Public Post Development

May 4, 2017
1,759
Plano TX
#61
Yea, the TCU only sends the RPM request for launch control and downshifts. So I don't think there's anything I could do on dme side. Its too bad it doesn't seem to gets rpm signal from dme either. You could yellow box all the wheel speed signals going to the TCU but that might have some negative side effects.

Isn't M3 redline 8300? What would the max mph be at 8k with a taller tire vs. 8300?
You're looking at +/- 203mph depending on tire size.
 
Nov 18, 2017
780
#63
The coding guide remains as it was originally as far as I know
 

Stokes

Corporal
Jan 26, 2018
118
Houston, TX
#64
No other ratio will upshift properly. And the 3.15 comes with quite the low mph ceiling for anyone trying to 1/2mi
I've decided to experiment with a 3.08 ratio. Couldn't get decent pricing on a M3 diff right now. This one is bolted per the diff part #, so I can always sell it quick if it doesn't work out. Going to see if a short tire and this ratio works out and fingers crossed someone decides the M3 TCU is worth cracking in the mean time.
 
Feb 23, 2018
74
NW FL
#66
Crap, well... if there is any word on 3.15 rear diff, half-shafts and driveshaft... include me on the list of interested buyers. :sweatsmile:

The 335is software is driving me crazy and I tried my best to NOT consider a M3 rear end but it's the only hope for a fix with the software correction attempted but with no luck. :(

Thanks to the community for the great work on this investigation. I really appreciate the dedication!
 
Feb 23, 2018
74
NW FL
#68
What's wrong w/ the 335is software?
Hesitations are the main issue. Most of the time trying to accelerate from a 'slow down' (~10 MPH) the DCT refuses to allow power to be delivered to wheels. Takes 1-3 seconds to 'go' most times after slowing down, such as when leaving a business parking lot (slowing down before entering highway) or slowing down for a light that suddenly turns green. Can't trust the car to move!

The 'crawling' mode of the 335is vs the M3 DCT logic is also irritating. I daily my car in stop and go'ish traffic and the 'crawling mode' transitioning into 'driving move' usually jerks the car around, causing clunks going back into gear, etc...

D mode is usually lethargic, D without SPORT is far too feathered in between gears and likes to upshift at measly RPMs. D with SPORT is barely better; the RPM doesn't gain much but comfort is drastically reduced. D with SPORT has great throttle response but the car doesn't want to actually move or downshift when power is requested (although once finally IN gear it will respond OK). If driving normally, the car wants to be in D7 even at 40MPH... I want to accelerate and the car refuses to shift THEN shifts 2-3 gears down and ZOOOOOOM, wasting many seconds to go through this affair.

S mode is great for me but the throttle isn't sensitive enough. S mode with SPORT is the most annoying drive mode I can imagine. Like a horny bunny, the car wants to force high RPMs, twitchy as hell.... It also LOVES to shift in the middle of turns such as when turning at a stop light, upsetting the car.

Predictability is my main issue. I can't trust the car to do what I want. :(


[EDIT]: Forgot to mention. M mode is usually PERFECT. When shifting on my own, the car is great. Sometimes it doesn't respond quickly (takes a second or two to shift up/down) but 90% of the time M mode is DCT perfection. I KNOW the hardware can do it! :)
 
Nov 18, 2017
780
#69
Hesitations are the main issue. Most of the time trying to accelerate from a 'slow down' (~10 MPH) the DCT refuses to allow power to be delivered to wheels. Takes 1-3 seconds to 'go' most times after slowing ..........
[EDIT]: Forgot to mention. M mode is usually PERFECT. When shifting on my own, the car is great ............I KNOW the hardware can do it! :)
I feel the same way about mine. And i have a similar use case. That’s why im here trying to find something sportier....I Thought M3 adjustability might be the answer.
I find that after a few hundred Kay’s of manual driving that the s-mode works better. The adaptations on these boxes confuse also because just when you think that you can predict what it’ll do It’ll change.
 

RSL

Sergeant
Aug 11, 2017
275
#70
It's the same on an actual M. If you want it to downshift itself, it's going to take 2 secs and then be off like a shot. If it's on the same clutch (even and odd gears) might be a little quicker, but not much. It's an unnerving feeling during emergency driving. Drive in M mode all the time was my answer, plus more fun.

If you find a mode that mostly fits the bill for you and just isn't sensitive enough, could do some torque request changes in the DME to give it a little more pep. If you drive it more aggressively though, DriveLogic should respond.
 
Feb 22, 2018
11
US
#71
Well, looks like 1 series tire diameter and a 3.08 would work fine. Couple options.
What are you trying to achieve with different wheel size? Wheel size doesn't have anything to do with speed calculations. FDR has because it is "before" ABS wheel speed sensors.

Yeah there’s an RPM sensor on the input shaft according to the documentation I read.

Edit There’s 3 sensors. one on the input shaft, one on input shaft 1 post clutch and input shaft 2 post clutch.
It was discussed here https://bmw.spoolstreet.com/threads/bmw-dct-information.2950/

I think the reason the call routine for RPM during launch can be altered and this cant is because when launch is initiated the box is still waiting for the correct RPM and you can launch in that period. The shift behaviour requires RPM in a narrow band.

Sadly I dont think that DME logic is likely to fix this TBH but @jyamona youd know more what do you think ?
Guys, FDR has nothing to do with performing shifting per se. Transmission has its own sensors to determine slip or target RPMs.
The problem could come from shift scheduling (to early shifts) and engine stall protection function which could be implemented based on vehicle speed calculated from ABS.
Question is when does the slippage happen, low engine RPM only or also high RPMs? Maybe lowering idle speed level in MHD flash (beta MHD) could help if it is stall protection kicking in. Maybe engine idle speed reported on the CAN from DME?

I think, the ultimate solution could come down to editing BIN file. I tried finding known FDR calibration in some Jeep software using WinOLS and change it. In that software it is only used for vehicle mass estimation. Here it seems to be involved more.
I don't know how to get the BIN :( And whether, after changing the value checksums would need to be recalculated or something.

Hesitations are the main issue. Most of the time trying to accelerate from a 'slow down' (~10 MPH) the DCT refuses to allow power to be delivered to wheels. Takes 1-3 seconds to 'go' most times after slowing down, such as when leaving a business parking lot (slowing down before entering highway) or slowing down for a light that suddenly turns green. Can't trust the car to move!

The 'crawling' mode of the 335is vs the M3 DCT logic is also irritating. I daily my car in stop and go'ish traffic and the 'crawling mode' transitioning into 'driving move' usually jerks the car around, causing clunks going back into gear, etc...

D mode is usually lethargic, D without SPORT is far too feathered in between gears and likes to upshift at measly RPMs. D with SPORT is barely better; the RPM doesn't gain much but comfort is drastically reduced. D with SPORT has great throttle response but the car doesn't want to actually move or downshift when power is requested (although once finally IN gear it will respond OK). If driving normally, the car wants to be in D7 even at 40MPH... I want to accelerate and the car refuses to shift THEN shifts 2-3 gears down and ZOOOOOOM, wasting many seconds to go through this affair.

S mode is great for me but the throttle isn't sensitive enough. S mode with SPORT is the most annoying drive mode I can imagine. Like a horny bunny, the car wants to force high RPMs, twitchy as hell.... It also LOVES to shift in the middle of turns such as when turning at a stop light, upsetting the car.

Predictability is my main issue. I can't trust the car to do what I want. :(


[EDIT]: Forgot to mention. M mode is usually PERFECT. When shifting on my own, the car is great. Sometimes it doesn't respond quickly (takes a second or two to shift up/down) but 90% of the time M mode is DCT perfection. I KNOW the hardware can do it! :)
I know the feeling. Do you know whether M3 software fixes any of those issues ?
 

doublespaces

Administrator
Oct 18, 2016
6,758
AZ
#72
Guys, FDR has nothing to do with performing shifting per se. Transmission has its own sensors to determine slip or target RPMs.
Sure it does, because changing the fdr is causing issues with those sensors.

dzid_ said:
I think, the ultimate solution could come down to editing BIN file.
That is the problem, currently you cannot just change the bin file, which is why we are talking about the various sensors so that the values can be modified to appease the TCU. We are looking for solutions which we can actually attempt. We can talk changing the BIN all day long, but if you can't modify the BIN, that talk is pointless.
 
Feb 22, 2018
11
US
#73
Transmission doesn't have to have anything connected to the output and still perform the shift. Doesn't need to care about FDR.

FDR doesn't mess with transmission sensors in anyway. That's nonsense.
If anything it messes up with calculation from wheel speed (ABS) sensors.
Some of the safety/protection/scheduling (where accuracy is not required) can be based off ABS sensors, (but not shifting itself).

If someone can give an information on whether mushiness happens at all time or only low RPM, then we have a trace :)
 
Feb 23, 2018
74
NW FL
#74
Transmission doesn't have to have anything connected to the output and still perform the shift. Doesn't need to care about FDR.

FDR doesn't mess with transmission sensors in anyway. That's nonsense.
If anything it messes up with calculation from wheel speed (ABS) sensors.
Some of the safety/protection/scheduling (where accuracy is not required) can be based off ABS sensors, (but not shifting itself).

If someone can give an information on whether mushiness happens at all time or only low RPM, then we have a trace :)
Well, @dyezak is the originator of this mod and he, like the rest of us, had mushy shifts. The solution was to swap in the 3.15 M3 diff (FDR). The car shifts like lightning now.

This thread encompasses our investigation:
https://bmw.spoolstreet.com/threads...d-whilst-converting-to-m3-flash-and-gws.3011/

This page is when the rear end was installed:
https://bmw.spoolstreet.com/threads...g-to-m3-flash-and-gws.3011/page-32#post-41347

This page is where @dyezak compared FDR with what the car is expecting vs 2.56 output:
https://bmw.spoolstreet.com/threads...g-to-m3-flash-and-gws.3011/page-27#post-40190

Beyond that... the shifts are bad no matter what. I can't find exact posts that describe, but scan around in there. :)
 
Feb 22, 2018
11
US
#75
You guys went through a lot !

The eureka moment put more light into the problem:

I think what's happening is TCU internal signal replacement due to implausibility. One thing to check would be to disconnect ABS speed sensors. This way implausible trans output speed signal should not be replaced with another implausible signal. I think this would be interesting test.

Btw, I think the idea of converting target speed for the DME to revmatch to correct RPM will not work. Maybe for the manual downshift it would work where rev matching is used... Normal upshifts and downshifts rates are controlled using clutches. The engine torque reduction for upshifts is just to make shift faster and smooth.

It would be really good to get some knowledge on how to flash modify the BIN. Do we even have the BIN? I have a feeling they were in some different format..
 
Nov 18, 2017
780
#76
You guys went through a lot !

The eureka moment put more light into the problem:

I think what's happening is TCU internal signal replacement due to implausibility. One thing to check would be to disconnect ABS speed sensors. This way implausible trans output speed signal should not be replaced with another implausible signal. I think this would be interesting test.

Btw, I think the idea of converting target speed for the DME to revmatch to correct RPM will not work. Maybe for the manual downshift it would work where rev matching is used... Normal upshifts and downshifts rates are controlled using clutches. The engine torque reduction for upshifts is just to make shift faster and smooth.

It would be really good to get some knowledge on how to flash modify the BIN. Do we even have the BIN? I have a feeling they were in some different format..
Disconnecting speed sensors was thrown up as an idea but I don’t think anyone tested it.
Don’t think anyone’s ever seen a BIN for the GS40?
 

RSL

Sergeant
Aug 11, 2017
275
#77
I don't know anyone that has an actual bin, but an option may be 0da/0pa merge if we can determine the correct offsets and potentially write changes back through WinKFP. Not sure I'd want to be the guinea pig for that though lol
 

Stokes

Corporal
Jan 26, 2018
118
Houston, TX
#78
Transmission doesn't have to have anything connected to the output and still perform the shift. Doesn't need to care about FDR.

FDR doesn't mess with transmission sensors in anyway. That's nonsense.
If anything it messes up with calculation from wheel speed (ABS) sensors.
Some of the safety/protection/scheduling (where accuracy is not required) can be based off ABS sensors, (but not shifting itself).

If someone can give an information on whether mushiness happens at all time or only low RPM, then we have a trace :)
Mushiness is mostly in higher gears (higher speeds). Seems less pronounced at lower rpm. For the test removing the speed sensors, would all 4 need to be disconnected to conduct a valid test?

Being that the 3.15 ratio resolved the shifting, I'm making the assumption that wheel speeds started to match. The factory tire size is slightly shorter than the M3, so that combined with a 3.08 ratio gives something equivalent to a sensor reading as if a 3.28 ratio FD is in place for a M3.

245/35/18 + 3.08 vs. 265/40/18 + 3.15.

If it proves out, it'll add an alternative to the M3 Diff... the manual e90/e82 non-M diff with a MFactory or equivalent LSD.
 
Feb 22, 2018
11
US
#79
The problem is that with tyres you are not changing the ratio between wheel sensor and the transmission sensor. And this is where the calculation mismatch occurs.
So even though the car will physically accelerate faster, it still will be using mismatched FDR when comparing those sensors. So you may still have not perfect shifts. But there is a chance that 3.08 is within some sort of margin. Or it can be that only the highest gears will fail, because mismatch is then "amplified". Good luck!


As for removing sensors, I think, all of them would have to be disconnected to be sure. It's all ISO26262, so faulty signals are replaced for safety reasons.
 

Similar threads

Top