heinzboehmer's 2002 Topaz 6MT Coupe

[heinzboehmer]

Yeah the stiffness of the part will be an interesting thing to figure out. Too stiff and it won't fit through the door, too bendy and it'll flop around when driving, since it's not really held by anything in the middle.


Also, have started working on the CAD model and wow, it's been a struggle.

For the previous scan, I hung the part vertically from the ceiling, thinking that that would allow it to keep its shape. Unfortunately, mirroring the part along its short axis in the digital world made it very apparent that this assumption was incorrect.

I messed around with it for a while, thinking that I had my coordinate axes slightly tweaked, but eventually came to the conclusion that the part was deformed when I scanned it. This is as good as I could get it:



That gap is roughly 20 mm across. Nowhere near acceptable.

So, I gave up on that scan and redid it. This time, I built a professional fixture so that the part could lay horizontally on the ground without distorting:



For anyone following along at home, you'll need exactly these parts for max precision:

• 4x Dinner plates
• 1x Flooring sample
• 2x Varied candles
• 1x Folded over MF towel
• 3x Assorted rolls of tape

I set everything up so that the headliner was juust barely resting on the plates. They're not there to provide any support, but rather to provide flat references for plane alignment in CAD.

Newest scan is muuuch better (have updated the previous link):



Gray is the scan and yellow is the mirror. Not entirely perfect, but I didn't expect it to be. Can absolutely work with though.

Now it's just a matter of getting lost in surface land and coming out the other side with a model.

[bigjae46]

You can get the the stock one in that way, but it’s quite bendy.

Maybe that's why the 2M one never took off? The headliner would need to be installed when the roof is installed and most that would buy it already have a roof installed.

[Obioban]

Should be a little bendy but not much. Can't get it in through the door opening with the seats removed?

You can get the the stock one in that way, but it’s quite bendy.

[bigjae46]

Suspect you'd have to take out the windshield or black glass to install it, if it were completely rigid.

Should be a little bendy but not much. Can't get it in through the door opening with the seats removed?

[Obioban]

Got it...I'll see what I can find

My mistake, I was thinking the normal aluminum to carbon conversion.

It could still be a lot lighter than the fiberboard. With the edges being thinner then that makes it much easier. Since it will be upholstered, the part could be laid up with a 1k carbon which is less dense, a 4mm soric core with an extra layer of 3k around the edges. One side will have the exposed core but who cares. Not sure how much lighter. I don't have my soric core math lined up with reality well enough quite yet to make a close guess.

Suspect you'd have to take out the windshield or black glass to install it, if it were completely rigid.

[ac427]

I'll ask my metalurgist father, but I think it would apply to all or most alloys, not just copper. Some crush washers provided in various kits look to be aluminum at times, not always copper.

What did your Dad say about aluminum and annealing?

[bigjae46]

For sure.

But honestly, I think finding someone local with a CNC router and routing it out of pink insulation foam is gonna be way easier and way, way cheaper. That's how most of the RC plane people build their stuff, so maybe someone in that space can help out.

Random video pick, but jump to 4:20 for an example: https://www.youtube.com/watch?v=o6FMjOl0TRA

How did you get to the 33% lighter figure? Scan measures in roughly at 1.5 m2 for surface area of the part. A layup with 4 (to make the part thickness ~1mm) 210g 2x2 3k layers would weigh ~1260 g dry. Leaves very little margin for the weight of the resin + fabric, unless 4 layers is way overkill.

Also, the part is made opposite to what you're describing (and what I expected originally). The thickness is ~5 mm throughout most of the part (sans fabric), but the edges are "crimped" down to ~1 mm to fit under the other trims in the interior. I'm guessing the increased thickness in the middle is for either stiffness or NVH. If the carbon part is stiff enough, it could likely be made at 1 mm thickness throughout. NVH is easy to address with the "recycled jeans" insulation (51488229733 + 51488229734) that you can get from BMW.

Got it...I'll see what I can find

My mistake, I was thinking the normal aluminum to carbon conversion.

It could still be a lot lighter than the fiberboard. With the edges being thinner then that makes it much easier. Since it will be upholstered, the part could be laid up with a 1k carbon which is less dense, a 4mm soric core with an extra layer of 3k around the edges. One side will have the exposed core but who cares. Not sure how much lighter. I don't have my soric core math lined up with reality well enough quite yet to make a close guess.

[karter16]

Interesting. I didn't think it was very hard to get mine in and out through the door when I did it - only minor flexing / squashing of the corners required and that was with the CSL bucket seats in place. I did the headliner in my 318i sedan many years ago and that was horrible to fit through a sedan-sized door.

[heinzboehmer]

I was hoping you picked this up! I almost bought it but shrugged it off after he upped the price behind the scenes.

These definitely need to be replicated, even just as an non-upholstered blank. Though I've removed 3 of these recently and there does needs to be some flexibility in the part for R&R. Also, unless the windshield is pulled, it's a pain getting it through the door.

Yeah, it was silly money, but what are you gonna do. Seller got a higher offer than mine, came back to say that I was still first in line and that if I could match the offer, I could have it. Free market and all that. I blame BMW for not making these anymore more than anyone else

Also, I'm hoping the ridiculous price and non-existent availability will be a net positive for the community. Would have never thought to try and recreate this part had either of those not been true.

Anyway, appreciate the insight on the R&R! I've only ever done this job on a wagon, which is cheating cause the whole rear of the car gets out of your way. I know this is tough to quantify, but how flexible would you say it needs to be? Needs to shrink to 95% of its width? 80%? 50%?

Could you conceivably get it into the car with zero flex if you remove the door and seat?

(I know that last bit sounds kinda insane, but both of those components are surprisingly easy to pull on this chassis)

[bavarian3]

Few updates:

Update 3:

1. I managed to get my hands on some unobtainium:

[ATTACH=JSON]{"data-align":"none","data-size":"full","title":"20260504_200205.jpg","data-attachmentid":35356}[/ATTCH]
[ATTACH=JSON]{"data-align":"none","data-size":"full","title":"20260504_212759.jpg","data-attachmentid":35457}[/ATTCH]
[ATTACH=JSON]{"data-align":"none","data-size":"full","title":"20260504_214123.jpg","data-attachmentid":35458}[/ATTCH]
[ATTACH=JSON]{"data-align":"none","data-size":"full","title":"original_9b24949e-b155-46de-89f5-6d3277b4ce90_20260504_214139.jpg","data-attachmentid":35459}[/ATTCH]

A never installed, NOS, 2021 production date euro slicktop headliner, courtesy of ridebikes. Incredible!

Gonna be kept in storage until I get the carbon roof project going, but I did do one thing before putting it away:

[ATTACH=JSON]{"data-align":"none","data-size":"full","title":"Screenshot 2026-05-05 at 1.11.57\u202fAM.png","data-attachmentid":35430}[/ATTCH]

Can check out the scan here, if you're curious: https://s.digital3dcloud.com/space/f...b-56687ab66f9f

Interesting to see how different the euro part is to the USDM one, especially from the back. Here's an internet picture of the USDM headliner, for reference:

[ATTACH=JSON]{"data-align":"none","data-size":"full","title":"IMG_5909.jpg","data-attachmentid":35461}[/ATTCH]

This makes me think that the foam and aluminum honeycomb structures are there for crash safety and not NVH, like I originally thought. Maybe it has to do with the extra US regulations for unbelted passengers?

Anyway, it seems dumb that these parts are so hard to get now. I'll try to spend some time recreating the scan in CAD, both to practice my surface modeling skills and also to try to make it as enticing as possible for composite manufacturers to start making replicas of these parts.

I'm thinking that with an accurate CAD model, anyone can CNC route the shape out of something like pink insulation foam board and then pull a mold. Maybe bigjae46 or Karbonius or Mile End will want to get in on this? Would love to see a quality replica of this parts available in the future.


Also weighed the part out of curiosity. Came in at 1959 g. Not bad at all! I honestly doubt a CF version could be made lighter, but being able to buy something is definitely worth compromising a few hundred grams for.

I was hoping you picked this up! I almost bought it but shrugged it off after he upped the price behind the scenes.

These definitely need to be replicated, even just as an non-upholstered blank. Though I've removed 3 of these recently and there does needs to be some flexibility in the part for R&R. Also, unless the windshield is pulled, it's a pain getting it through the door.

[heinzboehmer]

The euro wagon/sedan pillar covers are also much tighter to
the chassis. Running euro pillars gave me notably better sight lines (A pillars) and more shoulder room (b pillars).

never checked the coupe’s Bs, but sadly there was no similar gains to be had for the As.

Looks like the same part numbers for all the coupe pillars between euro and USDM, so unfortunately no upgrades to be had.

Checked the E82 as well, since the giant A pillars on it are one of my biggest complaints, but again, no difference.

Interesting. I'd have to use a CNC service. Is it possible to use the CAD model and then the CNC service make it in 3-4 parts that I can bond together?

For sure.

But honestly, I think finding someone local with a CNC router and routing it out of pink insulation foam is gonna be way easier and way, way cheaper. That's how most of the RC plane people build their stuff, so maybe someone in that space can help out.

Random video pick, but jump to 4:20 for an example: https://www.youtube.com/watch?v=o6FMjOl0TRA

A carbon part should be about 33% lighter. You'd have to figure out what areas need to maintain the original part thickness and add some layers along those edges - which would add weight.

How did you get to the 33% lighter figure? Scan measures in roughly at 1.5 m2 for surface area of the part. A layup with 4 (to make the part thickness ~1mm) 210g 2x2 3k layers would weigh ~1260 g dry. Leaves very little margin for the weight of the resin + fabric, unless 4 layers is way overkill.

Also, the part is made opposite to what you're describing (and what I expected originally). The thickness is ~5 mm throughout most of the part (sans fabric), but the edges are "crimped" down to ~1 mm to fit under the other trims in the interior. I'm guessing the increased thickness in the middle is for either stiffness or NVH. If the carbon part is stiff enough, it could likely be made at 1 mm thickness throughout. NVH is easy to address with the "recycled jeans" insulation (51488229733 + 51488229734) that you can get from BMW.

[bigjae46]

I'm thinking that with an accurate CAD model, anyone can CNC route the shape out of something like pink insulation foam board and then pull a mold. Maybe bigjae46 or Karbonius or Mile End will want to get in on this? Would love to see a quality replica of this parts available in the future.


Also weighed the part out of curiosity. Came in at 1959 g. Not bad at all! I honestly doubt a CF version could be made lighter, but being able to buy something is definitely worth compromising a few hundred grams for.

Interesting. I'd have to use a CNC service. Is it possible to use the CAD model and then the CNC service make it in 3-4 parts that I can bond together? A carbon part should be about 33% lighter. You'd have to figure out what areas need to maintain the original part thickness and add some layers along those edges - which would add weight.

[Obioban]

The euro wagon/sedan pillar covers are also much tighter to
the chassis. Running euro pillars gave me notably better sight lines (A pillars) and more shoulder room (b pillars).

never checked the coupe’s Bs, but sadly there was no similar gains to be had for the As.

[0-60motorsports]

Thats awesome you found a NOS one. I tried to get one from BMW direct and they dont even have the beige one in stock. This is because i have done mine in alcantara and i want to revert back to stock fabric....So thought i could sell the alcantara one locally and......

[heinzboehmer]

Few updates:

Update 1:

Cerakoted parts are still not ready. Supposedly get them back end of the week.


Update 2:

With Ted's guidance, I was able to get D Bus -> I Bus message relaying working. The third and sixth messages below were relayed from D Bus, with `DE AD BE EF` being the payload.



This is what I sent over D Bus:

BB 07 DE AD BE EF 9E

It's essentially just a normal DS2 command (first byte is destination, second length, last checksum and the rest are the payload) that gets routed by the IKE to I Bus if the destination is on that bus.

Next step is to figure out how to get Gauge.S to send these messages out on demand. Ideally, I'd like to program in event-based sending, but not sure if that's possible. Will investigate.

The other thing I noticed is that the IKE relaying is super slow. I played around with sending the D Bus message at different frequencies (all the way from a couple Hz to a couple hundred Hz), but the fastest I saw the relayed message pop up on the I Bus side was at ~1 Hz. I can't imagine tools like Inpa are operating at speeds that slow, so I'm hoping that I was just overwhelming the IKE. Again, will investigate.


Update 3:

I managed to get my hands on some unobtainium:






A never installed, NOS, 2021 production date euro slicktop headliner, courtesy of ridebikes. Incredible!

Gonna be kept in storage until I get the carbon roof project going, but I did do one thing before putting it away:



Can check out the scan here, if you're curious: https://s.digital3dcloud.com/space/f...5-eb2c04e8a802

Interesting to see how different the euro part is to the USDM one, especially from the back. Here's an internet picture of the USDM headliner, for reference:



This makes me think that the foam and aluminum honeycomb structures are there for crash safety and not NVH, like I originally thought. Maybe it has to do with the extra US regulations for unbelted passengers?

Anyway, it seems dumb that these parts are so hard to get now. I'll try to spend some time recreating the scan in CAD, both to practice my surface modeling skills and also to try to make it as enticing as possible for composite manufacturers to start making replicas of these parts.

I'm thinking that with an accurate CAD model, anyone can CNC route the shape out of something like pink insulation foam board and then pull a mold. Maybe bigjae46 or Karbonius or Mile End will want to get in on this? Would love to see a quality replica of this parts available in the future.


Also weighed the part out of curiosity. Came in at 1959 g. Not bad at all! I honestly doubt a CF version could be made lighter, but being able to buy something is definitely worth compromising a few hundred grams for.