toddlamb

Some of that points to distance sensor but also could have a valve body issue or pump issue, or a combination.

PV997

Hi Todd - Any thoughts on how this series of faults could be caused by a distance sensor failure? My suspicion would be hydraulics (as you mention) but can a DS failure mimic a hydraulic issue? I realize we are all learning here so any thoughts (even spitballing) are welcome.

toddlamb

There's usually a cascading effect of one thing triggering codes for other things, sometimes seemingly unrelated. This is one of those in between situations where the failure could be several different things. We have seen some of those codes with a distance sensor failure, and we have seen some of those codes with a hydraulic failure. We have also seen them with wiring/connector failures. Further diag needed to confirm the root cause.

floatingfish

Thank you Todd and PV997.

jjrichar

I agree with what has been put forward regarding where the issue may lie. Below is some information that might help in the diagnosis problem. This is my best stab at providing some guidance on where to pin point the issue. We are still learning. Symptoms like this that don't immediately point to a smoking gun can be hard to work through.

Firstly, do you plan to do the work yourself or have a shop do it for you? I've owned a Durametric before and it was very limited in functionality compared to PIWIS? Can you see distance sensor response, activate solenoids, etc?

The CAN faults are just symptoms of another problem from what I have observed. The issue is the non selection of gears. When exactly did the fault occur? Immediately after start or when you selected a gear to begin driving? Which gear did you select, R or D?

If you clear the codes prior to starting engine and try again, what happens? Do you hear a clunk when you select a gear? If you clear codes and select R, is the result different to if you select D?

Diagnosis:
The symptoms point to one of two possible scenarios from what I see.

Scenario 1. The gear was selected correctly but the distance sensor is unable to detect this. The distance sensor response is still within the plausible signal bands so isn't giving a 173x code. Instead the TCU assumes the sensor response is correct and so it thinks the error is one of gear selection. Hence the gear selection codes. From what I read of the error codes, the transmission has attempted to disengage reverse to then select a forward gear. It needs to go to a neutral position first prior to selection of any forward gear due to the mechanical lockouts. It thinks this hasn't happened so it also logs a gear selection hyd fault.

Scenario 2. The gears aren't being selected as requested and the distance sensor is responding accurately.

Scenario 1. If a distance sensor problem I would be looking closely at the sensor responses. Do they look normal from what you've seen elsewhere? You may have seen the '$100 distance sensor' thread. There is a lot more information there about correct DS response, especially towards the end. https://rennlist.com/forums/997-foru...ce-sensor.html

Scenario 2. I'd be looking closely at the valve body solenoids. If they electrically fail the TCU should detect this and tell you, but they can fail without giving you an indication. They would need to be removed and pressure tested if this is the case. It's very unusual for valve body hardware to fail. They are just pistons inside bores that will eventually wear in time, but they rarely just fail. Performance just degrades. This is something we've seen little of in these transmissions.

Of the two scenarios I'd concentrate on the distance sensor failure first and get some readings of what the TCU is getting from the sensor.

I mountain more information I could put here but it would take forever and isn't really required until you go through the initial steps.

PV997

Good analysis. The thing I have trouble with is how often does the distance sensor (DS) fail without setting P1731, P1732, P1733, or P1734 faults? We have seen plenty of cases where the DS has some sort of internal component failure and continuously outputs ~95% duty factor signal. We know that when this happens it always sets a P173x fault. We also know from our testing that when the magnet is removed from a functioning DS that it also outputs a ~95% duty factor. So the reasonable conclusion is that in a DS failure it can no longer sense the magnetic field (and outputs 95% DF).

How often though, does a DS failure result in a sensor getting "stuck" at some plausible value (say 60% DF as an example)? That's what it would take to not set a P173x fault. Has anyone ever actually documented this happening? I'm asking as I don't know, but it must be extremely rare if it does happen. Further, it's very hard to for me to come up with a DS failure mechanism that would result in this signature. Perhaps some sort of wiring issue (high resistance or a leaky path to ground) could cause the signal to get distorted and misread by the TCU, but beyond that I have trouble explaining it.

If you look up the faults @floatingfish lists in the Porsche diagnostic trouble code documentation, they all point to a clutch or hydraulic (i.e. valve body) failure provided the P1731, P1732, P1733, and P1734 faults are not detected. Based on this I would start with the hydraulics as the main suspect (scenario 2). It's certainly worth monitoring the DS signals with a PIWIS to see if they are consistent and make sense, but I find it hard to believe the DS is the issue here without something much more definitive.

Also noting that valve body solenoid replacement is close to an order of magnitude cheaper than a DS replacement, if that is the actual cause.

Agree with jjrichar's comment regarding the Duramteric, it's PDK t-shooting capability is extremely limited, plus I've seen it misinterpret codes and provide the wrong text explanation. A PIWIS is far superior.

Prairiedawg

The first thing I think of is, it screams like an electrical fault. The canbus errors with the hydraulic errors simultaneously. The hydraulics need voltage to actuate and if a coil failed to ground or shorted across, could be drawing the voltage downward. We all know these cars don't like any type of low voltage and seen a myriad of issues due to failing batteries. It's possible it could be simply that. It seems like when a distance sensor fails, it tells you so. Could be a bad TCU. That's the brains of the operation and communicates with the rest of the car. Could be the source of the hydraulic and Canbus failures. If possible, I would try and swap that out with a known good one first before doing any invasive steps first.

floatingfish

Thank you all for your help and input, To answer some questions that have been put forth:

1. I plan to have a shop do this work - Atlanta Speedwerx.
2. I'm using a Durametric and it is not giving me any more detailed info or options.
3. The fault occurred as soon as I started the car; I then tried R, then D, no luck. When in R, the red indicator light blinked. When in D, it was steady red. In all cases neither R or D would engage a gear. P would put it into park.
4. I cleared the codes prior to starting the engine and tried again - the same result. No clunking or other noises when trying to select a gear. It doesn't matter if I select R, or select D, neither makes a difference.

Speaking with Atlanta Speedwerx, they suggested I consider a full rebuild since I drive this car a lot (15,000 miles annually, including monthly aggressive track days) and I plan to keep the car until I die. When they get the car and take a look, I'll update the thread with what is found.

jjrichar

Thanks for the info. Feedback of what exactly were the faults with the symptoms you have will be very useful for future reference.

BadsSgt

Hello,
I have a 2012 Panamera with 120k miles. I was starting to have a little noise on acceleration. after much reading I decided to get the piwis iii along with everything to do transmission oil and filter. my air conditioning went out also so I just got done replacing the whole system. then my job laid me off in the recently tech layoffs. its been bad luck after bad luck. I decided to do a transmission calibration before doing the fluid change as I just received my piwis iii unit and really wanted to see if it would make a difference. while doing the pdk calibration while sitting in it with my foot on the brake. it did exactly what you have described here. now I have the following codes. you can see the most I made with all the steps we have taken so far. here. https://rennlist.com/forums/panamera...-panamera.html do you know if there is a way to fix this? my car was running before I tried the calibration, but now its down. it will not shift into gear or do anything. I am a little limited as I was laid off. I could possibly get a used pdk for 3 to 4k from what I see with around 60k miles. but I am really hoping there may be a cheaper way to go as it was running but now done do to calibration.

GTSpure

Thanks for your sharing.
Good luck with the diagnosis and repair, I'm glad to hear good news.

12gauge

When I look at PIWIS Diagnostics under the PDK selection, it basically sub-menus down to two selections; one for presumably runability issues and one for when the PDK transmission has been replaced called Calibration after part replacement.

Since I've replaced the Distance Sensor, I suspect I need to compensate for the differences between the factory installed DS and the new one.

The logic chart in the Porsche Technical Document I have, surrounding the Distance Sensor diagnostics, has two footnotes referenced on the "Replace PDK. Erase fault memory" box.What the heck is "engine torque loss adaptation" and why would I have to perform it, since everything is from the factory except the Distance Sensor?

I'm not certain of the significance of s/w version 25.00 or higher or what version I have for that matter.

I'm interested in hearing from anyone with personal experience using PIWIS-2 to calibrate (or compensate) for "a replace part", in my case the new Distance Sensor.
What can I expect and what are the conditions I need to setup and prepare, i.e.:

• Do the rear wheels need to rotate, or is done on the floor with my foot on the brake?
• Will it shift into each gear, take a reading and apply a compensation value to the TCU?
• Does the PDK fluid need to be within a particular temperature range?
• How long can I expect it take?
• Does it compensate first time every time or might it take several passes?

PV997

Regarding torque loss adaptation, go back to the main post and see the three attached documents at the bottom of it. Download the first doc and see section 3.6, it's a seemingly silly procedure where you run the engine for 3 minutes with the AC off and then on. In the document it says to run it before calibration.

We've seen this procedure discussed before and it supposedly clears the TCU's learned memory (aka adaption tables) that have been developed over time (though I have not been able to confirm this). This is a good point, are people who are having cal failures trying this before running calibration?

Edit: Noting that I'm not saying this is always required as I know plenty of people have skipped this step without issue. Mainly trying to understand if it's actually helpful in getting past cal problems (which I've long thought might be linked to the adaptation tables). However, the doc does say "Engine torque loss adaptation must always be performed before carrying out calibration" (page 3).

jjrichar

Here's my take on your questions.

From what I understand, the torque loss adaptation is for the car to work out how much torque is lost when the a/c is activated so it can adjust power so you don't have noticeable power changes when the a/c cycles on and off. This is why is says to run the car at idle for 3 mins both without and with the a/c selected so it can work it out.

When doing a cal, is instructs you to turn the a/c off, so it being activated shouldn't have an effect. It can't hurt to do the TLA first, as you have to warm up the car for both procedures. Interestingly the doc that PV points to is a 987/997 doc, but the 981 and 991 workshop manuals don't talk about the TLA procedure as a prerequisite for a cal. No idea why.

Which cal to perform? I always do the one without part replacement. I've never done the one with a part replacement and never had issues doing the other. It has always worked. Well, it didn't work on one occasion and that was when our prototype sensor was giving outputs beyond the limits.

When you do the cal, it will list a bunch of prerequisites, which include leaving in P, park brake on, brakes pedal pushed (put a stick on it) and the temp to a certain level. The wheels are on the ground and the car doesn't move.

All error codes must be removed from the TCU, and I would ensure all are removed from the DME also. It will let you do the cal with the errors, but it will most likely fail and give you a big headache getting everything sorted afterwards for to you attempt another cal. Moral of the story is to ensure these control units have zero errors prior to starting. Note you can't delete TCU or DME codes with the engine running. It needs to be stopped with the key turned on.

Something that burned me on one occasion was that the filling procedure for the 981/991 requires you to run in 1st for 2 minutes to fill the gear oil cooler, and then top up the fluid. The problem is without a good cal, as soon as you select 1st the transmission will fail and you will be troubleshooting to get the TCU clear again to start the cal. There is plenty of gear oil in there to conduct a cal without needing to top up. That can be done once the cal is complete. This isn't a factor for a transmission without a gear oil cooler, but I thought I mention it for others reading.

The cal is done in two parts. First is the shift rod cal, which is getting all the distances from the distance sensor stored. It will clunk for a few minutes as it selects gears multiple times to gather all of this.
If the distance sensor is OK, then it will finish the shift rod cal and move to the clutch cal. It then activates the clutches over and over at different RPMs. You will hear the engine groan as it activates and tests what it needs to do for accurate clutch activation. If the shift rod cal fails it won't move onto the clutch cal.

Shift rod cal takes about 2 minutes.
Clutch cal takes about 7 minutes.
It's all automatic, with zero inputs required once it has started.
Of note, the clutch cal stagnates from what I've seen at the 24% and 74% mark for some reason. It looks like the whole thing has stalled, but there is clearly something going on behind the scenes, as this always happens and seems to be normal.

At the end it will instruct you to turn off the engine and turn back on without starting. It won't disconnect PIWIS, just do what it says.


Video here of the procedure.

PV997

That's the first reasonable explanation I've heard for the intent of that procedure. It's been floating around on forums as a "PDK reset procedure" for years (see link) but the AC on/off part never made a whole lot of sense.

https://www.6speedonline.com/forums/...procedure.html

I think it may also actually clear some of the adaptation tables too based on anecdotes and what comes after it in the procedure but that's speculation.

Note that the document I referenced is actually a service bulletin (PDK Transmission Diagnosis: Symptoms and Repair Procedures) so maybe it include info specific to problematic PDK conditions not covered in a service manual. After calibration it calls out a lengthy "adaptation drive" and "touchpoint adaptation" (sections 3.9 to 3.14) that likely updates the adaptation tables to improve drivability and shifting precision.