heinzboehmer's 2002 Topaz 6MT Coupe

[D-O]

How can we get some of those scoops?!?!

[Bry5on]

You had blanked the bumper brake duct holes for this one right? Ie: only clean air coming from under the car, not through the factory ducts, same as my testing?

[heinzboehmer]

Alright, time for a bit of science. Did some testing to qualify whether the brake scoops help the dust boot problem at the latest track day.

This whole experiment is based on the assumption that my driving is consistent enough to draw somewhat reliable conclusions from. Thunderhill 3mi is definitely the track that I know the best, so I don't see much of an issue with consistency. Also, the two track days I'm comparing did happen at different times of the year, but the day with scoops was MUCH hotter, so the scoops needed to be extra efficient to make a difference. Here's a quick overview of the experiment design:

No scoops track day:
- Mar 2024
- Thunderhill 3mi bypass configuration
- 3.5 hrs of track time
- Ambient temp: ~15 C
- Cloudy skies
- ~2:05 min/lap
- More pad life

Scoops track day:
- Sep 2024
- Thunderhill 3mi bypass configuration
- 3.5 hrs of track time
- Ambient temp: ~35 C
- Sunny skies
- ~2:05 min/lap
- Less pad life

Car setup (same both days):
- ~3200 lbs
- ~315 whp
- MK60
- RBF600 (purposefully did not flush fluid so that the only potential difference would be worse thermal capacity on the scoops day)
- Bridgestone Potenza RE-71RS
- Cerakoted 996 calipers (finish supposedly helps dissipate heat)
- Girodisc titanium pad shims
- Brand new girodisc dust boots installed before each day
- Same PFC08 pads front + rear (as in the same exact parts, not just the same compound)

The experimental results are meant to be qualitative and not quantitative. Basically just trying to see if having the scoops makes the dust boots survive.

I found one cooked dust boot on each side of the car after the no scoops track day. Scoops track day involved thinner pads and higher ambient temps, so zero cooked boots would point to working scoops and not increased thermal capacity in the system (in reality thermal capacity decreased).

Aaand the results:






Success! Dust boots seem totally fine We'll have to see how many more track days they last.

Also, a few entertaining things I found when dressing the car back up for street duty.

These used to be rotated 90 deg:



These pins and screws used to be centered on the rotor holes:



And these tires used to work:



Man, this last track day really made me remember how hard this sort of driving can be on the hardware. Tons of fun, but it really does age parts very fast.

(and yes I'm aware I need more camber, more on that soon)

[heinzboehmer]

Some more Gauge.S progress. Was made aware of these two repos recently, so decided to take the nuclear option and write my own validation firmware for the board:

https://github.com/handmade0octopus/Gauge.S-sd-updater

https://github.com/handmade0octopus/ds2

First thing I did was build the DS2 library as a standalone to unit test it. Fed the tests some real data coming out of my DME and fortunately the logic in that library seems good. Test code snippet:



I then set up the programmer and PlatformIO in VSCode to flash the board. First time using PlatformIO for Arduino stuff and it's pretty nice. Feels a bit more professional than the official Arduino IDE.

The Gauge.S board is surprisingly simple. At its core, it's just an ESP32 connected to a TJA1021 for handling D Bus communications. This means I can write some super quick Arduino code to send out DME requests and then listen to what comes back. Code is written, just need to find some time to test. I should start billing for the time spent debugging this...

Anyway, onto some other stuff.

Decided I didn't want the cooling efficiency tradeoff of the Rogue power pulleys anymore, so swapped them out for stock. Old stock PS pulley was cracked, so brand new one is in. Don't want to think about what that part is gonna cost in a few years.

Power pulleys:



Back to stock:



Didn't order new idler pulleys because the ones on the car are fairly new, but of course I found them both to be bad with the belt off. It's crazy how little these things have lasted. Think I've gone through four sets since I bought the car? Oh well, new ones ordered.

Then met up with Bry5on and Slideways to do some tuning. I mean, in reality I just provided another set of hardware to experiment with and Bry5on did all the work, but still

Newest version of the mullet tune is even better. There used to be a bit of a torque dip at 2-3k rpm, but now that range feels awesome. Need more seat time to find any nuances, but so far feels great.

And then on the way back home I heard the dreaded failed CSB rumble under load. Incredible timing with a track day less than a week away. Parked in the garage and immediately took things apart. Interestingly, I found the CSB in great shape. However, the real failure more was way, WAY scarier:



I have no idea how this happened. I always torque this bolt with the driveshaft in the car, so that I can get as much torque on it as possible. Maybe another case of the track rattling things apart?

Fortunately, both the bolt and the threads in the driveshaft were perfectly fine, so I poured two gallons of threadlocker on the bolt and did the double wrench thing to torque it to 750,000 lb-ft. Hopefully this doesn't come undone again.

Perhaps unsurprisingly, torquing this bolt back up cured about 95% of my drivetrain clunk. It's crazy how much quieter the car is now. I thought it had been extra noisy recently... Will keep an ear out for similar noises in the future.

And lastly, here's a photo dump of the dust boots on my front calipers. Want to run an experiment with the scoops, so swapped some fresh GiroDisc boots in.

Old GiroDisc on FL caliper:




Old GiroDisc on FR caliper:




New GiroDisc on FL caliper:



New GiroDisc on FR caliper:



Interestingly, the rear caliper boots are still the stock Porsche/Brembo ones and those are showing absolutely no signs of failure

[heinzboehmer]

More like 25 HP - plus the weight savings.

Oh man, how could I forget about the weight savings? Don't even need a carbon roof now.

[D-O]

More like 25 HP - plus the weight savings.

[0-60motorsports]

Yep! 61349316116

Cheers

[heinzboehmer]

Thats cool. Got a pn for the plug?

Yep! 61349316116

[0-60motorsports]

Thats cool. Got a pn for the plug?

[heinzboehmer]

Okay so INPA does not show EGT. However, I did learn that DIS can grab a ton of information from the DME under DME -> Expert Mode -> Diagnosis Requests.

Here's EGT (amongst other sensors):



And a sample of other interesting things like oil level and VANOS timing:




Now just need to sniff D Bus while I open up all the different screens to reverse engineer all this data. Honestly sounds like a lot of tedious work, but having all that data in a log does sound very appealing...


Also, I can't believe I forgot to show off my most important mod to date! Easy 20 hp from this thing:



In all seriousness, the open 12V socket was annoying me so I bought one of the modern BMW 12V socket plugs. Don't smoke, so no need for the cigarette lighter one and I like how this one is less busy.

[Bry5on]

Brilliant idea to record the traffic during an INPA session

[heinzboehmer]

Stuck home with a cold this weekend, so I knocked out some boring maintenance that I had been putting off for a couple months.



Swapped fuel filter, filter to fuel rail line, EGT, PCV line that goes to the airbox and a bunch of underside trim clips/screws that were starting to get old and worn. Always fun to do your best to evacuate the system of fuel beforehand, only to still have it spill everywhere.

Everything else was uneventful. I haven't seen the intermittent EGT code I was getting pop back up and EGT readings on Gauge.S are different now. I say "different" and not "correct" because it's clear the scaling is still way off.

Originally, I was planning on using this new sensor to calibrate the Gauge.S reading. DME has this table for the sensor:



And a third degree polynomial fits it perfectly:



So, it should just be a matter of exposing the sensor to a few known temps and measuring its output to correlate it with the graph.

I was about to head to the kitchen for some ice water when I realized that the DME is likely not sending out the raw sensor values over D bus. All other temp sensors are scaled linearly in the output data, so it would be very unlikely that this one is different. I'll check to see if INPA or DIS show EGT. If they do, it should be super easy to figure this out using a technique like this one: https://github.com/tomicooler/bmwe46...il-temperature

I did also set up the cheap logic analyzer I got and started trying to figure out why the V5 Gauge.S doesn't work with the DME, but the V4 does. Here's version 1 and 2 of my D bus probing setup:




I haven't really gotten anywhere yet. The V5 board on older firmware was sending out the correct DS2 commands, but would then hold D bus low and not allow the DME to respond. Updated the firmware and swapped the janky trimpot I had used as a D bus pullup for a properly sized 10K ohm resistor, then tested again.

Janky:



Not janky:



Now Gauge.S sends out the correct command and the DME responds with a valid response, but for some reason the device still does not display anything. It communicates just fine with the MK60, so I'm fairly convinced this is a FW issue. Unfortunately, Gauge.S FW is closed source, so I can't really debug this...

Anyway, I'll keep messing with this and update with any resolutions.

[heinzboehmer]

Small addendum to the phone holder project. Drove the car a bunch last week and found two issues with the bumpers that hold the phone in place:

• Top bumper is designed to be a press fit into the top tab. The top tab has a rib in it for rigidity, but because of packaging reasons, it had to face down. This means when the tab + bumper are pressing against the phone, the bumper is also being forced forward. After enough vibration/time/etc. the bumper will fall off. Was fun trying to catch my phone lol. Some visual references of the design:

One good solution would be to add a ledge to the end of the tab that will hook into the bumper once installed. However, this would leave overhangs inside the bumper and I really didn't feel like spending a bunch of time trying to print an overhang without support using that super soft stuff. So, a dab of 3M emblem and trim adhesive was the solution I picked.

• Turns out the super soft filament was too soft for the bottom bumper. Again, after enough vibration cycles, the phone would scoot forward just enough and bend the bumper out of the way. Was also fun to try and catch it when this happened. To fix, I extended the bumper forward by 2mm (to cradle the phone a bit better) and printed it out of the harder TPU. This bottom piece is really just acting as a rest for the phone, so the hardness of the material doesn't really make much of a difference with how well it's held in place.

Besides that, holder has been great! Pretty happy with how it turned out.

[heinzboehmer]

Annual wheel bolt replacement came a bit early cause I noticed some slight surface rust on some of the rear bolts. It's very orange (i.e. newer), so likely just from heat cycling at the track and getting washed.



The ease with which I can replace these is the number one reason why I don't have a stud conversion kit installed

[heinzboehmer]

If no one else tells you this let me, this development of the vent phone mount has been awesome! Also if you ever feel bothered to print another I would love to buy one

Thanks!

Here's the CAD if anyone wants to print it out: https://cad.onshape.com/documents/aa...abcccfec208abd

Anything suffixed with "Bumper" is made to be printed out of a flexible material (regular 95Aish TPU works, but the super soft 60A stuff works better) and everything else is made to be printed out of ABS or similar.