heinzboehmer's 2002 Topaz 6MT Coupe

[heinzboehmer]

Track day went awesome. Weather was perfect (cool all day, no rain) and there ended up being a great group of people there. Tons of fun.

M3 friends:




And of course the girodisc dust boots did not survive. Found a melted one on each front caliper:




Also, got my 3D printer and continued with the brake duct development. First prototype on the left and latest on the right:



Test fitted the third one from the left:





Fits great! Newest version has the fan clocked differently for better clearance and has a slightly different front mounting tab design, but is otherwise the same.

Speaking of the fan, I decided that it would be much easier to test the duct performance with one, so I went this route instead of the originally planned passive vacuum generation. I'm still not sure if the fan will live on the duct or if it will live in the belly pan and be connected to the duct with some hose. The fan itself isn't too heavy, but it spins up to 16k rpm, so it puts a decent amount of torque on the duct when the wheel is turned. Having it in the belly pan means I can go to a more powerful fan too if needed. I'll have to do some (destructive) testing and decide if I want to move it or not. More to come

[heinzboehmer]

Love the 3D scanning and CAD work. What program are you using? Maybe I'm thick headed but FreeCAD has not been that easy to understand.

Are you trying to extract air from the brakes? That area is probably high pressure and turbulent. The air coming from the brake ducts comes from a relatively low pressure area which is what makes brake ducts so challenging. People say that to not aim brake duct air to the rotor face, which is correct, but the reasoning of distributing the air through vanes from the hub is part of the reason. The goal is to vent into a low pressure area to encourage air flow, I'd guess that the hub is the lowest pressure area.

You might find that you're not getting enough cooling.

Maybe if you could run that air to a really low pressure area, you could use that duct to extract air and then run a more convention setup to feed air to the brakes.

I used a Revopoint Pop 3 (I think) to scan the car and the Revo Studio software to generate all the meshes and merges. Then imported them all into Onshape and have been using that to design the duct. There's definitely a learning curve to Onshape, but I do like it. Coolest part is that it's cloud based, so I can open up models on my phone, which is really handy when I'm down in the garage.

And yep! Plan is to pull air out through the calipers. I was hoping I'd be able to duct to a low pressure area or build a passive venturi vacuum generator to get the air out. But the more I think about it, the more it seems like I should just make these things active. Stick a fan on the flange and build a small duct to guide the air out of the wheel well nicely. If it's going to be active, then might as well do it properly and add temp sensors and a microcontroller to control fan speed.

This whole idea came to life because the design of the CSL rotors doesn't promote the most airflow from hub to rotor veins. Will likely still install some Bry5on style scoops (will need to adapt his design to my modified dust shield, but that should be simple) to get some extra air to the hub, but this duct I'm designing should help with pad and dust boot temps/life even without those.

[bigjae46]

Love the 3D scanning and CAD work. What program are you using? Maybe I'm thick headed but FreeCAD has not been that easy to understand.

Are you trying to extract air from the brakes? That area is probably high pressure and turbulent. The air coming from the brake ducts comes from a relatively low pressure area which is what makes brake ducts so challenging. People say that to not aim brake duct air to the rotor face, which is correct, but the reasoning of distributing the air through vanes from the hub is part of the reason. The goal is to vent into a low pressure area to encourage air flow, I'd guess that the hub is the lowest pressure area.

You might find that you're not getting enough cooling.

Maybe if you could run that air to a really low pressure area, you could use that duct to extract air and then run a more convention setup to feed air to the brakes.

[heinzboehmer]

Well I completely forgot to account for the weights that come on street pads. Scanned the car with track pads and designed the duct around those. Fortunately the street pads are on the bench, so I just measured them and added to the CAD:


Definitely not going to work with those:



Design needed some tweaking anyway (thinner walls, reworked caliper reliefs), so V2 it is:











Since I'm 3D printing these, my approach is a very software-engineery approach. That is to say I'm tweaking a few things at a time, not validating anything besides a quick eyechrometer clearance check and printing. There's bound to be more iterations of this thing, so I'll just update with a pic of the different versions when I settle on a final one.

[heinzboehmer]

First duct prototype (a.k.a. the vacuum cleaner attachment) done:







I just followed the radius of the caliper, so there should be no clearance issues with the wheel:



Here it is placed into the merged scan:





The merged scan has the wheel at full lock to check clearances. Looks like it will be fine!

Now just need to find some time to print and test.

[bigjae46]

Same connector as for vanos solenoid pack

Thanks! Note to self…don’t plug the auto level harness into the VANOS…🤣

[maupineda]

Same connector as for vanos solenoid pack

[heinzboehmer]

Do you have a part # for that?

I didn't find the connector keyed for that specific location, so I just used the universal one: 61138383300

[bigjae46]

Front xenon leveling sensor connector was crumbling apart and being held on by zip ties, so I changed that out for a new one:

Do you have a part # for that?

[heinzboehmer]

Track day in a week, so prepped the car over the weekend. Swapped pads, swapped wheels/tires and addressed a couple small things as well.

Front xenon leveling sensor connector was crumbling apart and being held on by zip ties, so I changed that out for a new one:




Also swapped out the rest of the dust boots on the front calipers for GiroDisc ones. Not fun to do.

Here's the condition of the Brembo ones that came off:



Three are ripped, but I'm fairly certain that happened when I was removing them. Regardless, they were definitely cooked and very brittle.

All the pistons were still clean though, so not worried about the pressure seals:



Rear Brembo boots looked perfect, so I just left those alone.



I also finally reached out to my friend with the 3D scanner and we scanned the caliper/wheel/wheel well. Scans came out pretty awesome:






Was even able to create a decent merge of all the scans with the scanner software:



Looks like a mess in that pic, but promise the 3D version is totally usable.

Looking forward to finally designing the caliper evacuation duct. Current plan is to have it be two pieces: one that attaches and seals against the caliper and another that will hang out somewhere that sees high velocity air flow. First piece will just serve as an air guide out of the caliper and will connect to the second piece with a short section of brake duct hose. Second piece will be a venturi vacuum generator that will create the necessary low pressure to pull air from the caliper.

I've already started designing some different venturi designs to 3D print and test the effectiveness of. None of my designs are based on any hard science, really. Instead, I plan to use the guess and check method on this . Print a couple different designs, test with a leaf blower, tweak things and keep iterating until I'm happy (or get stuck and have to resort to some actual theory-backed design).

I do also intend to try and see if I can make the final version a one piece thing. Would be a much more elegant solution, but I'm not sure that I can make something that can stick out into fast moving laminar airflow and attach to the top of the caliper and not be gigantic. Plus, the two piece design with definitely simplify testing of the different venturi designs.

Anyway, excited to start messing around with this. I have access to a 3D printer at work, but this project is also making me want to just go out and buy one...

[heinzboehmer]

To get DME data onto the bus, you have to command the k-bus to send it. Same thing testo does, takes control of the k-bus logging.

On my list of things to do is to find/replace everything I can from k-bus to CAN bus, then disable k-bus logging. It’s low on the list though, as I’m pretty sure all of the data I want is not sent over CAN.

Aah that makes sense. Completely forgot that Gauge.S is primarily, well, a gauge

[Bry5on]

Hmm, this is surprising. The k bus is usually decently active, so not sure why those are causing issues. Any idea what the commands being sent out are? Steering wheel button things?

To get DME data onto the bus, you have to command the k-bus to send it. Same thing testo does, takes control of the k-bus logging.

On my list of things to do is to find/replace everything I can from k-bus to CAN bus, then disable k-bus logging. It’s low on the list though, as I’m pretty sure all of the data I want is not sent over CAN.

[heinzboehmer]

Super clean. Remember to always use twisted pairs and the shortest wire lengths you can for CAN, and as you know, try not to run CAN in parallel to power wires. This is mostly for people who might want to recreate this but run the wiring a little differently. If you’re pulling CAN from the steering angle sensor, this should be a non-issue since the run is so short. Also, don’t forget to run CAN to your OBD2 connector while you’re in there doing all of this. It’s so nice to be able to grab the data without doing any wiring if needed.

Yep! Have a junkyard harness for the steering angle sensor on the bench waiting for me to wire it up. Thanks for the reminder on the OBD plug CAN wiring, need to order the correct pins for that.

And one more note for future people. Gauge.S is actively sending commands to the k-bus, which means your code reader will not work while Gauge.S is active. Either kill power to Gauge.S or reboot it into WiFi mode (can be done with steering wheel controls). This one got me a couple weeks back as I couldn’t figure out why I couldn’t pull codes from my DME. -_-

Hmm, this is surprising. The k bus is usually decently active, so not sure why those are causing issues. Any idea what the commands being sent out are? Steering wheel button things?

[Bry5on]

Super clean. Remember to always use twisted pairs and the shortest wire lengths you can for CAN, and as you know, try not to run CAN in parallel to power wires. This is mostly for people who might want to recreate this but run the wiring a little differently. If you’re pulling CAN from the steering angle sensor, this should be a non-issue since the run is so short. Also, don’t forget to run CAN to your OBD2 connector while you’re in there doing all of this. It’s so nice to be able to grab the data without doing any wiring if needed.

And one more note for future people. Gauge.S is actively sending commands to the k-bus, which means your code reader will not work while Gauge.S is active. Either kill power to Gauge.S or reboot it into WiFi mode (can be done with steering wheel controls). This one got me a couple weeks back as I couldn’t figure out why I couldn’t pull codes from my DME. -_-

[heinzboehmer]

Going to be replicating Bry5on's wideband/logging setup and parts have started coming in. Last night I spent a few hours getting the wiring for the wideband in place.

Easiest way to wire it would be to just run the wire through some grommet in the drug bin area, but that would require wires to be floating around the engine bay and I wanted to avoid that. Since I pulled the precat O2 and EGT wiring out of the narrowband harness during my header install, there was just about enough room to feed it through there.

Getting the AEM wiring through required a decent amount of messing around, but it wasn't too too bad in the end. Depinned the O2 connector and used some coated stainless steel wire rope to guide it through:




Some time later it was in place:





Luckily, the wire length was just about perfect to reach a wideband O2 plugged into the headers. No need to be wiring up any extension harnesses or anything.

I did also fit that rubber wire grommet on the passenger's side of the valve cover (12517831185). Hopefully that keeps the top of the valve cover a bit cleaner. I spent some time trying a bunch of different wiring routes until I settled on something I liked. I wish the SS headers didn't make the narrowband O2 wiring so tight...

I then drilled some holes for zip ties in the transmission ECU compartment and mounted the wideband controller there:





In the future, I'll cut the loop of wire going between the wideband controller and the DME housing grommet in half and add some connectors to it. This is mainly so that getting the O2 harness out of the car isn't a huge pain.

And through to the interior:



And that's where I stopped. I have yet to install the O2 sensor onto the headers and have also yet to wire the controller in. My Gauge.S still hasn't come in, so probably won't continue with this until it does. Pretty happy with how stealthy this turned out though!

I did do a bit of thinking and I'm pretty sure I'm going to wire everything to the eject box power supply harness that I have in the center console. It's on a 10 A fuse so should be fine. Here's my messy current draw calculations for everything that's going to be connected to that supply harness:

• Front dash cam: 1 A peak
• Rear dash cam: 1 A peak
• Wireless charger: 1 A peak
• AEM Wideband Controller:
  • 3 A peak worst case
  • 1.5 A peak measured
• Gauge.s:
  • 500 mA peak worst case (estimated)
  • 150 mA peak measured (by Bryson)
• ---------------------------------------------------------
• Worst case peak current draw: 6.5 A
• Realistic peak current draw: <4.65 A

Think it should be safe! Will make everything super easy to wire as well. If I start popping fuses (I shouldn't), I'll run a separate fuse for the wideband/Gauge.S setup.