SDU Board error message overcurrent

[johu]

yes the faults must all read high, maybe no power into the gate driver?
The gains are hidden so they aren't changed accidentally. Yoz can display them with "stream 1 il1gain" on the command console on the dashboard

[MoonUnit]

Yes, I realised late last night after I'd posted that I need to actually put the board back into the inverter to connect it to the power stage and then I should get something back via conn5 to make the fault lines go high .. I am an idiot ...

I will try again later.

[MoonUnit]

OK so, the OVERCURRENT error is fixed and I will update the wiki with the steps taken but in summary:

The board does need to be connected to the power stage of the inverter to bring values for faults A, B, C etc up to the OI board. The HVIL disable also needs to be bridged - if any of these inputs to the logic chip are low the OVERCURRENT error is thrown.

Also (for v9 board and version 5.35R of sin release) when first testing the board, you may see 'UdcBelowUscSw|BrakeCheck' which prevents you from starting up the inverter. Setting UdcSw to zero and cruisemode to switch in the parameters should put the board into the 'awaitstart' mode and allow startup.

So things are better but I do still have an errant signal from a current sensor, plot attached. Interestingly the range has halved from 14A to around 6A now that I have plugged the OI board into the power stage of the inverter. I would like to understand what the problem is before I button everything up - does anyone have any ideas?

EDIT - wiki updated.

[MoonUnit]

I am getting nowhere with this, I can't pin down a reason for the different readings from IL1 and IL2. Voltages from the sensors themselves are the same, there is a slight variation (6mV) once they have been through the conditioning circuitry on the board but that's not enough to account for the range difference. The averages of the signals are much closer to each other, so maybe the problem is illusionary. Maybe when there is current on the HV lines any noise that might be causing the differences is swamped by signal.

[crasbe]

You said on Discord that you measured the signals with your analog oscilloscope, can you show some photos of how the signals look there?

Can you also measure the +5V power supply rail?

[MoonUnit]

Hi - I was measuring around 4.93V at the points on the board I thought might be relevant - so at current sensors and at the fault logic chip I5. I didn't measure 5V everywhere it is used ... maybe I should. The scope images are dreadful - very hard to read, faint, unstable. I can have another go at it, but I don't want to be going down the wrong rabbit hole unnecessarily. All I could really determine with the images was that the signals were both noisy and looked very similar.

Johu had pointed me at the IL1gain variable which is hidden from the user but can be exported via the webinterface, however the web interface prints out exactly 0 (not 0.00), and I can't get it to show IL1gain. I attach a short log file from the web interface data logger which averaged the values over 100 readings for IL1 and IL2, and when averaged like this the two values are close together so maybe there's nothing to worry about ...

[johu]

Yes you can also choose to ignore it. The values are not used in any kind of control loop so it's ok for them to be inaccurate. Plot them again with some load on the motor

[MoonUnit]

Thank you Johu, that gives me confidence to ignore it which is by far the easiest 'fix'! As and when I do get a load on the motor I'll see what they read then.