Black & Tan 332iT

[bigjae46]

Well on D-Bus like you said, sure. I was looking into reading directly from sensor voltage but I was looking for at least 10hz

MK60 Brake Pressure Sensor Data – GSR Motorsports

[bigjae46]

It doesn’t seem to have affected when TC kicks in - and my guess is actually the inverse of yours as there are separate settings for enabling torque control by brake. I figured it was part of the model for expected yaw calculation. Wish we could decompile this!

I actually dug back into my datalog and it does look like the LF wheel speed gets significantly noisier during DSC control. That would mean that it’s the oversteer model that’s fighting me, which is what logic would dictate with the stiffer chassis and eliminated steering bushing. Essentially, the car is more responsive to my inputs and that is telling the DSC that it’s oversteering because it’s yawing ‘too much’

Now that I think about it, you can use wheel speed readings. When a wheel spins are slips you'll see a deviation from the other tire speed readings. I use that to see how much I'm into the ABS when I overcook a turn or go off track...lol.

I can lend you my AiM Solo2DL for a week or two. It can log wheel speed DATA through CAN. It is a quick and easy 4 wire install.

[YoitsTmac]

It doesn’t seem to have affected when TC kicks in - and my guess is actually the inverse of yours as there are separate settings for enabling torque control by brake. I figured it was part of the model for expected yaw calculation. Wish we could decompile this!

I actually dug back into my datalog and it does look like the LF wheel speed gets significantly noisier during DSC control. That would mean that it’s the oversteer model that’s fighting me, which is what logic would dictate with the stiffer chassis and eliminated steering bushing. Essentially, the car is more responsive to my inputs and that is telling the DSC that it’s oversteering because it’s yawing ‘too much’

Would you want Mk60e5 firmware if you could? Thats the last CAN based unit out there. It’ll project pressure and needs a wake and translator, but could technically work. Has internal pressure sensors too. Bluebimma on Instagram overcame the WSS with a factory WSS but didn’t state which one. Actually, I’m on mobile but you may already know from your wheel hub on your custom knuckle.

I only ask because I was looking into it this weekend. Really wanted mk100 but flexray is a black box to me

[Bry5on]

I guess I was just thinking about it from an acceleration perspective…Where I’d think the open diff model would be braking the spinning tire sooner, to route power to the other tire… and the m variable lock model would be designed to allow a bit of one wheel peel to allow the diff to function (as you’d also want with the wavetrac).

It doesn’t seem to have affected when TC kicks in - and my guess is actually the inverse of yours as there are separate settings for enabling torque control by brake. I figured it was part of the model for expected yaw calculation. Wish we could decompile this!

I actually dug back into my datalog and it does look like the LF wheel speed gets significantly noisier during DSC control. That would mean that it’s the oversteer model that’s fighting me, which is what logic would dictate with the stiffer chassis and eliminated steering bushing. Essentially, the car is more responsive to my inputs and that is telling the DSC that it’s oversteering because it’s yawing ‘too much’

[Obioban]

I think the brake pressure sensors are the upstream ones on the master cylinder, right? I’m not sure if they’d show any pressure while DSC is cycling an individual wheel. Are you prompting the D-bus to get it to send out the pressure data? I could do this temporarily - I don’t like to use the d-bus for my regular logs because it’s so slow.



I think the wavetrac looks more like an open diff than a clutch diff, as it doesn’t have a torque distribution by speed differential. Clutch diffs are always trying to straighten out the car by making the rear wheel speed the same, where the wavetrac is totally alright with differential speed. But without knowing what’s in the model, it’s hard to say honestly.

I guess I was just thinking about it from an acceleration perspective…Where I’d think the open diff model would be braking the spinning tire sooner, to route power to the other tire… and the m variable lock model would be designed to allow a bit of one wheel peel to allow the diff to function (as you’d also want with the wavetrac).

[YoitsTmac]

Well on D-Bus like you said, sure. I was looking into reading directly from sensor voltage but I was looking for at least 10hz

[Bry5on]

When I was looking into advanced vehicle dynamics logging, brake pressure was unavailable on the e46 bus.

You just have to know how to ask for it

[YoitsTmac]

When I was looking into advanced vehicle dynamics logging, brake pressure was unavailable on the e46 bus.

[Bry5on]

MK60 has dicrete front and rear brake pressure sensors. I pull front brake pressure sensor data...could you use that to figure out which axle is activating? Or would ABS activation not affect the pressure at the sensor since it is before the ABS unit.

I think the brake pressure sensors are the upstream ones on the master cylinder, right? I’m not sure if they’d show any pressure while DSC is cycling an individual wheel. Are you prompting the D-bus to get it to send out the pressure data? I could do this temporarily - I don’t like to use the d-bus for my regular logs because it’s so slow.

I would think you’d want to keep the stock m3 dsc diff setting? Don’t the wavetrac and m variable lock both require wheel spin to send power to the other tire?

I think the wavetrac looks more like an open diff than a clutch diff, as it doesn’t have a torque distribution by speed differential. Clutch diffs are always trying to straighten out the car by making the rear wheel speed the same, where the wavetrac is totally alright with differential speed. But without knowing what’s in the model, it’s hard to say honestly.

[Obioban]

Thanks man, this project absolutely would not have happened without you!


I'm still waiting on parts before I can drive it faster/harder/longer, but I did manage to datalog a couple sweeping right corners and play with some DSC settings.

So far, I have done the following:
1) Disabled the limited slip diff model as I've got a wavetrac that is torque based
2) Set the understeer threshold values to non-CSL settings as these appear to be looser than CSL - maybe because of the sporty tires and the need for them to be safe in the wet?

I'm still seeing DSC kick in on at least one of the wheels in .6-.8G sweeping corners and highway offramps at speeds above ~50mph. It's not pulling throttle and it's not dragging hard enough to bring down an individual wheel speed (proof below, including an individual wheel speed datalog).


I don't know if it's understeer intervention or oversteer intervention until I can figure out exactly which channel is getting braked, as none of that data is reported over CAN and it's hard to tell by feel. I'll poke around to see if the DS2 bus/INPA has anything useful.

I also cleaned up my Gauge.S config so I can view individual cylinder retard time-history on the ash tray display. Because that'd long been on my to do list.

Focused on the wrong thing here but you get the idea, nice to get some seat time with the car again.

I would think you’d want to keep the stock m3 dsc diff setting? Don’t the wavetrac and m variable lock both require wheel spin to send power to the other tire?

[bigjae46]

I don't know if it's understeer intervention or oversteer intervention until I can figure out exactly which channel is getting braked, as none of that data is reported over CAN and it's hard to tell by feel. I'll poke around to see if the DS2 bus/INPA has anything useful.

MK60 has dicrete front and rear brake pressure sensors. I pull front brake pressure sensor data...could you use that to figure out which axle is activating? Or would ABS activation not affect the pressure at the sensor since it is before the ABS unit.

Thanks man, this project absolutely would not have happened without you!

My pleasure to be a part of this...it is certainly worth the effort. I just wish I had a better donor car. I now see the dent you were referencing.

[karter16]

Yes I am, and I think that’s actually the DSC’s internal modeled yaw. What makes you think it’s muxed?

Yep I think you're right looking at that - thanks for this. I was going off information I have for that CAN message for other DMEs. Those bytes aren't used in MSS54 so there were no leads in the code. But that log looks pretty conclusive to me that it is indeed yaw.

[Bry5on]

Do you have 0x1F3 logged? and in particular bytes 0 and 1? I *think* they are multiplexed information, but I don't know what. Curious to see if you're logging any variance on those bytes.

Yes I am, and I think that’s actually the DSC’s internal modeled yaw. What makes you think it’s muxed?

https://github.com/Bry5on/332it/blob/main/config-9-20-26_knockCAN04.json

Data is in my datazap log linked above too.

[karter16]

as none of that data is reported over CAN

Do you have 0x1F3 logged? and in particular bytes 0 and 1? I *think* they are multiplexed information, but I don't know what. Curious to see if you're logging any variance on those bytes.

[Bry5on]

Wow! Looks amazing!

Carbon trimming...not fun.

Thanks man, this project absolutely would not have happened without you!


I'm still waiting on parts before I can drive it faster/harder/longer, but I did manage to datalog a couple sweeping right corners and play with some DSC settings.

So far, I have done the following:
1) Disabled the limited slip diff model as I've got a wavetrac that is torque based
2) Set the understeer threshold values to non-CSL settings as these appear to be looser than CSL - maybe because of the sporty tires and the need for them to be safe in the wet?

I'm still seeing DSC kick in on at least one of the wheels in .6-.8G sweeping corners and highway offramps at speeds above ~50mph. It's not pulling throttle and it's not dragging hard enough to bring down an individual wheel speed (proof below, including an individual wheel speed datalog).


I don't know if it's understeer intervention or oversteer intervention until I can figure out exactly which channel is getting braked, as none of that data is reported over CAN and it's hard to tell by feel. I'll poke around to see if the DS2 bus/INPA has anything useful.

I also cleaned up my Gauge.S config so I can view individual cylinder retard time-history on the ash tray display. Because that'd long been on my to do list.

Focused on the wrong thing here but you get the idea, nice to get some seat time with the car again.