Discussion about issues encountered whilst converting to M3 Flash and GWS

[derekgates]

Awesome Derek! Glad its up and working.
If i send you a word version of the coding doc would you be able to take the time to add the relevant changes ?

Thanks! It's a big process. I can definitely help with changes. Google Docs works well for collaboration if you are OK with that? Happy to help with Word. I believe the last steps we need in the guide is the discussion about the SPA (gear indicator) settings, Motorsport setting and HSA (start assist). Add a blurb about resetting modules (such as this guide from @LMB335is ).


For my last struggle with this retrofit, I am trying to figure out how to reset transmission adaptations. My DCT is not happy in the least and is struggling to upshift. This is expected, but I can't figure out how to change adaptations.

I've tried:

Put key in car
Turn on electronics (Mode 15) with cluster activated but car NOT started
Hold accelerator for 30 seconds

I also saw talk of using Tool32 to reset:

http://www.e90post.com/forums/showpost.php?p=19184537&postcount=4
"You can reset the GS40 adaptation values with Tool32. (C:\EDIABAS\ECU\DKG_90.PRG):"

<JOB>
  <JOBNAME>STEUERN_ADAPTION_ZURUECKSETZEN</JOBNAME>
  <JOBCOMMENT></JOBCOMMENT>
  <JOBCOMMENT>Adaptionswerte Zurücksetzen</JOBCOMMENT>
  <JOBCOMMENT>( Service ID 0x30, Identifier 0x04 )</JOBCOMMENT>
  <JOBCOMMENT>Voraussetzung für ordnungsgemäße Funktion ist, dass</JOBCOMMENT>
  <JOBCOMMENT>man in Anschluss die Zündung abschaltet und den</JOBCOMMENT>
  <JOBCOMMENT>Steuergeräte-Abfall abwartet, damit die Defaultwerte</JOBCOMMENT>
  <JOBCOMMENT>im NVRAM abgespeichert werden.</JOBCOMMENT>
  <ARG>
  <ARGNAME>ADAPTIONSWERT</ARGNAME>
  <ARGTYPE>int</ARGTYPE>
  <ARGCOMMENT>Eingebe des Hexwertes</ARGCOMMENT>
  <ARGCOMMENT>0xA0 = Adaptionswerte Kisspoint K1</ARGCOMMENT>
  <ARGCOMMENT>0xA1 = Adaptionswerte Kisspoint K2</ARGCOMMENT>
  <ARGCOMMENT>0xA2 = Adaptionswerte Kupplungsmomentenkennlinie K1</ARGCOMMENT>
  <ARGCOMMENT>0xA3 = Adaptionswerte Kupplungsmomentenkennlinie K2</ARGCOMMENT>
  <ARGCOMMENT>0xB1 = Adaptionswerte für DRM</ARGCOMMENT>
  <ARGCOMMENT>0xB2 = Adaptionswerte des Semischlupfs für Gänge 1-7</ARGCOMMENT>
  </ARG>

There is also the ISTA/D method (I don't have this on my VM, unfortunately):

Any thoughts?

 

[aus335iguy]

I’ve been told that resetting adaptations can lead to premature clutch wear. So I’d hold off. This seems to be related to coding of M Drive but I’m not clear on exactly which step is the cause. It may be tune or even DSC related. Derek, what state of tune is your car at?

 

[derekgates]

I’ve been told that resetting adaptations can lead to premature clutch wear. So I’d hold off. This seems to be related to coding of M Drive but I’m not clear on exactly which step is the cause. It may be tune or even DSC related. Derek, what state of tune is your car at?

I also saw premature clutch wear as one of the concerns; a big reason I am asking around before attempting it. Unfortunately, my car is riding the clutch like CRAZY. It's 'feathering' the whole time until it finally settles down. It feels it is going to destroy the clutch pack during daily driving.

I am tuned by @WedgePerformance but I haven't had a tune update or data logs in a long while (need to!). My last logs:

https://datazap.me/u/dg/wedge-performance-dct-v6


I am FBO but with stock turbos and outlets (my MMP outlets blew up) coupled with an MFactory LSD. Stock diff ratio.


[EDIT]: You mention the coding of MDrive but I skipped the MDrive VO coding in NCS Expert. My car is only set to $2MK. Do I need to set $2MD for this retrofit?

 

[aus335iguy]

We need to isolate which modules are causing this and then which parameters are the ones responsible.
To find out which module. Take a trace of how the car is right now with all of the “correct” settings
Revert to stock, And then slowly add in what we need until the car starts exhibiting the weird shifts. Once the slow shifts appear go back to the last step and analyse what we changed last

 

[aus335iguy]

Sorry i was typing while you were.... That’s really weird.

There has to be something different about @dyezak ‘s car to make it shift normally most of the time.

 

[derekgates]

We need to isolate which modules are causing this and then which parameters are the ones responsible.
To find out which module. Take a trace of how the car is right now with all of the “correct” settings
Revert to stock, And then slowly add in what we need until the car starts exhibiting the weird shifts. Once the slow shifts appear go back to the last step and analyse what we changed last

Hmm, the only changes I have are in the KOMBI and DSC.

DSC has start assist disabled (HSA).
KOMBI has SPA settings enabled (gear shift indicator), MOTORSPORT set to wert_01, DZM_DISP_RPM to wert_02.

All other modules (KOMBI, DSC, 2JBBF, NFRM) were reset to VO settings ($2MK added, $2TC removed).


Could this be DSC coding? I didn't see anything else that was needing to be done

 

[aus335iguy]

When you change your VO and default a module it adds the parameters that you need for that option. So although you’ve only manually changed two or three parameters many more others could have appeared or disappeared.
That’s why we might need to compare the before and the after VO changes.

Does that make sense ?

Edit also 2MD is m drive, 2MK is drivelogic are you sure that’s what you added ? You may want to check.

 

[Hpaula]

We need to isolate which modules are causing this and then which parameters are the ones responsible.
To find out which module. Take a trace of how the car is right now with all of the “correct” settings
Revert to stock, And then slowly add in what we need until the car starts exhibiting the weird shifts. Once the slow shifts appear go back to the last step and analyse what we changed last

Is not about tue coding, ones you flash the transmission with the gts flash hesitations starts , I thought it was coding at the beginning but no we put the car back to stock coding only gts flash it did the same , it looks like is shifting faster because the kombi does not have the correct coding and the needles move so fast but the shifting problems and hesitation always been there from 3 gear on

 

[aus335iguy]

If that’s the case @Dyezaks car is the odd one out... edit so is @TireDeathRacing ... have you noticed this laggy shift behaviour as well? What about you @Redi132kw ?

 

[derekgates]

When you change your VO and default a module it adds the parameters that you need for that option. So although you’ve only manually changed two or three parameters many more others could have appeared or disappeared.
That’s why we might need to compare the before and the after VO changes.

Does that make sense ?

Edit also 2MD is m drive, 2MK is drivelogic are you sure that’s what you added ? You may want to check.

Right. I wonder if that's something going on. There are .TRC files posted by @LMB335is ... is this something we can compare in those files? Something in the DSC module that needs to be changed? Is this even a DSC problem or the transmission unit itself?


And yes, I meant $2MK. I was previously wondering if I did need $2MD as well..

Is not about tue coding, ones you flash the transmission with the gts flash hesitations starts , I thought it was coding at the beginning but no we put the car back to stock coding only gts flash it did the same , it looks like is shifting faster because the kombi does not have the correct coding and the needles move so fast but the shifting problems and hesitation always been there from 3 gear on

Ah, so you've done the research already. I only backed the car out and back into garage, didn't drive (on just transmission flash).

I really wonder why @dyezak seemed to have a great first experience with fast shifts? Mine feels like a torque converter or CVT when it upshifts...



What about E92 M3 GTS DCT Flash 7848442 that I have seen mentioned? It was commented to be a later ZB for the GTS flash... Is this something we can try?

 

[dyezak]

Is not about tue coding, ones you flash the transmission with the gts flash hesitations starts , I thought it was coding at the beginning but no we put the car back to stock coding only gts flash it did the same , it looks like is shifting faster because the kombi does not have the correct coding and the needles move so fast but the shifting problems and hesitation always been there from 3 gear on

This is likely true. My initial test drives never got beyond 3rd gear.

 

[dyezak]

I really wonder why @dyezak seemed to have a great first experience with fast shifts? Mine feels like a torque converter or CVT when it upshifts...

I think it's exactly like @Hpaula stated. My initial test drives didn't have coding...and the KOMBI moves the tach needle incredibly fast without the coding...so there's that subconscious thing there. You see that needle move that fast and "obviously" the shifts are incredibly fast. Also, I never went above 3rd gear during aggressive acceleration, so I never saw the 4th gear + slow hesitations.

Move forward and I have all the coding done and I attribute this hesitation to the coding.

Obviously not the coding, it's something else.

 

[dyezak]

Could this be DSC coding? I didn't see anything else that was needing to be done

I don't think there's anything you can do with the DSC via coding to fix this. If you compare your 335 DSC TRC file to the posted M3 DSC TRC file you'll see what I mean. The M3 DSC is *completely* different than our DSC. Wildly different options in coding. That means the flash itself is massively different.

I will go pick up an M3 DSC tomorrow from my guy. I don't know how long it'll take me to get it installed however as it's fairly involved in the brake system.

 

[derekgates]

This is likely true. My initial test drives never got beyond 3rd gear.

I think it's exactly like @Hpaula stated. My initial test drives didn't have coding...and the KOMBI moves the tach needle incredibly fast without the coding...so there's that subconscious thing there. You see that needle move that fast and "obviously" the shifts are incredibly fast. Also, I never went above 3rd gear during aggressive acceleration, so I never saw the 4th gear + slow hesitations.

Move forward and I have all the coding done and I attribute this hesitation to the coding.

Obviously not the coding, it's something else.

Ah, this is something we all overlooked. My car feels great when starting, reversing, 1st/2nd gear and then it turns into a CVT.

I don't think there's anything you can do with the DSC via coding to fix this. If you compare your 335 DSC TRC file to the posted M3 DSC TRC file you'll see what I mean. The M3 DSC is *completely* different than our DSC. Wildly different options in coding. That means the flash itself is massively different.

I will go pick up an M3 DSC tomorrow from my guy. I don't know how long it'll take me to get it installed however as it's fairly involved in the brake system.

Oh yeah, you've mentioned that. It's a bit beyond my understanding how all of these modules interact. Is the common thought that the DSC unit, module flashing and coding will alleviate these issues? The DSC is in play when the transmission shifts?

 

[dyezak]

I want to see if @jyamona's custom DME tune where the engine ignores torque reduction requests from the DSC resolves this. If it does then we know it's the DSC, and we just have to figure out how to resolve it. If the @jyamona hack doesn't fix it then we start looking elsewhere. But right now I'm kind of stuck on the DSC for some reason.

 

[aus335iguy]

I think it’s final drive DSC related. If it gets worse at higher speeds it’s a giveaway. I was surprised that this ever worked without the change.

If true, There’s two ways to fix it. Change final drive and move to m3 DSC or try and fool our DSC into thinking speed is different.

Given that final drive in the 3.0 and over range would be less than ideal for daily drivers.

 

[LMB335is]

And this is why I think I'll get the M3 drivetrain installed before I do the shifter and coding changes. I can source a 2.93 gear for the M3 diff and I think I know where 210mm 2.81 and even 2.65 comes from if necessary, just finding one may be a pain.

Terry went from the 2.56 to 2.81 which is a 9.7% difference and the only two ratios offered on diffsonline for the 335 DCT and have no issues but the 3.15 does. Dropping to a 2.93 from a 3.15 is 7.5% and to 2.81 is 12.1%. Thing I want to know is where the final drive difference is seen in a M3. If it's a TCU/DSC thing I think we can get around it. If it's a signal from the DME then we may need Jake's help on that, shit maybe the tune Darrick's trying to get may do the trick. There is a wide range offered for the M3 DCT, from 2.81 to 3.62 so it seems the M3 wasn't nearly as sensitive to the FD change as the 335is but that perhaps the 2.56 falls outside that buffer zone at 23%. The 3.62 is at 15% lower. The 2.81 may be the sweet spot either way, not perfect gearing but livable.

 

[aus335iguy]

OK. Im home from work which interestingly is strikingly similar in a lot of ways to what were doing here but anyway... as im not in a comfortable space to go messing with my daily driver the only thing i can contribute with is organisation and methodical fault finding.(well i think that’s what I’m good at) If you guys are happy with that, that’s what ill contribute.
The objective is to quickly identify what the problem is and resolve it so we can get on with the business of going fast(er)

Below is what i think should happen from here. Im not the boss here and none of you are under any obligation to me or anyone else and if im treading on your toes please say so and ill back off. It’s your car not mine. Im just some crazy dude on the internet.

Problem
M3 GWS and flash conversions done on non M cars exhibit drawn out or laggy shifts

Hypothosis

1. Final drive ratio disparity is causing it(likely)
2. DSC is intervening during shift(less likely)
3. Torque limiters being hit (Least likely)

Propsed test for hypothosis 1
The only way to test the first one is to actually change final drive ratio. @Hpaula said he had a diff on the way so maybe he can run with that and let us know ?

Proposed test for hypothesis 2

By using @jyamona DSC intervention prevention(TM) flash we can test the second hypothosis fairly quickly. We could also install an M3 DSC.
@dyezak may have both of these covered as long as Jake is prepared to provide the flash and he gets a DSC.

Proposed test for hypothesis 3
Any one of the four people who have converted may already be running stock tunes or can easily switch back to stock to test.

What if that doesnt fix it?
I suggest we look at better logging. A CAN logger would be helpful in understanding the messages going back and forth.

Im going to see if i can grab that spare 335i DSC this weekend and have a crack at flashing it with M3
I need to figure some stuff out first... It’s a LOOONG shot though so dont get your hopes up. Ill also make some updates to the coding guide that needed doing.

 

[dyezak]

Well, hypothesis 3 isn't a hypothesis. During logging I do see torque reduction requests while I'm experiencing these lethargic shifts. Fact. The question is where are they coming from? The DSC as during a limited traction event? I do have DSC completely off but we all know there could be other things in the DSC module that might not be turned off when you fully disable it. Or is it coming from TCU due to final drive issues and the transmission trying to save itself?

I don't know yet but we will figure this out.

 

[Jake@MHD]

Hey guys. I can get the logging params to figure out where the torque reduction requests are coming from, and then after determining that can disable them if need be. However depending on the cause, it may not be best to disable them, will have to see.

Can someone compare the torque reduction request of a normal DCT shift vs. one of these laggy M3 GWS + software shifts?