Fullyield

Normal at high rpm’s, max torque and in Sport/Sport +. I was surprised and impressed first time it happened to me. That is why PDK is faster than a MT. It minimizes rpm/torque loss during shift.

t-design

Great news! We had faith in you!
Did you end up taking the gearbox out of the car or not?
Did you do distance sensor calibration? Did calibration went smooth?

workkyle

Thanks, I feel better. I also went out and did various WOT take offs in sport, sport+, manual and recorded them. They sound just like a YouTube video of a 997.2 PDK doing launch control.

Niklas Vemdal

I kept the gearbox inside the car, not a big work if I should redo it. No transmission calibration done. Worked well on Saturday evening, I only did the steering wheel calibration and all fault code could be deleted and they didn’t come back on short drive. I took a longer drive on Sunday morning and got some issues, gear fixed in 3rd gear. Fault code P17D9, impossible to shift. After restarting car it worked again, but after a while similar issues on different gears. I believe I need to calibrate so I have ordered time for that at workshop next Wednesday, my Autel can’t manage it. I will come back with updates. If someone have other ideas… please write here.

Niklas

PV997

Great work Niklas, and thanks for posting.

It would be very surprising is the T-design distance sensor worked flawlessly without calibration. T-design has stated that their device uses a new design and new components, it's not simply a copy of the ZF design. So it's electrical characteristics won't exactly match the ZF part under all conditions (though they are probably quite close). In fact, normal component variation will mean even the actual ZF sensor will vary from unit to unit.

I suspect calibration works much like that of the power windows after a battery disconnect (i.e. runs into mechanical stops and then correlates the sensor data with these known positions). It's a pretty basic tried and true concept that uses the known end points to determine the zero (middle) position and fine tune the appropriate +/- scaling thus compensating for sensor variability. If anyone can confirm this or has additional details please post.

A Chinese PIWIS2 clone will do the calibration if your shop doesn't have one, it's about $425 shipped. There's a thread on it here and another good one over at Planet 9.

t-design

Sir, what you described is all correct. Still, we know some cases where it works fine without calibration. Yet, most of the cases do require calibration.

Niklas Vemdal

I realized that I should try to add some more clutch / hydraulic oil - maybe it wasn’t completely full after one time. I did it yesterday in “fill oil mode” and drove 20 min without any problems, almost at home it happened again. Stuck in 3rd gear and overheating clutch oil. Stopped, restarted and came home with no issues. Today I refilled again, not sure that I got in any more oil but believe I closed with a bit more oil. Have now been driving for 30 min, slowly, fast and accelerated hard - came home without any issues at all. Should I calibrate or not… doesn’t feel that it’s needed and I’m a bit afraid of getting new issues when doing calibration. What do you think? I will take car to my work tomorrow (45 miles), I have booked calibration at workshop tomorrow afternoon - need to decide before that.

t-design

Can you post PIWIS readings from yesterday?

Niklas Vemdal

What is coming first? P1870 or P17D1? My believe is that there was some issue with hydraulic (not enough oil?) creating P17D1. My guess is that P1870 is a generic code when shift change doesn’t work. With this the Temperature increase after issue with shifting, clutch has to work harder on high gear with low speed, thus setting P17F0/1.

It runs so good when it runs so I don’t believe calibration should be needed… but I don’t know for sure

thanks!

amet

1870 indicates that the calibration should be done, I have only seen that error after failed calibration attempts.

using autel, you can check the distance sensor position values. Could you shift through few gears and show us those values please

Niklas Vemdal

Started up car this morning, no messages in DIM but 1870 set after yesterday’s drive.

Driving 45 miles (mostly 7th gear) without issues.
Slowing down / parking resulted in P17D1 again, has to restart car and it works again. I will go for calibration today.

Below pictures shows positions in different gears, P, R,1,2,3… if I got them in correct order.

PV997

Those values for the distance sensor don't look right to me. On first glance they seem to have the correct gears engaged so that doesn't look like an issue. But look at those that aren't engaged or set up for the next shift (for example 4-6, 5-7, and 1-2 in the first photo). Those should be closer to zero in my experience, within a few tenths of a millimeter, yet every one except 1-3 is -0.7 mm to -0.9 mm. Curious what others have seen. I don't think there is anything wrong with the distance sensor but simply that its electrical characteristics don't exactly match the previous sensor. This is exactly the type of stuff calibration is intended to correct (part to part variation) as I described up in comment #620.

The overtemps would have me a little concerned as they could be indicative of clutch slipping. Unlikely I think, but keep an eye out for speed mismatch faults that indicate the input and output shaft speeds don't match. You really don't want excessive clutch wear.

Obviously I cannot guarantee anything and move forward at your own risk, but I think this is pretty much classic evidence of why calibration is required and how it manifests if it isn't done. The whole point of calibration is to characterize the effect of part to part variation within the PDK and apply compensation in response. If calibration is skipped not only isn't this done, but you still have the wrong compensation from the previous parts.

t-design

Niklas,
Those readings seems to be correct. But we still recommend calibration.

t-design

Here's sample from one of our 991's :

PV997

Nice video T-design, should be helpful going forward.

Notice that the shift rods that aren't engaged (zeroed) read in the 0.3 mm to 0.5 mm (+/-) range. This is almost exactly what I see in my car which has the factory distance sensor. Niklas's photos show his in the 0.7 mm to 0.9 mm range for 3 of the 4 shift rods (the other one was 0.3 mm). This is what caught my eye as they look out of family, I'll be very interested to see if calibration changes these to be closer to zero. Again, I don't think there is anything wrong with the sensor but noting that these minor differences are likely corrected via the cal.