[harkes]

Anyone in for donations to the first one figuring this out?
Would the solution be to just raise the limit or can clamping be increased like the DCT also?

Harkes: 100usd

[bradsm87]

Originally Posted by harkes

Anyone in for donations to the first one figuring this out?
Would the solution be to just raise the limit or can clamping be increased like the DCT also?

Harkes: 100usd

enemigo13 has already flashed the M135i calibration to a 335i and it works. 550nm limit in every gear instead of 440 in 3rd and 400 in 6th.

Martial at MHD is aware of the limit differences too. I'm sure a good guide will pop up any day now.

[harkes]

Originally Posted by bradsm87

enemigo13 has already flashed the M135i calibration to a 335i and it works. 550nm limit in every gear instead of 440 in 3rd and 400 in 6th.

Martial at MHD is aware of the limit differences too. I'm sure a good guide will pop up any day now.

550 limit isn't nearly enough running meth/ethanol. We need minimum 700 or basically just eliminate it.

[bradsm87]

Originally Posted by harkes

550 limit isn't nearly enough running meth/ethanol. We need minimum 700 or basically just eliminate it.

It's easy to under-report torque on the DME to get under limits. Under-reporting too far (like to get under the 400nm 6th gear limit if you're actually running ~630nm), then you have problems, but under-reporting ~630nm to get under a 550nm limit should be fine.

Any more than 650nm of real torque and the 8HP45 just isn't going to do well. That's a huge amount of torque. If you're genuinely pushing that much torque, time to look into an 8HP70 swap.

[afranke]

Originally Posted by bradsm87

It's easy to under-report torque on the DME to get under limits. Under-reporting too far (like to get under the 400nm 6th gear limit if you're actually running ~630nm), then you have problems, but under-reporting ~630nm to get under a 550nm limit should be fine.

Any more than 650nm of real torque and the 8HP45 just isn't going to do well. That's a huge amount of torque. If you're genuinely pushing that much torque, time to look into an 8HP70 swap.

What about the ActiveHybrid 3, we already have the 8HP70 (as the electric motor outputs torque). Any chance the TCU files for the AH3/8HP70 would work on the regular 335?

[bradsm87]

Originally Posted by afranke

What about the ActiveHybrid 3, we already have the 8HP70 (as the electric motor outputs torque). Any chance the TCU files for the AH3/8HP70 would work on the regular 335?

8HP70 files won't work on 8HP45. All the control pressures relating to the hardware would be very, very different.

[bradsm87]

Another thing we need to work out before flashing other car's stuff to our cars is what are the hardware differences between the car models?

2 different mechatronics units for 8HP45:
http://www.realoem.com/bmw/enUS/part?q=24347647840
http://www.realoem.com/bmw/enUS/part?q=24347647849

It seems like the 840 one is on the N55B30M0 (225kw) and the 849 on is on the N55B30O0 (235kw and 240kw) variants. One of the differences in the files is the system pressure EDS valve values are different between 335i/535i and 640i. Maybe the EDS valves in the mechatronics units are different between the M0 and O0 variants? If that's the case, flashing files from the other variant isn't ideal. Maybe a mechatronics unit swap from a higher power variant could allow more clutch holding pressure?

[enemigo13]

Originally Posted by bradsm87

Is that using E-Sys?

Did you need to re-code any other modules after flashing to make any errors go away or regain gear display or paddles or anything?

How is the M135i calibration? Any noticeable difference in shift points or firmness?

yes, with esys. after flashing i coded to my car FA and then options I want like launch control etc. did not notice any difference in gear changes, running this flash for almost year, currently with 1.1 bar of boost (stage 1 turbo).

[bradsm87]

I got some old version SWFL files for both the F30 335i (N55B30M0) and F20 M135i (N55B30O0) that are the same firmware version as the WinOLS project that I have. This makes it extremely easy to compare the differences between the flash files. It made it easy to define the later M235i differences too. The only differences between early version M135i and 335i ROMs are:


- Rear axle ratio
- Tyre circumference
- Torque limit per gear
- System pressure EDS valve calibration
- Track width of front axle. Important for lateral acceleration calculation.

The later versions have many more changes but that's probably just minor tweaks per-vehicle that happened later on.

The system pressure EDS valve calibration is the most concerning. BMW have likely fitted a different system pressure EDS valve to N55B30O0 models (hence the different mechatronics part numbers) for better power holding. It really isn't the best idea to flash a file made for one mechatronics to a different mechatronics. Maybe adaptations can counteract it? It might still be worth it to get around the torque limit. I don't know.

Here are both system pressure EDS valve calibration values overlayed. Y axis is system pressure. X axis is power through the EDS valve. Blue line is 335i. Green line is M235i. What this means is that if you flash an M235i file to a 335i, the file has values that assume it's using the M235i system pressure EDS valve, but you're actually running the 335i EDS valve. If the TCU requests say 10 bar system pressure, it will apply around 570mA to the EDS valve when in reality, it's only producing around 8 bar system pressure because it's actually the 335i system pressure EDS valve.

[enemigo13]

another (and most correct) way would be to find a way for engine ECU to do not listen to TCU limits. Similar problem was with older VAG 3.0tdi, where it was torque limited in 5th gear. But after setting some switches off in engine ecu, this gearbox limitation was bypassed. It can be not easy though, as MHD and bootmod3 still cannot do it, all sw versions logs I've seen so far use heavy load to torque tables decalibration for ECU/TCU to think that torque is ~400Nm. Which I don't like at all. I'm using -10% decalibration max at high loads (stock at low-mid load levels) along with +10% compensation to sport displays to show correct (not -10%) values.

Originally Posted by bradsm87

I got some old version SWFL files for both the F30 335i (N55B30M0) and F20 M135i (N55B30O0) that are the same firmware version as the WinOLS project that I have. This makes it extremely easy to compare the differences between the flash files. It made it easy to define the later M235i differences too. The only differences between early version M135i and 335i ROMs are:


- Rear axle ratio
- Tyre circumference
- Torque limit per gear
- System pressure EDS valve calibration
- Track width of front axle. Important for lateral acceleration calculation.

The later versions have many more changes but that's probably just minor tweaks per-vehicle that happened later on.

The system pressure EDS valve calibration is the most concerning. BMW have likely fitted a different system pressure EDS valve to N55B30O0 models (hence the different mechatronics part numbers) for better power holding. It really isn't the best idea to flash a file made for one mechatronics to a different mechatronics. Maybe adaptations can counteract it? It might still be worth it to get around the torque limit. I don't know.

Here are both system pressure EDS valve calibration values overlayed. Y axis is system pressure. X axis is power through the EDS valve. Blue line is 335i. Green line is M235i. What this means is that if you flash an M235i file to a 335i, the file has values that assume it's using the M235i system pressure EDS valve, but you're actually running the 335i EDS valve. If the TCU requests say 10 bar system pressure, it will apply around 570mA to the EDS valve when in reality, it's only producing around 8 bar system pressure because it's actually the 335i system pressure EDS valve.

[bradsm87]

Originally Posted by enemigo13

another (and most correct) way would be to find a way for engine ECU to do not listen to TCU limits. Similar problem was with older VAG 3.0tdi, where it was torque limited in 5th gear. But after setting some switches off in engine ecu, this gearbox limitation was bypassed. It can be not easy though, as MHD and bootmod3 still cannot do it, all sw versions logs I've seen so far use heavy load to torque tables decalibration for ECU/TCU to think that torque is ~400Nm. Which I don't like at all. I'm using -10% decalibration max at high loads (stock at low-mid load levels) along with +10% compensation to sport displays to show correct (not -10%) values.

If the DME ignores TCU torque lowering requests, I'd be worried that the TCU would go into some sort of safety mode if it's requesting a torque drop and not getting it. I'd also be worried that I'm losing some safety features such as the TCU requesting torque drop if it detects slip for example.

Hopefully someone can modify the logic to ignore RSA signature soon and we can start editing the TCU ROM. I believe it will be very hard to crack the RSA key so modifying the logic to ignore it is probably more realistic.

[BMWcurious]

If there was strong "lock up" at the TCU torque limits then the DME could be made to limit the reported torque to the TCU to that level, but the DME could use correct values internally and respond to a request to torque down from the TCU to below its limit. However, it would be better for the TCU to request and tolerate real values.

[bradsm87]

Originally Posted by BMWcurious

If there was strong "lock up" at the TCU torque limits then the DME could be made to limit the reported torque to the TCU to that level, but the DME could use correct values internally and respond to a request to torque down from the TCU to below its limit. However, it would be better for the TCU to request and tolerate real values.

I don't know of a way for the DME to only report torque lower to the TCU and still use correct values itself. The tables you modify lower the DME's perceived torque. This is normally OK but to lower it to keep under 400nm is too much. I think until we get around the TCU torque limit, the best approach is to lower DME torque to just below the 440nm 3rd gear limit and limit load target on the DME a further 10% in 6th gear only to get under the 400nm limit in 6th. I'd prefer to sacrifice a little power in 6th gear only to get better shifts in all gears by not lowering reported torque too much.


Alternatively, if you're really serious about holding maximum power, you could source a mechatronics unit from an M135i, M235i, 640i or 740i and fit that then flash it with the latest M135i ROM if you have an F3x or a 640i ROM if you have an F1x. That way you get both the 550nm limit in all gears AND you get correct system pressure.

[enemigo13]

lots of 3.0tdi going with 5th gear limit disabled for many years, some have sw where it’s disabled from factory. it’s not easy to crack rsa and you probably know that F series obd writing was not cracked but leaked some person sold bunch of btld files with correct rsa and these btlds are not checking calibration space rsa. probably they were used in bmw ag internally for testing/calibrating. if there will be such leak for 8hp gearboxes - no one knows.

Originally Posted by bradsm87

If the DME ignores TCU torque lowering requests, I'd be worried that the TCU would go into some sort of safety mode if it's requesting a torque drop and not getting it. I'd also be worried that I'm losing some safety features such as the TCU requesting torque drop if it detects slip for example.

Hopefully someone can modify the logic to ignore RSA signature soon and we can start editing the TCU ROM. I believe it will be very hard to crack the RSA key so modifying the logic to ignore it is probably more realistic.

[BMWcurious]

bradsm87, I was thinking of altering the code that puts torque on the two PTCAN buses, but only if it would be useful. There are a handful of torque values. You could have gear Vs torque in Vs torque out and tune it in realtime if you really want. The most important thing for me would be deciding on an approach with a realistic prospect of success. Would need to work out the communication from TCU to DME so that a 1000Nm DME variable is not still limited by a regular TCU torque request too.

I don't have an F30: my 8HP70 seems happy with 920Nm in an Audi.

[bradsm87]

Originally Posted by BMWcurious

bradsm87, I was thinking of altering the code that puts torque on the two PTCAN buses, but only if it would be useful. You could have gear Vs torque in Vs torque out and tune it in realtime if you really want.

I don't have an F30: my 8HP70 seems happy with 920Nm in an Audi.

I think it would definitely be useful for 3rd and 6th gears at the moment. It might pay to get an 8HP timeline from XHP first because if they will have an 8HP flash by end of the year, it might not be worth it. Are you thinking custom DME logic or a physical device?

[BMWcurious]

DME logic. It is what I do. If TCU can be done best to try that first.

[bradsm87]

Originally Posted by BMWcurious

DME logic. It is what I do.

That's awesome!

[BMWcurious]

I don't go by the same name on every forum, but you've seen my thread on another forum, apparently I like Lambos.

[bradsm87]

Originally Posted by BMWcurious

I don't go by the same name on every forum, but you've seen my thread on another forum, apparently I like Lambos.

Aaah I did wonder if it was you!

[bradsm87]

I spent another few hours decompressing and checking more files from random cars like F07, F15 etc and the solid pattern is that every ROM that has the higher torque limits also has the different system pressure EDS scaling which would cause lower system pressure than it should with our F30, F10 mechatronics.

[BMWcurious]

What processor do these use? How large is the total flash image? Is there an A2L matching any of them?