Mrmerlin

Bill IIRC the sunroof harness goes up the right side of the rear quarter panel, ( it should be bundled with the roof antenna wire so to see it it will be remove the rear seats time, to be able to move the front part of the quarter panel........ has there been any speakers installed? the harness also should be running down the console/ hump to the rocker so maybe the front seat pull should be first

Bill Ball

Stan:

Yep, that is where the harness runs. Connector T42 that carries the sunroof motor wires is near the door opening edge near the speaker, so if you pull the rubber bead that covers that seam/opening, you can get to T42. I disconnected it and the short was still present, so the short is before it passes behind the speaker and up to the sunroof, between T42 and the window controller. If I can't spot it in the harness traced from the T42 end as far as I can go, I'm going to pull the driver seat out.

Alan

Bill - which pin is it that is shorted to ground pin5 or pin6?

I'm going to guess pin 5

Alan

Alan

OK just for some completeness here - I've been thinking about the switch input details Ed posted - they confused me signigicantly at first because I expected simple switching - Switch poles biassed to 6v & Red wire & Green wire at 12v & 0v.

I then further assumed the processor would just read the switch values & drive the relays...

Porsche does other things this way (e.g mirrors) so it seemed the obvious path.

Its pretty apparent to me it doesn't work this way now. I believe the switches actually drive the relays directly and the processor monitors the motor outputs to determine whats going on and then also has a path of its own to drive each relay too. To me this is actually more complex to implement (and costs an extra relay) - however it does have the benefit of an extra failsafe - even if the microprocessor circuit dies - you can still operate the windows ~directly. Probably Porsche's reason to do it this way.

This would suggest that the switches always override the processor - and this seems to be true testing my car - even when the processor has detected a stall or timed-out you can always make the window motor drive by pressing the switch (until the thermal cut-out engages).

This may also explain Bill's behaviour of the inoperative express modes. If the processor is monitoring the motor outputs it is looking for specific conditions on these outputs that tells it a buttom is being pressed. As in: I see the motor common, passenger motor & sunroof are at 12v and the drivers motor is at 0v - so the drivers close button must have been pressed...

This only affect the express modes because the rest are directly driven via the relays.

However if the sunroof motor is always at 0v it will be very confused about what is happening and will likely just ignore all the invalid cases this generates. Thus it will miss all the cases where the sunroof motor should not be at 0v:

This includes:
Rest (it always thinks the sunroof close button is being pressed - makes no real difference)
Driver Close (this is an express mode so it won't work)
Passenger Close (this isn't an express mode so it won't make any difference)
Sunroof Open (this is an express mode so it won't work)

It will not affect the cases where the Sunroof motor should have been at 0v anyway:
Driver Open (this is an express mode - it should still work fine)
Passenger Open (this is an express mode - it should still work fine)
Sunroof Close (this isn't an express mode - so it won't make any difference)

I think this matches what Bill saw...?

The direct switch to relay drive scenario implies the motors are actually connected slightly differently than I showed - conceptually it is very similar but based on Ed's voltage readings I'm still figuring a few thing out... I'll post an updated diagram later.

Alan

Bill Ball

Yes, it's pin 5 on T42. Initially when I found that, I thought it should be ground and pin 6 was power. BUT pin 5 comes from from pin 32 on the controller, which is the sunroof motor output, and it should be 12V when at rest. Pin 6 is the common motor lead and also should be 12V at rest. I may have mixed them up in previous posts, so I will go back and clean that up.

I read your immediate preceding post that details how you believe the system works. It certainly isn't as intuitive, and the short in the controller output from 32 (sunroof motor) wreaked havoc with a lot more than the sunroof. I'm hoping that all it is. I'm tempted to cut 32 and run a temporary jumper just to find out now if that is responsible for all the varied anomalies.

Alan

Bill - yes a temporary repair there would be a good check that it solves everything. Based on what I showed in my last post I'm highly confident it will in fact solve everything. If it does - see what the effect of the pin 38 connection is plugged & unplugged.

Alan

Alan

OK here goies an approximation - Ed mainly for you !

In addition the processor will monitor CM, DM, PM, SM to know whats going on.

Bill Ball

UPDATE:

The repair to pin 32 output to the sunroof motor fixed the major problems, but not all. I reconnected 33-38 jumper as well. See critical information below about 33-38.

FIXED:
- There is 12V on window controller pins 32-35 at rest (correct)...note below that pin 33 (passenger window motor) is getting 12V from pin 38.
- The passenger window no longer spontaneously goes down (the primary problem)
- The door lock auto close sequence works properly including the driver window which did not go up before.

EDIT: IGNORE MOST OF THE STUFF FROM "NOT WORKING PROPERLY" BELOW AND SKIP TO THE NEXT SECTION
Things that are NOT working properly that I BELIEVE should work differently:
- Express Up does not work on either window (same as before)
- You cannot operate the windows with the doors closed and the ignition not in the run position (same as before).
- Both windows and the sunroof operate like express down anytime you push the buttons for more than a split second until you push them a second time. (LATE EDIT: NORMAL according to the owner's manual)
- A relay clicks anytime you change the ground state of pin 15 (door pins). Pin 15 still puts out 12V from the controller as soon as the controller gets power to pin 37 (battery power) with nothing else plugged in (just using a jumper wire to pin 37). So, that must be normal, BUT the same controller taken from this car and installed in a GTS does not show a relay click with change in ground state. However, ground state is not affecting the ignition off operation of the windows.

NOTE: The ignition feed (pin 36) goes to ground after sitting off for a few minutes. Is this correct? Regardless, the windows will not operate anytime the ignition is in other than run position.

With the plugs removed from the controller, the plug pins 32-35 each rise from zero to 400 mv then drop to zero and start over. This may be a fault with my multimeter. If I leave the multimeter +V probe detached, it marches up miliV, but at a much slower rate. If I pull the probe out of pin 32-25, the miliV reading at the time lingers, fluctuating a little, until I ground the probe.

Any ideas now? I'm tempted to wrap it up as is, but there appears to be another fault.

NOTE: If I cut 33 to 38 jumper, the driver window no longer lock auto closes. Also, pin 33 no longer has 12V at rest. However, the passenger window still operates normally, but 33 only shows 12V when the passenger window switch is pushed to OPEN the window. Everything else is as with 33-38 connected above. So, I'm still wondering what 38 is supposed to represent and if it is supposed to supply 12V to pin 33.

EDIT: IGNORE MOST OF THE STUFF FROM "NOT WORKING PROPERLY" ABOVE AND READ BELOW

I started to do measurements on the controller box with just power (pin 37) attached, and I got some very weird numbers. So, I swapped in the OLD controller. Things changed dramatically.

- pin 33 now has 12V with or without the 33-38 jumper. A whole mess of other pins now have 12V that didn't before.
- The relay no longer clicks when pin 15 ground state is changed.
- The windows now operate with the doors closed and the ignition off

If I cut the 33-38 jumper, the driver window no longer lock auto closes. Other than that I can't see any effect from that jumper.

The only anomaly still present is no express CLOSE to the driver's window. According to the owner's manual, there is express open on all switches (including sunroof) if the switch is held down for more than 0.5 seconds. There should be express close on the driver's window if the ignition is on.

If someone can deduce how express close is supposed to work on the driver window, please let me know.

So, at this point, it appears we had a short in the wiring from controller pin 32 and a bad brand new controller, and something else is still wrong preventing driver side window express close. That is the only missing function. If we have to leave it there, so be it.

Alan

Bill - good - based on what should have failed due to effects of Pin32 shorted - these things should all work now...

I think maybe your expectation of how things should work is off a little...

Express close will only ever work on the drivers window AND only with ignition on (not in door open mode). It is only functional in ROW mode - is the jumper right?

Express open should work on both windows and the sunroof - including in the doors open mode. I think it may all be perfect!

Alan

Bill Ball

Alan:

You got me excited there for a minute. I went out and checked. Still no express close with the door closed. ROW jumper is right. I'm going to run through Ed's voltage findings now that I am real close.

EDIT: Jumper was intact, set for USA/Canada. If I open the jumper, I get driver express up and windows still operating with the doors open. So, things are perfect!

Alan

Bill - did you reconnect the Pin 38/33 wire link - I'm still interested in what effect that seems to have anyway?

Alan

Ed Scherer

Ignition feed dropping to 0 V over time seems OK (I'd expect that various loads across ignition and ground would cause this); can't easily test it anymore right now, sorry. My windows and sunroof (still not in ROW mode, Alan! ) will still operate after ignition off until I open a door (i.e., a pin switch is closed, i.e., connected to ground).


That might be a ROW feature; none of my stuff express closes, only express opens (which does work).

I think this is one of the "DOT mandates" that Alan was complaining about. He's cool with body parts being chopped off by auto-closing windows.



Other interesting tidbit: the stall current detection doesn't work very well on my passenger side window (which, interestingly enough, is the one of the three motors that I replaced a few years ago). I just noticed while testing the "close everything by holding the key in the door in the 'locked' position for awhile" feature that — after having all three things open — the passenger window would close... then 5 seconds or so (controller timeout rather than stall current detection!) would elapse before the sunroof would close and then (no significant delay) the driver window would close.

I think I noted the stall current unreliability in one of my earlier posts (and Alan commented on the timeout fallback that the controller employs).

Ed Scherer

I'd really like to see your best-guess source code for the microcontroller.

Bill Ball

Yes, the only effect I could observe was the lock auto close for the driver side window. 33-38 was needed for that to happen. Everything else appears unchanged. I rechecked after opening the ROW jumper, and 33-38 has no effect on driver express up or windows still operating with doors open.

Ed Scherer

Meaning? We need to know what mode you're choosing: ROW "Who needs all their body parts, anyway?" (22/17 not jumpered) or US/Canada/Australia "Save our limbs!" (22/17 jumpered).