Discussion about issues encountered whilst converting to M3 Flash and GWS

[dyezak]

A thought.... how does the syvecs standalone deal with this issue? Whatever it is needs to be able to be fixed at the DME because they can use it in a supra to control the dct. The chances of the final drive being same as an m3 in that car are slim....i reckon they offer the DCT an altered wheel speed signal. for us we would need to do that but just for the DCT not the DSC.

Our DCT is getting it's data input from the LIN bus and the PT-CAN bus. These buses are multi-use data streams that many other modules use to consume data. RPM for instance comes from the DME and both your KOMBI and DCT use it. That means our capability of lying over the CAN bus(s) have other negative impacts.

During a full DCT transplant it is a direct PT-CAN connection from the Syvecs to the DCT. No other module uses the data that the Syvecs outputs. So you can have the Syvecs lie its ass off and manipulate the DCT into thinking whatever you want it to think. No negative impacts.

In one way, a full DCT transplant is actually easier than what we are dealing with.

 

[aus335iguy]

Our DCT is getting it's data input from the LIN bus and the PT-CAN bus. These buses are multi-use data streams that many other modules use to consume data. RPM for instance comes from the DME and both your KOMBI and DCT use it. That means our capability of lying over the CAN bus(s) have other negative impacts.

During a full DCT transplant it is a direct PT-CAN connection from the Syvecs to the DCT. No other module uses the data that the Syvecs outputs. So you can have the Syvecs lie its ass off and manipulate the DCT into thinking whatever you want it to think. No negative impacts.

In one way, a full DCT transplant is actually easier than what we are dealing with.

I used to think the same way and now I might be crazy

The can shield would need an input and an output. I now have the opinion that a can shield connected just before the tcu can take all of those messages and pass them all through....except the ones that relate to road speed we intercept those and alter them to suit our final drive. Essentially every module gets the right message, except one. Does that make sense?
The following diagram is how it works in my head and any resemblence between it and reality is purely coincidental



Am I crazy ?

 

[doublespaces]

Our DCT is getting it's data input from the LIN bus and the PT-CAN bus. These buses are multi-use data streams that many other modules use to consume data. RPM for instance comes from the DME and both your KOMBI and DCT use it. That means our capability of lying over the CAN bus(s) have other negative impacts.

During a full DCT transplant it is a direct PT-CAN connection from the Syvecs to the DCT. No other module uses the data that the Syvecs outputs. So you can have the Syvecs lie its ass off and manipulate the DCT into thinking whatever you want it to think. No negative impacts.

In one way, a full DCT transplant is actually easier than what we are dealing with.

Can bus doesn't have to be broadcast messages, I'm pretty sure you could send the altered data directly to the transmission node itself. Each can node will parse the can packet and look for a few things. If it's not addressed to them, it is ignored and not processed further. However in the case of a broadcast message or a message sent directly to the node, it will read the can message and process it accordingly based on the packet commands found inside this "can wrapper".

The trick would be to get the trans to ignore the other messages, or have the dme send this data out to each can node specifically, but modified to the trans alone.

If this data is not sent out by the dme, we would need to hack or replace the device that does. Potentially put a man in the middle interpreter box that would do this for us. A sophisticated yellow box if you will.

 

[doublespaces]

And even if sending the messages directly wouldnt work, I don't see why we couldn't talk to the transmission through an external box like syvecs. It would listen on the pt can and modify accordingly and be the sole communicator to the tcu.

 

[aus335iguy]

So... I’m not crazy...?

 

[NoQuarter]

So... I’m not crazy...?

No. You and doublespaces are correct.

 

[aus335iguy]

So, in order to figure this out we’d need to record the data stream from the various buses into and out of the TCU. Find the relevant ones relating to wheelspeed. Then we need to write a program to take these and change them to suit our final drive and voila! We have our solution.


Does anyone here know how to do that?

 

[dyezak]

Can bus doesn't have to be broadcast messages, I'm pretty sure you could send the altered data directly to the transmission node itself. Each can node will parse the can packet and look for a few things. If it's not addressed to them, it is ignored and not processed further. However in the case of a broadcast message or a message sent directly to the node, it will read the can message and process it accordingly based on the packet commands found inside this "can wrapper".

The trick would be to get the trans to ignore the other messages, or have the dme send this data out to each can node specifically, but modified to the trans alone.

If this data is not sent out by the dme, we would need to hack or replace the device that does. Potentially put a man in the middle interpreter box that would do this for us. A sophisticated yellow box if you will.

There are 2 types of messages. Broadcast and addressed. I know for a fact the RPM and Wheel Speed messages are broadcast. Those are the messages that will dictate your rear drive ratio options.

The rest of the messages may be addressed, I'm not sure (they probably are). But that doesn't matter as they don't apply to the problem at hand, being locked to a 3.15 gear ratio (+/- a small percent).

All that said @aus335iguy has a great concept for the CAN shield idea, because even broadcast messages can be intercepted and altered and the only impacted module is the transmission. This was one of my ideas when I bought my CAN shield, but I want to knock this single turbo conversion out before I commit any real time to looking into it.

 

[doublespaces]

So... I’m not crazy...?

No, I actually didn't read your reply, this was right before I crashed. But more or less, we're on the same track.

There are 2 types of messages. Broadcast and addressed. I know for a fact the RPM and Wheel Speed messages are broadcast. Those are the messages that will dictate your rear drive ratio options.

The rest of the messages may be addressed, I'm not sure (they probably are). But that doesn't matter as they don't apply to the problem at hand, being locked to a 3.15 gear ratio (+/- a small percent).

I'm not familiar with the 'can-bus shield' but it seems promising, but what does this yellr box do, is it can capable?

Also, addressed messages might apply, unless you are sure that the TCU does not respond to addressed wheel speed messages specifically? I don't think it would matter if it is a broadcast or addressed message. The issue comes from the need to send wheel speed to every other can node that requires it, in an addressed form. This adds latency, but you do not need a box or any hardware assuming its coming from the DME or another area we can write custom logic. If this wheel speed comes from another sensor or requires hardware, then the can-shield option becomes the best bet again.

 

[LMB335is]

Just got my car back with the M3 drivetrain installed with the 2.81 gear. So far, so good. I can't really get on it too hard until I get the diff broken in since it was essentially rebuilt, but just running through the gears at a brisk pace felt good. I'm dying to try that LSD though. I'll get some logs ASAP, especially of the higher gears, to compare with @dyezak and @Stokes and their 3.15 gear. No adaptions have been done yet, probably do that tomorrow time permitting. If it's still a little off after all this then fuck it, four yellow boxes it is.

 

[aus335iguy]

Just got my car back with the M3 drivetrain installed with the 2.81 gear. So far, so good. I can't really get on it too hard until I get the diff broken in since it was essentially rebuilt, but just running through the gears at a brisk pace felt good. I'm dying to try that LSD though. I'll get some logs ASAP, especially of the higher gears, to compare with @dyezak and @Stokes and their 3.15 gear. No adaptions have been done yet, probably do that tomorrow time permitting. If it's still a little off after all this then fuck it, four yellow boxes it is.

Great to hear you got it all in. does it feel as good as we think it does ?

 

[NoQuarter]

Great to hear you got it all in. does it feel as good as we think it does ?

Dude... Keep that kind of talk in a PM or something. I'm sure he will call later.

 

[Stokes]

Just got my car back with the M3 drivetrain installed with the 2.81 gear. So far, so good. I can't really get on it too hard until I get the diff broken in since it was essentially rebuilt, but just running through the gears at a brisk pace felt good. I'm dying to try that LSD though. I'll get some logs ASAP, especially of the higher gears, to compare with @dyezak and @Stokes and their 3.15 gear. No adaptions have been done yet, probably do that tomorrow time permitting. If it's still a little off after all this then fuck it, four yellow boxes it is.

Good to hear. I'll capture new logs tonight as it is shifting better each day. I wouldn't worry about adaptations, it'll adjust on its own as you break in the diff.

What lsd are you running? Assuming a clutch style if you are breaking it in.

I've been looking at the possibility of transplanting 2.93 gears over with a classic clutch style lsd. But it looks a bit complicated with the 215k/l case setup.

Btw, I'm 3.08, so would be nice to.see the difference across the 3 ratios.

 

[dyezak]

Hahahaha

 

[LMB335is]

Good to hear. I'll capture new logs tonight as it is shifting better each day. I wouldn't worry about adaptations, it'll adjust on its own as you break in the diff.

What lsd are you running? Assuming a clutch style if you are breaking it in.

I've been looking at the possibility of transplanting 2.93 gears over with a classic clutch style lsd. But it looks a bit complicated with the 215k/l case setup.

Btw, I'm 3.08, so would be nice to.see the difference across the 3 ratios.

Stock M3 M variable LSD. I used all new bearings and my diff guy said the ring looked like it was new, no wear whatsoever so he said to just go easy on it for a few hundred miles and I'll probably change the fluid and let 'er rip. Drivetrain felt super smooth, no vibration, even with black powerflex diff bushings.

 

[Stokes]

New log for 3.08 gearing. Hopefully no spinning occurred. Three pulls.

Light load in D (level 3)
Moderate load in M (level 6)
Heavy load in M (level 6)

https://datazap.me/u/cstokes54/mdct-308-2nd-log

 

[Stokes]

Stock M3 M variable LSD. I used all new bearings and my diff guy said the ring looked like it was new, no wear whatsoever so he said to just go easy on it for a few hundred miles and I'll probably change the fluid and let 'er rip. Drivetrain felt super smooth, no vibration, even with black powerflex diff bushings.

@LMB335is you get that 2.81 broken in yet? I'm learning a lot about these BMW diffs and not liking the conversion options. To go 210 LSD requires different half shafts (flanges root cause). Plenty of 215l/k gear sets but no factory clutch LSD units that I can find exist. It seems like a fix BMW created to keep LSD units out of non-m cars. M factory and wave track still an option, but was hopeful for a 1.5 way clutch that isn't $2800.

I'll probably pick up a 2.81 and 2.93 in the 215 size to try out.

 

[LMB335is]

Well, I have an interesting error that I can't clear that's related to the DKG module (mechtronics in the transmission) expecting some data from the DSC and not getting it...and the most interesting thing is that it looks like the message is supposed to be routed through the KOMBI:

DKG00CF21 Message (raw data for longitudinal accelleration) error from DSC, receiver DKG/EGS, transmitter DSC/KOMBI

ISTA+ is pretty neat...but I'm wondering how much of this is useful.

Hey Darrick, what did you do that ended up clearing this error? It's the only one I'm getting and ISTA doesn't really address a fix for it unless I'm overlooking something. I'm installing the M3 DSC/SZL next week FWIW. Did that address it?

 

[LMB335is]

@LMB335is you get that 2.81 broken in yet? I'm learning a lot about these BMW diffs and not liking the conversion options. To go 210 LSD requires different half shafts (flanges root cause). Plenty of 215l/k gear sets but no factory clutch LSD units that I can find exist. It seems like a fix BMW created to keep LSD units out of non-m cars. M factory and wave track still an option, but was hopeful for a 1.5 way clutch that isn't $2800.

I'll probably pick up a 2.81 and 2.93 in the 215 size to try out.

Now I'm waiting for this sub tropical storm to move up the coast so it will dry out enough for me to try it. That's why I just went with the M3 drivetrain, yes it added $2k to the bottom line but it included the axle and driveshaft upgrade. No need for aftermarket axles and driveshaft which would be much more than the M3 pieces and offers a huge selection of gear ratios and LSD options if the M variable LSD doesn't cut it for you. The 215 pumpkin just doesn't have many options available that I could find. I can find a 2.81, 2.93, or a 3.07 in 210mm right now for $200-300 and they'd all bolt right in my M3 diff.

 

[dyezak]

Hey Darrick, what did you do that ended up clearing this error? It's the only one I'm getting and ISTA doesn't really address a fix for it unless I'm overlooking something. I'm installing the M3 DSC/SZL next week FWIW. Did that address it?

Once I did the DSC/SZL I didn’t have the error any more.