Euro MDM Coefficient of Friction

[d0o0fy]

Hi all,

So I coded my car to Euro MDM (US_VEHICLE->nicht_aktiv, MDYNAMIC_MODE_USA->nicht_aktiv) recently and noticed that the coefficient of friction values for the oversteering can take multiple values (M3_UEBERSTEUERN_MUE_0 and M3_UEBERSTEUERN_MUE_1_2_M).

NCS Dummy has the following Translations for M3_UEBERSTEUERN_MUE_0:
wert_00: !(E90+(S65B40+US,S65B44+!US+MAN_UMF_GTS),(E92,E93) +S65B40)
wert_01: E90 + S65B40 + US
wert_02: E92 + S65B40 + !US + !MAN_UMF_GTS
wert_03: E92 + S65B40 + US
wert_04: E93 + S65B40 + !US
wert_05: E93 + S65B40 + US
wert_06: E92 + S65B40 + !US + MAN_UMF_GTS
wert_07: E90 + S65B44 + !US + MAN_UMF_GTS

And for M3_UEBERSTEUERN_MUE_1_2_M:
wert_00: !(E90+(S65B40+US,S65B44+!US+MAN_UMF_GTS),(E92,E93) +S65B40)
wert_01: E90 + S65B40 + US
wert_02: E92 + S65B40 + !US + !MAN_UMF_GTS
wert_03: E92 + S65B40 + US
wert_04: E93 + S65B40 + !US
wert_05: E93 + S65B40 + US
wert_06: E92 + S65B40 + !US + MAN_UMF_GTS
wert_07: E90 + S65B44 + !US + MAN_UMF_GTS

I left my car to its default value of wert_03 (US ZCP E92) but I'm wondering if to make the real switch to Euro MDM I should not also switch those two parameters to wert_02.

I'm also intrigued by that MAN_UMF_GTS and maybe switching to wert_06 to get the GTS friction coefficients? although those might be based on stickier and wider tires (not sure).

What might be useful would be for people with different spec'd cars to read their DSC modules (MK60_M3.C08) and post their trace files here so we can compare them. As an idiot I didn't make a copy of the full trace for my own car, but I'll post it later today. Ideally an GTS owner could chime in

Disclaimer: Use at your own risk. I'm not responsible for any consequences if you damaged your vehicle as a result of coding.

[terahertz]

This is good info... I'm going to have to test some of these. I don't like how DSC cuts power on the 1-2 shift.

[gazosnic]

wow thanks for the info...sounds like we may be able to fine tune the DSC limits a bit more. Curious to see how the GTS settings would be.

[d0o0fy]

Here is my trace file from the DSC module, in the state it was before modifying it:

RPA
	nicht_aktiv
MSA
	aktiv
M3_BAUART
	wert_08
M3_GETRIEBE
	wert_00
M3_UEBERSTEUERN_MUE_0
	wert_03
M3_UEBERSTEUERN_MUE_1_2
	wert_03
MDYNAMIC_MODE_USA
	aktiv
M3_PRE_MSR
	nicht_aktiv
LENKWINKELKENNLINIE
	wert_02
	wert_03
LAT_EBD
	aktiv
US_VEHICLE
	aktiv
POHO
	nicht_aktiv
DDS_PLUS
	nicht_aktiv
ABSCHALTBARKEIT_DSC
	aktiv
DRUCKMODELL_HA
	wert_00
DRUCKMODELL_VA
	wert_00

I since changed the mue coefficients to wert_02. I would be curious to see what the default values are for other M3 configurations.

[BPMSport]

The ! is an operand meaning "not".

02 would likely be the standard setting for euro spec, non GTS vehicles.

Selecting 02 will also be like selecting 06 or 07 - the data byte for each one of those is the same in hex, "CD".

[d0o0fy]

Mike,

Here are the values for the other settings:
wert_00: 0x00 (0)
wert_01: 0x66 (102)
wert_02: 0xE6 (230)
wert_03: 0x33 (51)
wert_04: 0x26 (38)
wert_05: 0x7F (127)
wert_06: 0xE6 (230)
wert_07: 0xE6 (230)

Now the interesting bit, there is an equation that seems to govern the value of coefficient used (in pseudo code):

if value < 128 then

coefficient = 2.5 + (value/64)

else

coefficient = 2.5 + ((value-256)/64)

Here are the coefficient of friction that are associated with each value:

wert_00: 2.5
wert_01: 4.09375
wert_02: 2.09375
wert_03: 3.296875
wert_04: 3.09375
wert_05: 4.484375
wert_06: 2.09375
wert_07: 2.09375

So when you go from wert_03 (51) to wert_02 (230) your coefficient goes from 3.296875 to 2.09375 and therefore decreases.

Now it seems that the available values are limited to those 6 here. But I am almost certain that it is possible to fine tune the coefficients by modifying the nettodat directly. Because of the way the equation works, the minimum coefficient possible is not obtained at the maximum value (0xFF) but at 128 (0x80). That would lead to a coefficient of 0.5! After that it's direction the skidpad and see how much of a difference those values make.

[terahertz]

I set wert_07 and euroMDM and it's pretty cool! I was able to drift out of the car wash with 90 degrees of opposite lock, and it pretty much held the drift for me.

I need to set it back to wert_01 and see what the difference feels like for a direct comparison.

[tdott]

Hmm, awesome info thanks for posting. Will try playing around with it.

[tdott]

I changed mine to:
M3_UEBERSTEUERN_MUE_0
wert_02
M3_UEBERSTEUERN_MUE_1_2
wert_02

After executing then reading it for confirmation, I got this:
M3_UEBERSTEUERN_MUE_0
wert_02
wert_06
wert_07
M3_UEBERSTEUERN_MUE_1_2
wert_02
wert_06
wert_07

Did you got that also?

[d0o0fy]

Originally Posted by tdott

I changed mine to:
M3_UEBERSTEUERN_MUE_0
wert_02
M3_UEBERSTEUERN_MUE_1_2
wert_02

After executing then reading it for confirmation, I got this:
M3_UEBERSTEUERN_MUE_0
wert_02
wert_06
wert_07
M3_UEBERSTEUERN_MUE_1_2
wert_02
wert_06
wert_07

Did you got that also?

This is expected since wert_02, 06 and 07 are actually the same hex value. When NCS decodes the data is doesn't know which one of those value it corresponds to and outputs all matching strings.

[tdott]

ahh okay, sounds good, I suspected something along those lines. Thanks

[jphughan]

Originally Posted by d0o0fy

Mike,

Here are the values for the other settings:
wert_00: 0x00 (0)
wert_01: 0x66 (102)
wert_02: 0xE6 (230)
wert_03: 0x33 (51)
wert_04: 0x26 (38)
wert_05: 0x7F (127)
wert_06: 0xE6 (230)
wert_07: 0xE6 (230)

Now the interesting bit, there is an equation that seems to govern the value of coefficient used (in pseudo code):

if value < 128 then

coefficient = 2.5 + (value/64)

else

coefficient = 2.5 + ((value-256)/64)

Here are the coefficient of friction that are associated with each value:

wert_00: 2.5
wert_01: 4.09375
wert_02: 2.09375
wert_03: 3.296875
wert_04: 3.09375
wert_05: 4.484375
wert_06: 2.09375
wert_07: 2.09375

So when you go from wert_03 (51) to wert_02 (230) your coefficient goes from 3.296875 to 2.09375 and therefore decreases.

Now it seems that the available values are limited to those 6 here. But I am almost certain that it is possible to fine tune the coefficients by modifying the nettodat directly. Because of the way the equation works, the minimum coefficient possible is not obtained at the maximum value (0xFF) but at 128 (0x80). That would lead to a coefficient of 0.5! After that it's direction the skidpad and see how much of a difference those values make.

This is an intriguing theory. If true, I like the fact that not only is the E92 the most permissive (as it should be ), but the gap between Euro and US tuning is also the smallest there compared to the other body styles, so it doesn't get neutered quite as heavily. This also would suggest that BMW really doesn't want US E93 owners having any fun at all, especially in the US! Actually, it's probably set that way because the E93 will always have a more total friction than the other body styles in any given driving situation because it weighs so much more -- but it wouldn't have a higher coefficient of friction....

Which leads me to my point that I'm not sure it's reasonable to assume that those values are coefficients of friction. First, "mu" in German is still just "mu", not "mue". Second, these values don't make sense as mu values for rubber on pavement, though that could be because the theorized calculation formula is wrong. But third, except for gradual changes due to tire temperature fluctuations, sudden changes from moving from one type of road surface to another, and the difference between static vs kinetic friction, coefficients of friction don't actually change based on driving situation. Your grip to hold your intended line can change, of course, but that's a function of the strength and direction of your momentum compared to where you want to go. And again, what matters when tuning for specific body styles is total friction, not coefficient of friction, the latter of which wouldn't change based on the weight of the car. So that makes me wonder even more what this setting means.

I'm also curious what the MUE_0 option changes compared to MUE_1_2_M. Maybe the former is for when DSC is on and the latter is for MDM? I'm just curious why the latter is labeled "1_2_M". Even if you count DSC Off as a mode (which wouldn't make sense here), there are still only 3 modes on the car, not 4 as the 0, 1, 2, and M naming pattern suggests.

And lastly, if these settings do in fact control DSC leniency, I wonder what changing only MDYNAMIC_MODE_USA does. Having been to the track several times before and after changing that setting, my car already allows more yaw, but if these other settings do indeed represent hard-coded friction limits, I wouldn't have expected that change without having changed these settings. I know that DSC takes other factors into account to decide when and how to intervene (speed, steering angle, etc), but this still puzzles me. "UEBERSTEUERN" seems to translate to "overdrive", so maybe MDYNAMIC relaxes the threshold for more subtle intervention and the "UEBERSTEUERN" settings determine the threshold for more drastic intervention?

[jphughan]

I edited the post below to remove the redundant listing of the options for both settings since they translated the same way, and also to replace the operand logic with just the model each setting is meant for:

Originally Posted by d0o0fy

Hi all,

NCS Dummy has the following Translations for M3_UEBERSTEUERN_MUE_0 and M3_UEBERSTEUERN_MUE_1_2_M:
wert_00: Euro E90
wert_01: US E90
wert_02: Euro E92
wert_03: US E92
wert_04: Euro E93
wert_05: US E93
wert_06: Euro E92 GTS with 4.0L engine (never existed)
wert_07: Euro E90 GTS with 4.4L engine (never existed)

Curiously, there's no setting for the E92 GTS with a 4.4L engine that DOES exist. But considering that the two GTS settings here both translate to the same hex as the Euro E92 anyway, I guess it doesn't matter.

There also don't seem to be any ZCP-specific options anywhere in here either. So where does the reprogrammed DSC that BMW advertised with ZCP come from? I know ZCP has a VO code, but typically the VO is used to determine defaults for settings in these modules, so you'd still expect to see differences in the modules between ZCP and non-ZCP cars.

[d0o0fy]

Originally Posted by jphughan

This is an intriguing theory. If true, I like the fact that not only is the E92 the most permissive (as it should be ), but the gap between Euro and US tuning is also the smallest there compared to the other body styles, so it doesn't get neutered quite as heavily. This also would suggest that BMW really doesn't want US E93 owners having any fun at all, especially in the US! Actually, it's probably set that way because the E93 will always have a more total friction than the other body styles in any given driving situation because it weighs so much more -- but it wouldn't have a higher coefficient of friction....

Which leads me to my point that I'm not sure it's reasonable to assume that those values are coefficients of friction. First, "mu" in German is still just "mu", not "mue". Second, these values don't make sense as mu values for rubber on pavement. And third, except for gradual changes due to tire temperature fluctuations, sudden changes from moving from one type of road surface to another, and the difference between static vs kinetic friction, coefficients of friction don't actually change based on driving situation. Your grip to hold your intended line can change, of course, but that's a function of the strength and direction of your momentum compared to where you want to go. And again, what's more important when tuning for specific models is total friction (mu * weight), not coefficient of friction, the latter of which wouldn't change based on the weight of the car. So that makes me wonder even more what this setting means.

I'm also curious what the MUE_0 option changes compared to MUE_1_2_M. Maybe the former is for when DSC is on and the latter is for MDM? I'm just curious why the latter is labeled "1_2_M". Even if you count DSC Off as a mode (which wouldn't make sense here), there are still only 3 modes on the car, not 4 as the 0, 1, 2, and M naming pattern suggests.

And lastly, if these settings do in fact control DSC leniency, I wonder what changing only MDYNAMIC_MODE_USA does. Having been to the track several times before and after changing that setting, my car already allows more yaw, but if these other settings do indeed represent hard-coded friction limits, I wouldn't have expected that change without having changed these settings. I know that DSC takes other factors into account to decide when and how to intervene (speed, steering angle, etc), but this still puzzles me. "UEBERSTEUERN" seems to translate to "overdrive", so maybe MDYNAMIC relaxes the threshold for more subtle intervention and the "UEBERSTEUERN" settings determine the threshold for more drastic intervention?

You raise some interesting points and I have been wondering myself what those settings were actually intended to change. I agree that I might have made a poor chose of terminology, the coefficient of friction is not going to change due to a software setting, but to environmental factors. The values being above 1 are also not coherent with coefficient of friction. However they could be pre-computed values -- based off slip value limits -- used in the DSC/MDM subsystem saving the subsystem some expensive and repetitive floating point operations. My current hypothesis is that if the internally computed value crosses the threshold of the set value then the DSC/MDM quicks in. I got the equations from NCSDummy, but have no idea where they come from in the first place.

I'm also not sure why there is 2 different values for mu (BTW mu is "my" in German and UEBERSTEUERN means oversteering, not sure about "mue"...), my guess was DSC and MDM. When I first fiddled with the EuroMDM mod, I wanted to make sure that the car was as close as possible to the Euro version as possible and that difference in value picked my interested and I wanted to document it and get the community's feedback.

[jphughan]

Originally Posted by d0o0fy

You raise some interesting points and I have been wondering myself what those settings were actually intended to change. I agree that I might have made a poor chose of terminology, the coefficient of friction is not going to change due to a software setting, but to environmental factors. The values being above 1 are also not coherent with coefficient of friction. However they could be pre-computed values -- based off slip value limits -- used in the DSC/MDM subsystem saving the subsystem some expensive and repetitive floating point operations. My current hypothesis is that if the internally computed value crosses the threshold of the set value then the DSC/MDM quicks in. I got the equations from NCSDummy, but have no idea where they come from in the first place.

I'm also not sure why there is 2 different values for mu (BTW mu is "my" in German and UEBERSTEUERN means oversteering, not sure about "mue"...), my guess was DSC and MDM. When I first fiddled with the EuroMDM mod, I wanted to make sure that the car was as close as possible to the Euro version as possible and that difference in value picked my interested and I wanted to document it and get the community's feedback.

Interpreting the MUE value as some calculated "overall grip level" is definitely plausible; maybe MUE is an acronym of some kind. I doubt we'll ever get a definitive explanation of what exactly these settings and the MDYNAMIC_MODE_USA option unless the engineer who built this is lurking and decides to speak up. But I definitely think it's awesome that you've pieced all of this together!

[terahertz]

To update you guys, I also changed two values DRUCKMODELL_HA and DRUCKMODELL_VA to the GTS option in NCSDummy (there are only two, so i'm guessing it was wert_01 since I don't have my laptop with me). NCSDummy describes wert_00 as !MAN_UMF_GTS and wert_01 as MAN_UMF_GTS.

It may just be in my head, but normal acceleration with full DSC on (not MDM) is less intrusive, doesn't cut as much throttle, and keeps you closer to the grip limit. I notice this in 1st and 2nd gear powering out of turns. Can someone else test and verify?

Google translate says DRUCKMODELL means Pressure Model. Not really sure what it means still.

[eeghie]

Can anyone confirm that the LAT_EBD is related to the Corner Break Control/CBC (or other ...) functionality?

[socrates42]

Very cool read, I just did the euro mdm and was happy with the additional angle. I'd love to have a GTS setting though.

[terahertz]

Originally Posted by eeghie

Can anyone confirm that the LAT_EBD is related to the Corner Break Control/CBC (or other ...) functionality?

Sounds possible. EBD could mean electronic brakeforce distribution. Lateral EBD maybe?

[meeko]

Hi d0o0fy,

Ultimately I'd like to enable M Track Mode on my Z4M Roadster, but the MK60 module doesn't have the required coding - i.e. TRAKTIONDIFFERENZREGELUNG and DTC_LAMPE. But it does allow the Over/Understeer Frictioin Coefficients and Understeer Threshold to be changed.

As for the M3, the coefficients can be set for 0_2 and 1_2. I presume one of these is for DSC on and one for DSC off, but not sure which is which.

The reason for the post is I'd like to understand what friction coefficient values the MUE "wert" values relate to. You mentioned you can calculate these from the Pseudo code. Would you mind explaining how you did this so I can calulate the values for the E85.

Regards,

Meeko.

[d0o0fy]

Originally Posted by meeko

Hi d0o0fy,

Ultimately I'd like to enable M Track Mode on my Z4M Roadster, but the MK60 module doesn't have the required coding - i.e. TRAKTIONDIFFERENZREGELUNG and DTC_LAMPE. But it does allow the Over/Understeer Frictioin Coefficients and Understeer Threshold to be changed.

As for the M3, the coefficients can be set for 0_2 and 1_2. I presume one of these is for DSC on and one for DSC off, but not sure which is which.

The reason for the post is I'd like to understand what friction coefficient values the MUE "wert" values relate to. You mentioned you can calculate these from the Pseudo code. Would you mind explaining how you did this so I can calulate the values for the E85.

Regards,

Meeko.

I got the equation for the coefficients from NCS Dummy. I'm not really familiar with the Z4M but I would guess that if it uses the MK60 ECU than the equation governing the different values might be the same.

[marconi118]

is it possible to code the Euro MDM to a 335i?