1997 m3 sedan with s54 swap

[3staxontheradio]

I spent an afternoon fighting an upstream O2 sensor in my headers and couldn't get it out, ended up rounding the sensor pretty well too. All my tools just slip. Torch, specially O2 sensor tools, and a big impact on deep well O2 socket wouldn't take it. I currently have Megan Racing headers with vbands in place of stock flanges- The vband welds leak badly and they are placed too close together so it makes fitting clamps a huge pain in the ass. I also want to be on stock flange locations to make header development easier.

I'm going to cut these headers off of the car with a reciprocating saw so I can do my scan plus I'm going to trash them for the new flange locations anyway. Planning another set of cheap headers for now to keep the car on the road (megan racing seems like a good option for a euro header clone). Does anyone know where I can find the correct 2 bolt flanges and details on how the tubing should be flared to correctly seal to OEM style flanges and gaskets? Does anyone sell formed stubs I can tig onto rather than finding/buying a die to do it myself?

[3staxontheradio]

I was still having issues with my selector rod loosening up over time so I went to pins and clips like OEM and autosolutions uses. The reboot engineering selector rod actually fit pretty well. I had to switch to a shorter bushing from mcmaster and then use a few shims and wavy washer on each joint to get the shoulders of the clip grooves to line up with the edge of the edge of the bushing. Results are good though and it should prevent anything from loosening.


I'm also in the process of pulling headers to do a packaging scan for header development.

[3staxontheradio]

My ltft continues to drift down. Now at .93 and .91. I do have two O2 sensors I plan to swap in today. Also given the way the car coughs on a hard start after it sits for more than a week I think I need to take another look at the injectors.

What else should I look into? Evap purge valve? I'm not actually sure which evap purge valve is in the car and I need to look at how the whole evap system is set up right now.

[heinzboehmer]

heinzboehmer You had a noticable improvement in ltft and drivability with that line being installed. I did a school dropoff run this morning and the car felt great again. Both resetting adaptions and I guess that drain hose?

I don't really understand why this constitues a vacuum leak when you aren't running a maf. The airbox before the throttle bodies isn't a vacuum source and it's before the air is "metered" by throttle position or MAP. The oil pan is getting essentially atmospheric pressure from the airbox too so 🤷

Oh, messing with the line under the airbox didn't contribute (negatively or positively) to driveability.

I swapped out the cracked original ICV hoses at the same time and those absolutely made a difference.

[samthejam]

more info about that

https://nam3forum.com/forums/forum/special-interests/coding-tuning/246025-ecuworx-issue-with-clutch-switch-direct-wire-and-gps-delete

Notes for when I did my manual swap below. Keep in mind the locations/addresses are for 2701 SW version. Please correct for your SW Version and keep in mind I modified the XDF to be in decimal for these, as I just found that easier to edit.


1) Modify K_SKRAFTS_CONTROL to 0x02

• Location: 0x4028
• Edit: Set to 0x02 → 2 (default is 0x01 → 1)
• Purpose:
  • The DME currently treats the clutch sensor and Gear Identification Sensor as one, since they are wired in serial.
  • This setting aligns the behavior with how the Z3M is wired, where the clutch switch is directly connected to the DME.
  • With this change, the DME will recognize that the circuit can still be closed even when the car is in neutral with the engine running.

2) Modify K_S_GANG_LL_ED to 0xFFFF

• Location: 0x5952 on 2001/2701
• Edit: Set to 0xFFFF → 65535 (default is 0x0032 → 50)
• Purpose:
  • This prevents the DME from throwing an error when the car is idling with the clutch engaged (up).
  • It essentially adjusts the logic for idle behavior, ensuring that the ECU does not falsely detect an issue when the car is stationary?

3) K_MD_MIN_VERH_KRAFTS

• Purpose: This parameter defines the minimum torque threshold required for certain powertrain functions to be active?
• Edit: Set to 0xA0 → 160 (default is 0x80 → 128)
• Location:
  • 0x67E0 on MSS54
  • 0x88A2 on MSS54HP
• Default Values:
  • US cars: 0x80 (128 in decimal)
  • Euro cars: 0xA0 (160 in decimal)

https://forums.m3cutters.co.uk/threads/smg-to-manual-conversion-complete-with-pics.150793/post-1991521

[3staxontheradio]

Also that DME utility from Martyn is very useful. Is that cylinder ignition showing ignition timing advance/retard ?

[3staxontheradio]

About to drive to Atlanta to pick up a 500mm bed 3d printer to use for fdm header test parts but I took a quick look at ltft. They don't look far off (yet?) There have about 45 minutes of city and highway driving on them. I'm going to keep watching to see if they drift.



Also my clutch switch was set to switch+gps so I guess I need to change that.

I also need to code out the dsc stuff and figure out why my brake switch is complaining. Brake lights come on just fine and the error count for that when I have been tracking it is huge.

[samthejam]

I have had the karbonius box on the car for about a week- the sound is great but I'm having some issues I think are related to my map sensor install. I also updated the bootloader with MpowerE36's tool and am currently running Bry5on mullet tune. The car is down on power, especially on the low end, and throttle response is lazy. When I first did the airbox install everything felt great except for some hiccups with the HTE tune that happened to be on the DME before I got it- not faulting the tune, its not for my car and I know nothing about it other than it says its an HTE tune for someones CSL car.The lazy low end power and throttle response happens with both tunes and if I reset adaptations it gets better temporarily. Long term trims are picking up bad data and throwing the car fueling off. Also with both tunes I get some on throttle hiccups and hesitation at medium to high loads and 4000ish RPM. I think this is just a fueling issue in line with whatever is throwing long term trims off.

I wired my MAP sensor by splitting the IAT ground off of the MAF harness (its all properly crimped open barrel splices, not worried about wiring and crimp quality). Martyn has a very helpful MAP sensor testing guide here:

https://www.ecuworx.co.uk/2016/03/15/checking-your-map-sensor-is-working-when-running-a-csl-clone/

I'm getting slightly different voltages when the car is off and when the car is idling, I think my ground location is causing a signal offset and I'm curious if other people have also grounded both MAP and IAT off of the same wire, its pin 25 on X60003.

Key in position 2 I get 4.02 volts and idling at about 900rpm I get close to 1.3v for and a 360 mbar reading. I live at about 400 feet of elevation so that doesn't really account for the difference in voltage when the car is off at position 2. Atmospheric pressure where I am should be 998.6 mbar which is a much smaller voltage difference.

Also very curious to hear what other people with correctly running CSL cars are getting when they run the testo tool linked from martyns article. FWIW the tool is x86 not x64 and it took me a minute to figure out why I was still getting MSVCP100.dll errors when I have Visual Studio C++ redistributable packages installed (I only had x64 version).

What did you do for the clutch switch DME settings? jayjaya29 we did yours like the Z3M right? Asking because that can create throttle lag like issues and stalling issues if it has GPS and clutch switch logic, unlike Z3m only looking for the clutch switch

[Bry5on]

That shouldn’t constitute a vacuum leak on the s54 - it’s got to be something else. ICV maybe?

[3staxontheradio]

heinzboehmer You had a noticable improvement in ltft and drivability with that line being installed. I did a school dropoff run this morning and the car felt great again. Both resetting adaptions and I guess that drain hose?

I don't really understand why this constitues a vacuum leak when you aren't running a maf. The airbox before the throttle bodies isn't a vacuum source and it's before the air is "metered" by throttle position or MAP. The oil pan is getting essentially atmospheric pressure from the airbox too so 🤷

[heinzboehmer]

Someone said that BMW might have updated that hose at some point to make it a tighter fit in the airbox. A hose from a later S54 or a new one from BMW should fix the issue.

New line worked on my car for a bit, but eventually started falling out again. Think it was just the newer o ring holding it in better.

Ended up modifying the boss on the airbox. Hasn't fallen out since: https://nam3forum.com/forums/forum/m...403#post328403

[Slideways]

I just found out that the crankcase drain hose was disconnected again (the small one under the airbox). I think the fitting is worn out and its popping out of place. Not sure that really constitutes a vacuum leak anymore since its before the throttle bodies so TPS based airflow calc doesn't care and MAP won't care either.

Someone said that BMW might have updated that hose at some point to make it a tighter fit in the airbox. A hose from a later S54 or a new one from BMW should fix the issue.

[karter16]

karter16 any chance you could load up testo and see what your key position 2 MAP voltage reading is when the car isn't running?

Even better, found a photo from a few months ago.

[3staxontheradio]

karter16 any chance you could load up testo and see what your key position 2 MAP voltage reading is when the car isn't running?

[karter16]

I'm confident in my splices but I'm also splicing off of a whole bundle of already extended MAF wires from the previous owners adapting the e46 harness to the e36 and I can't see what's under the shrink tubing. It's only 3 wires and I have like 15 different colors of tefzel so probably worth just running 3 new wires and not sharing a ground with the IAT.

Also on map sensor scaling: I have the Bosch sensor and martyns tool shows it on the default CSL map scaling. I assume that's correct?

Sharing ground with the IAT shouldn't be a problem - it's the exact same setup I'm running. I believe when I checked the IAT, MAF and MAP are all driven from the same UB (supply) circuit anyway, so ground should be directly the same between them anyhow. But yeah probably worth trying a second set of wiring anyway.

Yep cool if you have the Bosch '101' sensor then the scaling is the default.