs to me is that "we" should work towards reproducing the euro headers through 3D printing, in addition to the work being done in this thread. B/c the design process is already done and tested in the real world, reproductions can move to manufacturing a lot quicker, and everything learned from that could be applied to producing the other header design. While I am not entirely sold on spending $1500+ for used euro headers, I would actually be willing to spend $2K on 3D printed versions. From what I understand, there is potential for these to be lighter than stock. Yes, I am volunteering to be the guinea pig.

[3staxontheradio]

I'm happy to scan a euro header bank 1 and add the scan to the github but I'm not personally interested in building off of that design. Megan racing headers are a passable copy and I would 10 out of 10 times spend $400 on them instead of $2000 on euro headers. The most significant difference I could find between Megan and the CSL bank 2 I scanned were the merges and the Megans probably need some tweaking so the flanges/primaries are well aligned with the exhaust ports. I think euro headers made a ton of sense when they were cheap but less so at these prices.

[Bry5on]

I have a scan of an e46 engine bay with an s54 loose in it, taken at George Hill shop. George, do you have the link handy by chance?

[George Hill]

Bry5on unfortunately I don't have that scan.

[Bry5on]

Bry5on unfortunately I don't have that scan.

Looks like I don’t have that one in the cloud. Well here’s the e46 engine bay by itself at least: https://s.digital3dcloud.com/space/f...d-03e18fed4c66

[Bry5on]

Would be great if we knew someone who had them. I believe Epic’s headers were the closest models I’ve seen to them.

This layout is definitely the way to make it package. I tried a few quick spline layouts and only this one really made sense to me.

[Cronenberged]

This layout is definitely the way to make it package. I tried a few quick spline layouts and only this one really made sense to me.

They are also equal length. My understanding is they don’t offer good low end torque but that’s something could be fixed by stepping them like SSV1.

[Bimma360]

I'm happy to scan a euro header bank 1 and add the scan to the github but I'm not personally interested in building off of that design. Megan racing headers are a passable copy and I would 10 out of 10 times spend $400 on them instead of $2000 on euro headers. The most significant difference I could find between Megan and the CSL bank 2 I scanned were the merges and the Megans probably need some tweaking so the flanges/primaries are well aligned with the exhaust ports. I think euro headers made a ton of sense when they were cheap but less so at these prices.

Completely understandable. Ironically enough, based on your Megan feedback is why they are completely off the table for me. I just don't have any tolerance for that stuff anymore. Anyway, if you do end up with a scan of the other bank I would have no problem being the 3D printing ginuea pig, depending on where I am at with all of this at the time.

[3staxontheradio]

This was a pretty challenging scan for my setup, especially on quickjacks. The camera setup is to bulky to get into most positions and then there are very few surfaces that can be clearly seen from above and below that can be used to tie together image from both sides. Also shooting through the top into a small gap to try to see the top of the engine mounts, etc most of the light is blocked.

So far pretty happy with the result. I think the two sets of images stitched relatively well:

[Obioban]

A couple of years ago the silencer community largely moved to metal 3D print printing, with with hugely beneficial results. surface texturing of the metals gave them a surprising amount of extra benefit. Is there anything to gain for us here, similarly?

[digger]

IMO you'd be looking to flex hone the ID to smoothen them up