Making a 3D printed Inconel turbo manifold - Building a 13B BMW i8 : EP3 

Inconel. The “order”was part of the demo to show what the typical buyer might choose. The manifold design reviewed is not the same as in the demo or checkout and you can see the manufacturer on the box is also different. Prices for the reviewed manifold are in the “i8 swap” column of the cost breakdown.

@@CouchBuilt Ah okay just confused me slightly and the colors in the pictures for the price comparison also made my brain think it was different. That's so cool!

Likewise. Way impressive and some knowledge 'bout 3D printing is mega appreciated ❣

I think the technology and price has finally reached a point where it can be a game changer to the garage builder or smaller fabrication shop looking to level up their builds. Nothing to hide or gatekeep. I wish it was at this point when I was in the fabrication game 10-20 years ago.

@@CouchBuiltagreed and only going to get cheaper going forward. Also CAD packages for enthusiasts like F360 and Solidworks for Makers edition will allow people to get right into this space.

Glad you liked it. My skills have a ways to go, particularly on workflow. I’ve learned a ton on this project but amassed so much technical debt that after I committed to sending the print files out, I exported them all as STEP files to free up the design timeline. Let’s just say simple changes were orders of magnitude too long to process toward the end.

Came to say the same thing - I learnt a *ton* from just that section alone.. I struggled doing a similar modelling task (by struggled I mean I gave up and went back to beating round tube into square flanges like some kind of caveman), and now I know how easy it should have been..

Not quite new tech, but newly affordable Id say.

Good insights, but bold of you to assume my bench is flat. There’s no gap with a straight edge.

@@CouchBuilt also bold of you to think a belt sander can give you a flat surface

Awesome! Thank you!

Good catch. I had to go back and look. Seems the stl from grabcad was scaled to inches and I assumed it was in mm like most sane people use 😅. That changes the price to $80. I didn’t think anything was amiss as I had reasonably similar size aluminum part that ran me $25.

I would argue your statement should be worded “may not”. If even pressure is applied you can get an impressively flat surface. If you’re wildly varying the angle of attack then you’ll end up with a banana. The purpose of this is to show the accessibility of a technology with post processing available to the average builder. Yes, milling would be more precise, but it’s not the only valid method. In a past life I built close to 1000 welded turbo manifolds with belt sanded flanges. Sealing was never a complaint. Wrench access to bolts was the usual gripe, but I’ll take that as an L with things I’ve learned over the years.

Came here to make a similar comment... especially with such a large part and small belt width and high belt speed. As long as he checks the flatness on a true flat surface, and uses dye removal to prove flatness, it can be done... but is time consuming.

Next video I’ll run a feeler gauge across this manifold and the intake manifold. That one was done the same way. Saying flat is like asking how long is a string. As long as an acceptable tolerance results, the method is moot.

It seems to be all about the details of how exactly it is done in my experience. The only conclusion I’ve reached is that results seem to vary wildly depending on the person who did it 😂

110%

The 13B model was by a guy on Instagram @mitchabbott_ I bought a STEP file from him and he also sells the STL. The DQ500 scan was from Domiworks. Neither were super cheap, but both were 100% worth the investment to ensure fitment before buying real parts.

304 and 321 are the usual alloys for header fabrication with 321 having better heat properties. 316 is comparable to 304 in heat resistance, but is more corrosion resistant.

One thing to keep in mind is that it’s an automated process and the prices shown are the lowest for a particular configuration. They might be through different vendors even though the material is the same. There might be limitations or ideal size thresholds for a particular manufacturers setup.

I have an acquaintance thats been involved in some turbine secondary machine work for such applications. It’s not an apples to apples comparison as those are usually moving parts at higher temperatures, pretty extreme rpm, and lightened to the max. This is just a static hunk of metal that’s twice as thick as it probably should be. I don’t foresee any issues, but time will tell.

@@CouchBuilt I will be watching for sure.

The problem, if you can call it that, is that "3D printing", or "additive manufacturing" is so diverse both in materials as in methods, that overall generic statements like that are meaningless. You can 3D print many different plastics, metals/alloys, some ceramics and even composites. The various methods use resistance heating, laser beam heating, electron beam heating (in vacuum) or electric arc heating. Some printing methods need debinding and sintering. Some are ready directly. Some benefit from a post heat treatment. You can print with powder, welding wire, filament, composite filament, concrete, 2-component materials and more... If this were Inconel with powder bed SLS printing (selective laser sintering) it could probably benefit from a post print heat treatment. If this was powder bed SLM (selective laser melting) it likely wouldn't need that. If this was made by binder jetting (metal powder printing with a wax-like binder) it would need debinding (removing the low melting binding agent) and sintering. Sintering will inherently stress-relief at the same time. So the short version is: It's complicated...

In this case it was DMLS.

@@CouchBuilt If that's on a printer from EOS, it's German for _Direkt Metall Laser Schmelzen_ so SLM, selective laser melting. SLS is older tech, resulting in slightly more porous prints.

The scans were done with a Shining 3D Einstar. I’ll probably show a little more on that in a future video. I have to make some exhaust stuff for someones E32 BMW that’s lurking in the background.

To make all this work with the factory hybrid components. I need to emulate a bunch of digital and analog signals that will be generated by my own hardware. I’m sure the scope will get some air time when I cover that stuff.

Probably not much. Still 10lbs lighter than stock twins from all the figures Ive seen.

They don’t need to be, especially on a turbo application where there’s positive pressure in the exhaust. You’re not relying on the scavenging effect for cylinder evacuation. I’ll always prioritize a shorter runner and better packaging over equal length. If you’re in a limited race series and shooting for absolute maximum power, or tuning the lengths to achieve a particular sound, then by all means the effort for equal length can matter

It’s a 2021 Asus G14 with 40gb RAM. It only performs like this when plugged in. I’d like to upgrade, but next real notable step up always specs out at like a $5k workstation, which I can’t justify. I need to buy a DQ500 DSG instead 😂

I believe .5mm was generic minimum recommended wall thickness for a DMLS metal print, so that should be doable.

@@CouchBuilt thanks for the info :)

They are, but I tried contacting the end manufacturer independently on two occasions, they were both more expensive. There’s also the convenience of one purchase point for the full array of vendors. I used 4 different sources for my project on one purchase.

Just wait until I get into the hybrid integration hell.

Fusion360. The same tools are available in most(all?) CAD packages.

Because this is going in a BMW i8….

The bench is not flat. The flange is .0015” flat. Covered in my next video.

Indeed. I explain a bit more about that in episode 1.

I clearly show what the i8 manifold cost in inconel. The whole point of the demo was to show that 316 is within the average budget. My manifold is inconel from KleRo in Germany.

@@CouchBuiltdid you need the extra heat or strength from inconel? What made it worth the extra cost to you? NA manifolds I could see going with stainless all day

@MPbdy I originally planned to use stainless and I wouldn’t hesitate to do so. Inconel is simple a better material and when I saw the price, it wasn’t an outrageous jump. The amount that I saved with the motor I bought also played a part in going this route.

What tolerance do you consider flat for a manifold sealing flange?

@@CouchBuilt +/1 1 thou 0.025 micron. unless your boosting only 5 psi in that case, your sanding will do.

Considering the general rule of thumb for a cylinder head is .001” allowable per cylinder with the expectation to seal 1000+psi of combustion pressure from coolant and oil passages, the .0015” that this measures will be just fine for even a godawful 3-4:1 back pressure ratio maxing out the MAP sensor.