Discussion about issues encountered whilst converting to M3 Flash and GWS

[Jake@MHD]

The DME will take the straight 1M bin flash via MHD or WinKFP if no special logics/conversions are needed (3-column scaling, etc.)?

Nope, flashing any rom type other than what came with your car will give 2FA4 and your car will either shut off injection or give you like a 60 load cap.

MAP conversion table shouldn't be too hard to find in the bin if needed for 3.5 bar.

That is all defined in every XDF. I think he just meant it's easier for someone running a basically stock tune to switch because you don't have to manually transfer over all of your table changes.

What all is involved in switching to IKM0S?

I have done IJ -> IN for a few of Terry's customers. The switch to IK is the same process. I need to make a custom MHD build for your VIN, and the tweaks are applied behind the scenes. Due to the ZUSB change, some of the auto-niceties of MHD won't work as expected like return to stock, etc. Also, flashing *.mhd files and OTS maps which apply delta changes to your oem ZUSB file may be missing certain regional settings if you don't use an IN or IK base bin from your same region. For those reasons this process remains a one-off I have done on a case by case basis.

Will PM you back shortly.

 

[RSL]

No problem, I tune myself, so no issue starting with whatever the base rom is. I do use 3 column, but no problem going back to standard 2 column for now extended period. I have a 3.5 bar too, but can switch back to the stock 2.5 bar if needed.

I saw the MAP conversion was defined in the XDF, hadn't looked at that XDF in like 4 years lol

As far as the M3 GSW is concerned and IKM0S seemingly working better, right off the bat, I see the L2T on IKM0S is significantly lower (probably because they're all MT), which is what I planned to do on Stokes N55 bin, which has not yet exactly happened as easily as I anticipated. Torque limiters on that thing just won't go away.

 

[Jake@MHD]

You can continue to use the 3-cell, that is no problem. Maybe for the first basic test just run the 2.5 bar for a more straightforward boost to load relationship. We'll talk next week

 

[RSL]

You can continue to use the 3-cell, that is no problem. Maybe for the first basic test just run the 2.5 bar for a more straightforward boost to load relationship. We'll talk next week

Sounds good, I'll reinstall the 2.5 bar in the meantime.

 

[doublespaces]

Oh, I think this needs a new thread! @RSL

 

[Stokes]

D mode is awesome. Once I reset DME Variants and adaptations the shifting went to shit. I've put 50mi on it and the 1-2 shift is pretty good again. I'll put a couple hundred miles on it and see what happens with the M mode shifting. Right now it's not good but I'm seeing if it has to do with adaptation.

Ditto. I've put a few hundred miles on and R, 1, 2, engagement feels good, but everything else is soft. As @RSL mentioned, we've been trying some things, but the N55 is a different base to work from. Trying to resolve this on a N54 where tuning knowledge is more mature might be easier.

 

[Stokes]

Oh, I think this needs a new thread! @RSL

Maybe. I believe this is still part of the "resolving the issues" when converting. The hardware and coding was the easy part (minus direct wiring that @dyezak incorporated for a complete swap).

On a related note .. Anyone have Testo working for the GS40? Or can you get torque requests from the TCU through the DME with Testo?

 

[dyezak]

On a different note I did got my Arduino CAN sniffer running. I have very little Arduino knowledge but it wasn't too hard. I then decided I needed to capture what I was seeing in a more manageable format....so I googled around and decided to touch my toe into python coding. I now have a python app running on my laptop that is sniffing my JBB merged CAN BUS. Still have a kink to work out on the arduino side...but it is fun/cool.

 

[NoQuarter]

On a different note I did got my Arduino CAN sniffer running. I have very little Arduino knowledge but it wasn't too hard. I then decided I needed to capture what I was seeing in a more manageable format....so I googled around and decided to touch my toe into python coding. I now have a python app running on my laptop that is sniffing my JBB merged CAN BUS. Still have a kink to work out on the arduino side...but it is fun/cool.

Reach out for any canbus sniffing stuff if you need it. Maybe I can find something from what I did a couple years ago.

I am a bit rusty for sure but getting back into the swing of things.

 

[doublespaces]

On a different note I did got my Arduino CAN sniffer running. I have very little Arduino knowledge but it wasn't too hard. I then decided I needed to capture what I was seeing in a more manageable format....so I googled around and decided to touch my toe into python coding. I now have a python app running on my laptop that is sniffing my JBB merged CAN BUS. Still have a kink to work out on the arduino side...but it is fun/cool.

What is your primary coding experience if you don't mind me asking?

 

[dyezak]

What is your primary coding experience if you don't mind me asking?

Coding? Hahahaha

I guess the only thing I know is SQL...but that's not coding. I know nothing of coding but I'm not skared of n-e-thing.

 

[RSL]

Oh, I think this needs a new thread! @RSL

Ditto. I've put a few hundred miles on and R, 1, 2, engagement feels good, but everything else is soft. As @RSL mentioned, we've been trying some things, but the N55 is a different base to work from. Trying to resolve this on a N54 where tuning knowledge is more mature might be easier.

+1. Bulk of what's going on right now is just figuring out WTF on N55 vs. N54 that I'm used to. Part of it is that N55 is much different than I could've ever guessed just looking the XDF and part is that info that is out there isn't exactly what I'd call extensive or conclusive. So, even though we're testing things, at this point, it's mostly a matter of figuring it out and seeing how things respond, if at all.

Still very much at step 0.5, so there's not much to say, but will make a separate thread at some point. If nothing else, it might increase the collaboration factor and decrease the learning curve timeframe.

 

[Stokes]

Well, I'm learning the hard way...

For Testo: GS19 does work to communicate, sort of. Ends up a lot of the parameters are available in the DME jobs anyway, so the correct GS program didn't matter, but I yet have figured out how to set up a custom job with the parameters I'd like. The learning curve is quite steep, but once I figure it out, it looks very promising. I found most of the parameters needed to log, but it errored out when I tried to connect and monitor.

I switched to a new VM with BimmerGeeks install and I have to say it's much nicer, but didn't fix my error with using Testo. I'm sure it's a user error. I'll try again tomorrow.

By the way, I'm still getting a persistent CF21 error in the TCU. Seems to be a missing CAN message from KOMBI. Not finding a more descriptive message. Stays even with a clear. Anyone else with this?

CF21

CAN message from instrument cluster: Mileage range

Signal time error

[TBODY] [/TBODY]

 

[dyezak]

Well, I'm learning the hard way...

For Testo: GS19 does work to communicate, sort of. Ends up a lot of the parameters are available in the DME jobs anyway, so the correct GS program didn't matter, but I yet have figured out how to set up a custom job with the parameters I'd like. The learning curve is quite steep, but once I figure it out, it looks very promising. I found most of the parameters needed to log, but it errored out when I tried to connect and monitor.

I switched to a new VM with BimmerGeeks install and I have to say it's much nicer, but didn't fix my error with using Testo. I'm sure it's a user error. I'll try again tomorrow.

By the way, I'm still getting a persistent CF21 error in the TCU. Seems to be a missing CAN message from KOMBI. Not finding a more descriptive message. Stays even with a clear. Anyone else with this?

CF21

CAN message from instrument cluster: Mileage range

Signal time error

[TBODY] [/TBODY]

Did you re-code your KOMBI? Are you seeing the Big D/M like an M3 or the little D/S like the 335?

 

[Stokes]

Did you re-code your KOMBI? Are you seeing the Big D/M like an M3 or the little D/S like the 335?

Re-coded. I did a read from the kombi first, then coded for my VO. Did a file diff and brought back in just a couple settings, like Motorsport. Most of the deltas were for the fuel gauge. Start/stop was already off and didn't get turned on when default coding for the E82.

I looked again and the German words say it's a warning, not enough to set an error or something like that. Sounds like no one else has it. I'll look at the kombi coding again.

 

[dyezak]

Uh, guys...

On a full factory 335is we have a sport button. And yes, that changes the DCT behavior. But it also does something to the DME and the engine tune as well, and the increased sportiness in the DCT shifts work in concert with the DME changes. I know this because MHD can enable burble with and without the sport button depressed.

Now that we don't have a sport button, what the hell are we missing on the DME side that may be impacting our shifts and exaggerating this slow soft shifting situation?

Should I get a test map with all the "sport" tables moved to my regular tables to test this out?

@jyamona @V8bait @doublespaces

 

[RSL]

There may be more behind the scenes things, but at the very least, it changes throttle mapping (torque request vs. pedal angle) and potentially line pressure. Not sure if the M3 TCU flash raises the line pressure by default. Anyone know if that can that be logged with anything? ISTA or Testo maybe?

It would take all of 5 seconds to move the sport throttle values to the normal throttle table if the map weren't locked. If it is locked, you'll need your tuner to do it for you.

 

[dyezak]

There may be more behind the scenes things, but at the very least, it changes throttle mapping (torque request vs. pedal angle) and potentially line pressure. Not sure if the M3 TCU flash raises the line pressure by default. Anyone know if that can that be logged with anything? ISTA or Testo maybe?

It would take all of 5 seconds to move the sport throttle values to the normal throttle table if the map weren't locked. If it is locked, you'll need your tuner to do it for you.

I'm going to do some logging runs this afternoon.

 

[Jake@MHD]

Yea sport mode doesn't change a whole lot in the DME on N54 cars. The main one as already mentioned is the throttle mapping. Also, on IK and IN I believe there is a slightly different WGDC adder table used somewhere. Other than that it's mostly negligible, and nothing that would effect torque output sent to the TCU. It's more likely that there could be different tables in the TCU used, if it even sees the sport mode variable. I'll have to double check if it gets sent.

 

[dyezak]

I don't know what happened to the rest of the log...but you can see in this log from today my 1-2 shift is back to being good:

https://datazap.me/u/dyezak/log-1524071595?log=0&data=0-3-12-17

That shows a .2sec shift time at 19% throttle. Seems right. I did a pull that somehow didn't get captured. I'll do another one shortly.