MDM retrofit

[VicS5]

I’ve retrofit MDM mode to my 09 M3 track car that did not have it from factory by installing a clock spring from an MDM car.

Problem is I have a detachable steering wheel so no button cluster on the wheel to activate it. With a little reverse engineering and too much time with my multimeter, I found that the MDM button on the cluster sends an analog message to the SZL with 82 ohms of resistance. All 4 buttons on each side of the cluster send their analog message through 1 wire, and each button has a different resistance. My assumption is the SZL interprets the different voltage or current it receives from the different button presses to determine whether to activate MDM vs. change the CD vs. alter the stereo volume.

I wired a switch on my wheel with an 82 Ohm resistor in-line. Good news is it works, but frustratingly only about 1/3 to 1/2 of the time. I’ve tried troubleshooting by removing variables like the switch (touching the 2 bare wires together) swapped out the resistor, and bypassing the 11 pin connector in the detachable steering wheel hub. Still inconsistent. Tried numerous combinations of cycling the ignition, turning MDM off before shutting down the car, dirty looks, and screaming loudly. None seemed to make a difference.

I’ve addressed it by adding a 2 pin connector in parallel so I can plug my stock 4 button cluster in to activate MDM, then unplug it and stow it. This works reliably every time when I use the stock buttons. MDM stays on for that session, but next session is another flip of the coin.

My guess is that that there’s more to the stock button than a simple resistor, and that there is something about the “shape” of the analog message that I am unable to resolve with my multimeter, but might be evident on an oscilloscope.

I am not an engineer, and I only pretend to be one in my garage. I’m hoping there’s an EE (or just someone with a better understanding of electronics) who can point out my folly.

Current setup is workable, but it drives me crazy that it works sometimes. I’d probably feel better if it didn’t work at all.

Thanks.

[kolosy]

Paging amg6975

[Scrippy]

Try the same ohm rated resister but with a different tolerance value. If 82 ohms is the sweet spot, buy some 82 ohm resistors in different tolerances and try them. I believe the tolerance is the 4th color ring. Resistor tolerance is the percentage of error a resistor's actual resistance may have from its stated value. Another option is to buy a variable resistor. I’d start with a small scale variable resistor that is like 1-200 ohms. Monitor the ohms with your multi meter while turning the knob and set it to 82ohms. Try your button for mdm and if it doesn’t work reliably, turn the knob as to tune in a more effective ohm rating. The bigger the scale of the variable resistor, the more ohm delta per degree turn of the knob, hence a small scale within your target of 82 ohms like a 1-200 ohm. Not a 1-1k or 1-10k…… Pick your variance scale as close to what you think the target is as to have a knob that isn’t excessively finicky and over or under shoots your target.

[amg6975]

Originally Posted by kolosy

Paging amg6975

You rang?

I don’t think this problem is as simple as measure and stick in a resistor. You need to take into account the resistance of the button and wires, etc. I think the OEM button circuits are actually current sources, not just plain resistances. The 12V can vary a lot and is very noisy so different resistances will look different at different voltages. There were a bunch of FETs in the OEM wheel but I never really bothered to look too deep into it when I looked at mine. I also think there is a base current/resistance and not just an open, but it’s been a while since I looked at it.

For a first pass let’s assume the button and stuff has a few ohms of resistance so give a 75ohm resistor a try. If that still only kind of works try 68 ohm. I’ll report back with more info sometime this week when I can get something set up in the lab to test.

[VicS5]

Thanks for the guidance. I’ve already tried multiple resistors, although I don’t have a 68 ohm resistor. I also tried it without the switch (just touching bare wire ends with resistor on one wire) and observed the same result. Sometimes it works and sometimes it doesn’t. Maddening.

The stock button cluster I’m using when the other 82 Ohm button fails is bare bones. It’s only using wires 1 & 3, and taps into the same circuit as my add-on button, but beyond (bypassing) the 82 Ohm resistor. The other 4 wires are trimmed and capped as I don’t need them.

[jeho]

Did you ever figure this out?

Trying to see how to program my aftermarket wheel with custom button to this

[VicS5]

No, not yet. One of buddy’s kids stopped by over the weekend to help. He’s an EE, so he at least has a better grasp of electronics than me - a monkey with a multimeter.
We pulled the cover off the stock button cluster and he confirmed that the button circuit has nothing more than a resistor in it. We tested voltage changes observed with presses of different buttons and found they vary by only a couple of millivolts, so the SZL is probably very sensitive to resistors with loose tolerances. I still can’t understand why it works sometimes and not others. Seems to me if I’ve got a bad resistor in there, it wouldn’t work at all.

[996dnz]

Just going off what the paddles need they have 2 resistors one for each pin which is then bridged.

The PCB for the multicontrols has 2 inputs that get bridged. The 2 gold posts where the button presses down. Have you check these in isolation? Its 100% possible as there are multiple companies that supply them do this now.

[996dnz]

So each of the gold circles get tied when a button is pressed. But the button presses two of these at the same time.



So in your setup you are using A single 82 Ω (±1%) i assume.

I would expect the PCB to have 2 for this which means they could be

47 Ω ±5%
+
33 Ω ±5%

Which would give a more flexible output for the SZL to understand or just try a 82 Ω (±5%)


Because then
82 Ω (±1%)
82 × 0.99 = 81.18 Ω
82 × 1.01 = 82.82 Ω

Or 82 Ω (±5%)
82 × 0.95 = 77.9 Ω
82 × 1.05 = 86.1 Ω

[jeho]

Does anyone know if there is canbus access, could it be possible to just program a new button through that for enabling/disable M button? I see that JQWERKS steering wheel has customization and accesses it via CANBUS. Also other aftermarket wheels like KMP that seem to have a M button

[kolosy]

Originally Posted by jeho

Does anyone know if there is canbus access, could it be possible to just program a new button through that for enabling/disable M button? I see that JQWERKS steering wheel has customization and accesses it via CANBUS. Also other aftermarket wheels like KMP that seem to have a M button

You need to send a message on the can bus. Very doable but requires an MCU of some sort to do that, a momentary button won’t be enough.

[rg1220]

Sometimes working and sometimes not working makes me wonder if there's a floating voltage issue. I wonder if you need a pull down resistor to basically "clear" voltage when nothing is pressed. Without a pull down resistor the voltage might be floating and would probably mimic just holding the MDM button down - but in your situation you'd be pressing the button multiple times but the computer would see it as being held down the entire time.

[amg6975]

Originally Posted by rg1220

Sometimes working and sometimes not working makes me wonder if there's a floating voltage issue. I wonder if you need a pull down resistor to basically "clear" voltage when nothing is pressed. Without a pull down resistor the voltage might be floating and would probably mimic just holding the MDM button down - but in your situation you'd be pressing the button multiple times but the computer would see it as being held down the entire time.

No, the SZL uses the combination of the resistances of the buttons to decide which are pressed, pulling the signal down would likely make nothing work.

[rg1220]

Ah gotcha - I just assumed and was hopeful it was a bit more "simplistic"

[amg6975]

Nothing is as simple as we hope. It'll be something more like this:





I'm not going to spam all the permutations here but they will have selected all the resistor values so they can detect any single button or any combination of buttons they want.

They very likely use a current source in the SZL for a couple reasons:

1. Then the voltage drop in the wiring doesn't matter
2. They can read the values with a simple/cheap ADC

Couple other thoughts:

1. I see what rg1220 is getting at, you may need a "no button pressed" resistor, not totally sure.
2. 996dnz's picture and description shows that each button actually pressed two "buttons" and if you count up the wires, it would not surprise me in the slightest if there were redundant circuits. Which could explain the issue.