heinzboehmer's 2002 Topaz 6MT Coupe

[Bry5on]

Make sure you put the carbon in compression if you want it to take load. You’ll need a shoulder area under the carbon on both sides as it will need to be ‘pinched’ - this means your machined bosses should certainly be thinner than the carbon sheet. Frankly, that sheet should probably just be aluminum so you don’t need to worry about bearing area on the carbon. But still make the boss shorter than the thickness of the sheet for sure. I can sketch this for you some time in person if you want the joint to perform as well as it can.

[heinzboehmer]

Make sure you put the carbon in compression if you want it to take load. You’ll need a shoulder area under the carbon on both sides as it will need to be ‘pinched’ - this means your machined bosses should certainly be thinner than the carbon sheet. Frankly, that sheet should probably just be aluminum so you don’t need to worry about bearing area on the carbon. But still make the boss shorter than the thickness of the sheet for sure. I can sketch this for you some time in person if you want the joint to perform as well as it can.

Yep! The bosses are slightly thinner. Once again picked some random number, it's one of the things on the TODO list that I still need to figure out.

Open to making it out of Al as well, just picked carbon cause of the weight.

And yes please! In person design review would be awesome.

[D-O]

"Just went with what the eyecrometer said looked good."

I am using that one. Far superior to the old "calibrated eyeball."

[bigjae46]

What's your timing? I have 2 more roofs, hood and a mold, and a dashboard mold, and front thrust plate for George Hill to do - 3-4 more weeks for me?

I have about 20lbs of casting resin and it has a limited shelf life. So far, I plan to a compression mold for race seat floor mounts, jack pad, and maybe seat mounts. I should have plenty of material left to do your mount. If you want to give it a try then all I ask is to get rid of the ribs on the backside...I'd have to fill all of that stuff in which would take forever. Estimate about 2-4 weeks to get done. Again, it's going to cost less than $50 in materials...if that.

It would be an honor to contribute something to this build!

[heinzboehmer]

What's your timing? I have 2 more roofs, hood and a mold, and a dashboard mold, and front thrust plate for George Hill to do - 3-4 more weeks for me?

I have about 20lbs of casting resin and it has a limited shelf life. So far, I plan to a compression mold for race seat floor mounts, jack pad, and maybe seat mounts. I should have plenty of material left to do your mount. If you want to give it a try then all I ask is to get rid of the ribs on the backside...I'd have to fill all of that stuff in which would take forever. Estimate about 2-4 weeks to get done. Again, it's going to cost less than $50 in materials...if that.

It would be an honor to contribute something to this build!

That timing should work. I can print you a piece and send you that.

To be completely honest, I'm not sure that forged carbon is gonna be the best material for this application. This piece will see a decent amount of load and the random nature of the forged carbon makes me a bit uneasy. Would gladly hang the piece on the garage wall if you're still interested in the practice!

[heinzboehmer]

Also, been learning as much as I can about designing this piece for strength, while also having it be light and relatively easy to manufacture. Lots of learning to go, but I'll get there. I'm very much the wrong kind of engineer to be designing these kinds of things, but luckily, I have a decent chunk of mechanical engineer friends (a lot of them automotive as well)!

There are a few things that are of concern to me right now (there's likely more points of concern, but I have yet to identify them...):

1. The ribs are placed randomly. I tried my best to put them in the load paths, but the corners of the piece don't line up with them, so there's a bit of a compromise there. I don't know what the implications of this are, need to figure that out.
2. The surface area for the CF to bond to is decreased because of said ribs. Again, not sure what the full implications of this are, maybe it's fine? But I will sleep better when I know what analysis to do to make sure it's good enough.
3. My fillets and bosses are picked kinda randomly. I just guessed at wall thicknesses and stuff for them in the version posted above. I've gotten some feedback from friends and they look much better now, but I need to keep these constraints in mind going forward.

Anyway, in an attempt to address the above concerns, I quickly drew up a couple (wacky) alternative ideas for the mount. Figured I'd document them even though I might not implement either.

First up, a split machined version of the mount:





The idea behind this is that I can keep the full surface for the adhesive to bond to on both the top and the bottom. The two halves would have to be machined with no threads, bonded together along the ribs, post processed (tap holes, clean up any excess adhesive) and then get bonded to the CF sheet.

The extra material on the bottom only adds ~50 g, which sound like a worthy tradeoff. The problem is that I don't actually know how splitting this thing in half and then bonding it back together will affect its strength. Again, more learning required.

Second alternative is to make the entire thing out of sheet metal. Would look something like this:



This design would also require two end caps to box in the sides, along with a bunch of locating tabs everywhere so that it's easy to fixture for welding. Some sheet metal ribs inside along the load paths would probably not be a bad idea either.

Big issue with this is that even without the end caps and ribs, the design already weighs roughly the same as the machined + CF version. Making it this way would really only help with cost and I'm not exactly trying to optimize for that factor with this project (especially after seeing that the CNC'd quote was <$500). Assembly would also be harder as it requires knowing how to weld well instead of just smearing a bunch of adhesive on.

Anyway, alternative #1 is appealing, but more thinking is required before any decisions are made.

[Obioban]

^that sheet metal version doesn't look like it's going to fit in a brake
(not manufacturable)

[heinzboehmer]

^that sheet metal version doesn't look like it's going to fit in a brake
(not manufacturable)

Oh it for sure won't. Would likely have to leave one of the bends halfway and finish it off on the bench.

[bigjae46]

That timing should work. I can print you a piece and send you that.

To be completely honest, I'm not sure that forged carbon is gonna be the best material for this application. This piece will see a decent amount of load and the random nature of the forged carbon makes me a bit uneasy. Would gladly hang the piece on the garage wall if you're still interested in the practice!

As long as there are no voids in the part and there is a high carbon:resin ratio - 60% - forged carbon will be stronger than the sheet metal it will be mounted to. Forged carbon would probably have better performance given the randomness of the fiber orientation.

The part won’t elongate or stretch which I’m not sure is an issue here. Carbon has poor elongation and strength after the carbon yields. Also is not as tough as steel.

[heinzboehmer]

As long as there are no voids in the part and there is a high carbon:resin ratio - 60% - forged carbon will be stronger than the sheet metal it will be mounted to. Forged carbon would probably have better performance given the randomness of the fiber orientation.

The part won’t elongate or stretch which I’m not sure is an issue here. Carbon has poor elongation and strength after the carbon yields. Also is not as tough as steel.

Interesting.

There's two very distinct load paths for this piece though, so wouldn't the random fiber orientation be a detriment? Compared to orienting the fibers so the part is stronger in those directions.

[heinzboehmer]

More feedback from people who know what they're doing incorporated:


Pocket is designed to be machined out with a 6mm diameter ball end mill. Fillets are 5mm radius.

With the new changes, there is a ton of material on the bottom surface for the CF sheet to bond to. CAD says the bonding surface is 48% of the entire bottom surface area of the part (3942 mm2 vs 8174 mm2). Feel a lot more comfortable bonding it like that.

Weight is up slightly, but still hovering around the 1-1.1 lb mark.

[heinzboehmer]

Designed a super quick fit test piece to check clearances and angles of everything:



I put the outer surfaces of the dark blue test piece right where I thought they would interfere with the HVAC components and gave myself juust enough clearance for tools. Here's everything on the car:



You can tell I measure/transcribed incorrectly because I had to remove the blower motor cover to test

Also, not gonna lie, swapping the blower motor is gonna suck with this in place. Totally doable (after unbolting the dark blue piece), but its gonna be really tight.

Next came the most exciting moment of this whole project. This is actually gonna work!



Also, test fit revealed a bit of extra room that I didn't know I had, which is great. It allows me to adjust the position of the fasteners slightly and put the rear fastener right in line with the load paths.



(the two circles along the dotted lines represent where the fasteners for the brace will be)

As you can see, I also got rid of that top rib, as it really wasn't doing much. The two diagonal ribs beside it should be able to handle all the longitudinal load the mount will see.

Next up was actually designing the attachment point for the braces. Here's what I came up with:



This will also need to be machined out of Al, but should be fairly easy to make.

Also, the brace mounting points are designed so that I can press in two M10 studs. Doing studs instead of threads for bolts allows me to bring the braces as close to the main mount as possible, which is highly desirable.

The funky shape of the surfaces the E86 braces bolt to are a function of fastener clearance as well as clearance for the braces themselves. Braces are not perfectly flat up top, so had to design for that:



And here's how everything will fit together:




This angle really shows just how tucked up the braces are, there's not a lot of room to go much higher:



Next round of prototype pieces are being printed now. Excited for the next test fit!

[bigjae46]

Interesting.

There's two very distinct load paths for this piece though, so wouldn't the random fiber orientation be a detriment? Compared to orienting the fibers so the part is stronger in those directions.

Now I’m clearer on what you are doing. Carbon could absolutely be superior to aluminum. Easy Composites did a test on forged carbon, its tensile strength not that much lower than a 2x2.

[heinzboehmer]

Now I’m clearer on what you are doing. Carbon could absolutely be superior to aluminum. Easy Composites did a test on forged carbon, its tensile strength not that much lower than a 2x2.

Interesting, I'll have to look into it

[heinzboehmer]

Mooore changes.

First up, moved the front fastener bosses to also be in line with the load paths. I hadn't done this before because I had misunderstood the rule of thumb for placing holes near edges of a piece (remember, I don't actually know what I'm doing):



Test fit for everything printed. Love that mcmaster has models for all their parts, makes designing the clearances super easy:




(Yes, those are M10 threads grafted onto an M8 bolt + washer combo. The piece that gets bonded under the windshield is designed with M10 threads for timeserts, but the bolts holding the brace mounting point will be M8. Easiest way to get everything assembled and test fit fastener head clearance at the same time is with these frankenfasteners.)

I also updated the brace cut templates and made the first modification to one E86 brace:



Unfortunately, the hole is a couple mm off center. I guess that's what I get for trying to eyeball the drill hole from a sharpie mark instead of double checking with calipers. I'll likely redo this in the future, but gonna leave for now cause I'm mostly using these parts to validate my designs. Good news is that these braces came on every E85/E86, so there are tons available.

Here's the assembly on the car. Note the washer to space the brace down. I modeled the boss as tight to brace as possible and left the studs long so that I could decide on optimum spacing in real life. Upcoming iteration has a slightly taller boss for better fastener clearance:



That angle actually exaggerates how close everything is, it's not terrible with the washer in there (still super tight in the grand scheme of things though!):



I ended up landing on approximately two washers of boss surface extension:



For the test fit above, I printed low profile cap head fasteners, but I'll be able to get normal sized fasteners in there with the extra clearance, for maximum strength.

And here's where the brace lands at the strut tower side after the trim and all the other adjustments:



The point where those two silver lines intersect represents the furthest (center) point I can put a fastener at, so that its bearing surface stays fully within the flat portion of the brace. It just barely works! Should also have enough space to slot the new hole in the brace to allow for chassis/alignment differences.

Finally, here's the latest version of the design. Boss extended, real ribs modeled and fastener clearances in check with the taller cap heads: