LDU Overcurrent Sense Issue/Question [SOLVED]

[P.S.Mangelsdorf]

So as many know I've had my car running with an LDU for a while, and 2 weeks ago had what appeared to be a catastrophic failure. I now don't think that's the case, but have an enigma on my hands. I'm running version 4B of the control board.

For background, here's the sequence of events: driving home with DC-DC on the fritz, I had a falling 12V voltage, and started seeing the CANBus temp gauge go nuts (jumping around all over the place), while slowing and making a sharp right turn, I lost power. Pulled over, initially thought the 12V battery was dead so went and got one and my multimeter, turns out the 12V battery was ok (not great, but 12.4V). Then used my phone to connect to the inverter and go step by step through the start process. When it got the start signal, it immediately showed an OVERCURRENT error. I assumed a contactor had dropped killing a power stage, and got a neighbor to tow me the mile home. Once home I opened up the bus bars and did a diode test on the inverter per Damien's video and everything tested out perfectly fine. I opened up the inverter and found a significant amount of coolant (I had an external leak that I think made its way inside) and the bottom of the control board had been fully submerged. I cleaned out the connectors and tested (several times) but still got the same instant OC error when given a start signal.

I've pulled the board out, and connected to just 12V power to test. The board powers up ok, and it appears that all the IC chips are getting 5V as they should. I didn't find any chips with pins connected to ground that shouldn't be. I decided to test the pins going out to the gate drivers, and found each 12V supply and ground to be fine, but found what seems to me strange readings on the Fault pins. Here is what I measured:

Fault A High 1.1V
Fault B High 2.9V
Fault C High 1.5V

Fault A Low 2.5V
Fault B Low 0.5V
Fault C Low 2.5V

I measured the capacitors attached to each of the fault lines and they appear ok.

That looks to me like there's some sort of issue on the Fault B circuit, but I don't know. This is getting beyond my knowledge, frankly I'm not entirely sure why there is voltage here at all; as I read the circuit diagram I would have thought these were inputs to the control board not outputs, but when it comes to this level of detail, I'm still hopelessly uneducated.

[Aragorn]

Those 6 lines appear to be inputs to a NAND gate.

The NAND takes the 6 fault signals, plus the output from the overcurrent comparator, and applies the NAND logic function. In effect, it will output low only if all 7 inputs are "high"

Thus in normal operation, all 6 fault lines, and the overcurrent line would be pulled up to 5v. If any of them are pulled down to 0v the NAND gate will switch on the PWM inhibit signal.

I dont know if its normal for those inputs to float like that. Seems a bit weird to me. However the datasheet does say unused inputs must be pulled to either 0v or 5v otherwise undefined states can exist.

For testing i guess you can perhaps pull the lines up to 5v yourself and monitor the output of the NAND gate.


Is the output from the overcurrent section behaving?

[P.S.Mangelsdorf]

Those 6 lines appear to be inputs to a NAND gate.


The NAND takes the 6 fault signals, plus the output from the overcurrent comparator, and applies the NAND logic function. In effect, it will output low only if all 7 inputs are "high"

Thus in normal operation, all 6 fault lines, and the overcurrent line would be pulled up to 5v. If any of them are pulled down to 0v the NAND gate will switch on the PWM inhibit signal.

Yeah that's why I was confused about them floating. Maybe I'm missing something but I don't see where the voltage is coming from.

Is the output from the overcurrent section behaving?

Well that's what I'm trying to figure out - is this signs that the NAND gate has failed and is falsely triggered and thus inhibiting PWM, or is it fine, and the problem actually lies in the LDU's gate drivers (i.e. are the gate drivers are signaling the fault)?

[johu]

I was under the impression they were pulled high on the control board but that's not the case. So there must be a local pull-up on the gate driver board. They should have a 5V level when all is well. This has been confirmed a few times when people did fault-finding.

Pull-ups are usually around 10k or more, so moisture on the gate drivers could pull the fault signals low.
I think there are also fault LEDs on the individual driver boards, maybe that helps?

[P.S.Mangelsdorf]

Pull-ups are usually around 10k or more, so moisture on the gate drivers could pull the fault signals low.
I think there are also fault LEDs on the individual driver boards, maybe that helps?

Good to know. I've been avoiding powering up the inverter with the cover off, but I think I'm going to need to do that. Question: If I power it up without HV connected, and there is a OC sensing related fault, is the board going to tell me, or does it need the HV to be present?

I may also try another LDU I have here, but it's untested so that may just introduce new variables.

[P.S.Mangelsdorf]

I think there are also fault LEDs on the individual driver boards, maybe that helps?

Yes, they do have fault lights, and I do have one phase showing a fault. This inverter is coming out for now, I'll try to rebuild it later. Looks like a bad gate driver, based on the phase passing a diode test.