Tesla LDU and incorrect UDC reading problem

[Boxster EV]

Hello All,

Whilst driving a few days ago, my large drive unit massively derated all of a sudden despite a full battery.

Upon power cycling and checking the spot values, the main contactor would not close because the last error was UDC below UDCSW. It was showing <200vdc. Got the car home and ran a few checks including reloading the sine firmware, but the issue remains. I also checked that I was getting 390vdc on the Tesla busbars, but still UDC reported at <200vdc (speradically, sometimes 160 another nearer 200).

Any educated guesses whether this could be a voltage sense issue on the open inverter logic board, or alternatively, could the issue point towards an inverter problem?

The car still drives, although all the firmware calculations are seemingly thrown off by the voltage discrepancy.

Either way I’m resigned to removing the LDU from the car, but in the meantime would appreciate anyone’s wisdom.

Other info: car has done about 30k miles in past five years with this v4b logic board.

params 5.json

(1.69 KiB) Downloaded 426 times

[johu]

Just a guess really: sounds like moisture either on the Vsense electronics or the connector or the logic board

[P.S.Mangelsdorf]

No idea why it would drop while driving like that.

I did look at your params, and you are running different udcgain and udcofs than I am. Obviously, if those are working for you, great, but its probably worth checking them against prior parameter sets of yours, just in case something was inadvertently changed.

Yours
"udcgain": 7.5,
"udcofs": 0,

Mine
"udcgain": 7,
"udcofs": 154,

[johu]

Does the voltage drop (close) to 0 when not allowing HV into the inverter?
186V sounds suspiciously like the floating value when Vsense is not connected to the OI board

[Boxster EV]

Does the voltage drop (close) to 0 when not allowing HV into the inverter?
186V sounds suspiciously like the floating value when Vsense is not connected to the OI board

I will check this evening. Thank you.

[Boxster EV]

Here’s the reading with HV:

Here’s the reading without HV:

[johu]

ah, that's with Mangelsdorf config then?
So the Vsense is still connected to the logic board then. Like said this could point to a moisture problem.

You could mitigate for now by setting udcgain in a way that the reading is kindof sane. Set udcofs=0 again. Set udclim to 1000V to prevent the inverter faulting out should the reading return to normal.

[Boxster EV]

[Boxster EV]

The car became more and more horrible to drive this week so I pulled the motor out. Whilst the inverter was completely dry, looks like I have some remnants of a previous leak from (3 years ago).

[Boxster EV]

To conclude this thread, I found the issue to be on the Tesla PCB that attaches to the main busbars. I think it manages the capacitors for the voltage sense aspect.

Having inspected this, I found that the board had burned out:

I managed to source a replacement from Ukraine part number 6009072-00-A for just £70. Interestingly Tesla seem to have updated the design and placed a plastic housing over the blue components. Anyway, I fitted it and put everything back in the car and it’s working perfectly again.