[Doyle]

Yeah, I was just reading through the x3 forum and it looks like there is a position sensor within the TC and the TCU cross compares what it thinks the position the servo is at and the % of lockup. In the thread I linked, since the servo was spinning, but not locking up the clutches, it threw the ABS/DTC/4x4 codes. There might be a way to control the delta the two sensors (servo & clutch) "see".

If you get downstream of clutch lockup (servo position) and up stream of clutch lockup (clutch position sensor) then the TCU might still think everything is where it should be. For example: the car may think you are in terminal oversteer, but it has locked up the clutch, so it throws its hands in the air and says you are just a really shitty driver...

[DannyZRC]

yeah, you will need to spoof the position sensor's readout most likely.

but the only real way to answer any of this is for some brave soul to start playing with it and putting their stash of magic smoke on the line.

[Doyle]

I have to drop off my car at the shop soon to get the tranny and TC flushed/filled. They are extremely reputable and get into funky engineering tasks, so they may be able to help us out.

[Pseudo Nim]

This thread has SO much more intelligence in it than the other thread about pulling fuses and the wipers going bonkers that I almost feel I'm reading some kind of Stanford publications on a completely different forum.

[Doyle]

Originally Posted by Pseudo Nim

entire post

I wouldn't say it is all that intelligent, lol. My coworkers would probably just call this "redneck rocket science".

Also, I think I we are getting closer!

This is a very pertinent read:

http://tis.spaghetticoder.org/s/view.pl?1/09/26/67

If I am reading this right, essentially BMW chooses a resistor from the factory that corresponds with the range of movement between the open and closed position of the clutches. The resistance is measured by the range of rotation on the cam gear that works with the worm gear and it's relative voltage in the electric servo motor. BMW defines a "curve" for that transfer case. I believe the "curve" simply defined as:

voltage at clutch position zero -> (measures resistance across servo motor and classifaction resistor terminal) -> voltage at clutch position closed

This accounts for variability in machined/mechanical tolerances. (note: My guess is that they don't measure every unique transfer case, but they just choose ones for each model.) Once that resistor is chosen, the ideal/actual "curve" is stored within the TCU. Everytime you start the car, the servo cycles and measures the real life "curve" in order to make sure that the full range of lockup is available. This also allows the TCU to apply some sort of offset in order to account for variance in oil viscosity, wear, and other real world issues.

This makes sense to me, especially after reading the x3 thread. Upon start up, the engine would cycle the servo and attempt to measure the change in resistance of the classification resistor that sits on cam gear. Since the cam gear wasn't moving, but the servo was, this popped the error lights.

Here is the image of the classification resistor and power source for the servo on an e83:

http://tis.spaghetticoder.org/s/view.pl?1/09/23/57

So...if we just measure the resistance on the classification sensor, create a dummy, and high jack the electric servomotor with our own potentiometer, we are in business. The even better news is that this looks like it could be completely plug and play, and as easily removable for factory service as a piggyback.

[DannyZRC]

they probably measure each individual unit, this is semi-common in automatic transmissions as well.

[ajsalida]

As Danny has pointed out, there is very complicated vehicle dynamics and traction control software layered on top of these simple analogue signals. I would not want to guess/assume spoofing such a servo signal is going to result in something you'd want to drive in all conditions.

What is happening is there is a model of sorts inside the controller that is inferring vehicle dynamics/state from sensor input & control outputs. You hack into that somehow you'd want to know what the controller thinks is going on under the conditions you modify.

For the same reason I'd want to know what the system thinks is happening when I drop a mechanical LSD into the whole xDrive/DTC/DSC/ABS/e-LSD system. Probably not as critical as spoofing the TC signal. But I'm an old controls guy, and swapping out a mechanical subsystem that duplicates part of the electro-hydraulic system function without telling the controller, well bad things can happen.

[DannyZRC]

It's possible that bad stuff happens, but there are more than a few instances of these types of systems being tuned/modified/bruteforced out there, and so far the trend seems to be that tuning AWD systems is certainly a solvable issue.

[ajsalida]

Originally Posted by DannyZRC

It's possible that bad stuff happens, but there are more than a few instances of these types of systems being tuned/modified/bruteforced out there, and so far the trend seems to be that tuning AWD systems is certainly a solvable issue.

Agree, it's solvable. It's not going to be more complicated than a piggy-back controller for example. It would be nice if someone with the cred cracked the tables or whatever is going on inside the box and said, yeah, OK, if I do XYZ it looks good.

[Doyle]

Danny- The reason that I say it is likely model specific is that there is only one part number for that resistor in the BMW parts diagram. However, the shop may just run out to Radioshack and grab the right resistor for the car for all I know...

AJ- That is the precise reason why I think having the push/pull solution is ideal. Essentially, there would be a switch that created two different circuit paths. Close the switch, and the circuit flows just like normal. The car would behave absolutely normally. Open the circuit, and that would give you the manual adjustment. That way you can determine your own risk level.

Without actually mocking a system up, it is all speculation. However, it is my belief after reading BMW's literature that all the car looks for with the transfer case is cause and effect. Does the servo turn the cam? Yes? We're good to go. If you treat these sensors as merely reactive and not prescriptive, then this should work out.

Give me some time to do some rigor with this paper system and I'll try to think of some weak links or potential issues. Feel free to do the same. However, mechanically, I believe the system will work flawlessly...you are spot on about the sensors!

[ajsalida]

Originally Posted by Doyle

Danny- The reason that I say it is likely model specific is that there is only one part number for that resistor in the BMW parts diagram. However, the shop may just run out to Radioshack and grab the right resistor for the car for all I know...

AJ- That is the precise reason why I think having the push/pull solution is ideal. Essentially, there would be a switch that created two different circuit paths. Close the switch, and the circuit flows just like normal. The car would behave absolutely normally. Open the circuit, and that would give you the manual adjustment. That way you can determine your own risk level.

Without actually mocking a system up, it is all speculation. However, it is my belief after reading BMW's literature that all the car looks for with the transfer case is cause and effect. Does the servo turn the cam? Yes? We're good to go. If you treat these sensors as merely reactive and not prescriptive, then this should work out.

Give me some time to do some rigor with this paper system and I'll try to think of some weak links or potential issues. Feel free to do the same. However, mechanically, I believe the system will work flawlessly...you are spot on about the sensors!

I am not sure what your background is so forgive me if I am stating the obvious. What you are talking about is the local sensors for the TC servo system, and then spoofing them. OK that may work.

What I am talking about is the higher level more complicated vehicle stability and traction control system (that includes e-braking & ABS, e-throttle, etc etc and the meta-system DSC that sits on top of all of it). This system takes sensed output from the TC sensors and infers/assumes its state, whether it is working or not and how well.

One would want to go into that entire universe of vehicle states and look at the ones where the controller is saying OK more power to the front axles please, it commands that, and then THINKS that is happening. But that isn't happening and so other things happen, and it senses that, ad infinitum, what does it do now?

Well shit I don't know. Does it think the front wheels have no traction? What then? Etc. This is not trivial.

edit: and yes you can turn off DTC/DSC, but you cannot turn off basic xDrive + e-LSD feature as far as I know.

[Doyle]

Yup. I'm in full concurrance, and I understand exactly what you are talking about and where you are coming from.

My reference points are the x3's that lost their cam/worm gear as well as the guy that dyno'd his car in RWD mode. All said that the car behaved entirely like a RWD vehicle, and even showed that there was no loss of power...except they all looked like Christmas trees. Now, you are absolutely correct that this is a noshitbigdeal. That is why I'm trying to run through scenarios in my head. The main one that I can think of is that the car will cut power when it thinks all is lost (for example: massive delta between front and rear wheel speed sensors). However, with the dyno and the servo failures, we saw that wasn't the case. However, they were throwing codes like crazy. Since I would likely only go into RWD mode on the track, I would be fine with the traction control systems off in RWD mode. My only concern is ABS.

I'm not trying to play fast and loose with thousands of dollars of equipment, I'm just trying to keep all options open.

[ajsalida]

Also must consider scenarios where lots of g's are being pulled in all possible directions, and maybe some big slip angles and wild counter steering. WTF the controller will do when it suddenly understands fronts are not getting any extra traction under those conditions.

There may also be other sensed failure modes beyond servo (e.g. broken front driveshaft/diff/halfshafts) that would get flagged and light up the same codes.

Anyway this is a worthwhile pursuit at all levels, even just thinking it though.

[ajsalida]

Oh yeah and don't forget, lore has it that the TC is full open over some HWY speed, like 70 MPH. So maybe there is a command in there that says, you know, just turn it off & no codes.

[Doyle]

AJ-

Per BMW: "...X-drive operates in rear wheel drive...at speeds in excess of 112 mph". That was my first train of thought, as brought on by member Amclint. Since that involved hacking the logic of the TCU, I didn't know how feasible that would be without aftermarket support. For example, with my JB4, via the user/steering wheel interface, I would be able to adjust the mph where that disengagement occured. Super cool and elegant if possible.

So you're saying that in the interest of science someone is going to have to go to an empty parking lot and hoon the shit out of their car?

[ajsalida]

A fundamental principle of science is hooning should be done with OPM. In this case service loaners.

[amclint]

At the end of some of my 1/4 runs I'm doing 130mph as I keep in it after the lights, I can't imagine slowing down from that speed with no ABS

[Doyle]

Aj- I like where your head is at, buddy.

Amclint-At that point it is all testicular fortitude. Pretty sure Jackie Stewart didn't have ABS, lol...

[NjPrince]

Thanks! Helped alot for my father's car!

[MTP335]

Blah blah blah... yes i want some rwd on my awd pls. Being able to toggle between awd and rwd would be greattt

[Doyle]

I'll just drop this here:

http://cobbtuning.com/products/?id=5676

[Yvette]

only complain for my xi is lack of aftermarket suspension parts.