Leaf Gen2 Heatsink temperature sensor

[johu]

When designing the Leaf expansion board I had clearly found a signal that correlated with the temperature of water I put into the cooling channels. The signal is PWM, 500Hz if I remember correctly, but it is not important here.
Here is what I found:
18°C 46,50%
25°C 48,60%
30°C 50,12%
35°C 51,90%
40°C 53,43%

I extrapolated a few more values from that curve, down to -10 and up to 100°C. But now it doesn't work. The value is now 30% at 20°C and thus the temperature is stuck at -10°C

To get at least something I just subtracted 15% from all dutycycles, now I get a very strange behaviour. The supposed temperature rises by up to 50°C within a second when accelerating hard but also drops at the same speed when coming off the accelerator. So I'm not sure if what I'm seeing is really the temperature.

Could anyone do a temp test with their Leaf inverter? The signal I use is on CN501 pin 14.

[IsakSoderlund]

Here's a screen shot of the Nissan leaf manual. You can see the resistance and the temp in both C and F. This is the temp sensor on the motor not the inverter. I don't know which one you are using here. The both might work the same way though?

Regards
Isak

[johu]

Yes the motor is the blue line, that works well. The inverter is giving me the headache. Knowing the resistance/temp curve wouldn't help, I need to know the resistance/dutycycle curve.

[Jack Bauer]

Pure guess but you could be seeing a die temp reading from within the igbt. They will heat up that fast on load internally. That's why current ratings at room temp in the datasheets are garbage.

[johu]

Yes I've thought of that. Of course then the 85°C derating would be much too soon. Will make that a parameter.

[SciroccoEV]

Arlin Sansome posted up a video recently which showed temperature rising and falling pretty fast and I think mentioned it being the die temperature of the centre IGBT.