openinverter forum

I have an SDU running openinverter revision ce8c1e55fb91cd928c4b5ad4ccc56940596640ee - parameters attached.

For some reason I cannot get regen to work except in a very small band at around 5mph. If I set offthrotregen to -30, it seems that potnom is set to -30 as expected when the throttle is released, but there simply isn't any noticeable regen at speeds above 5mph. If i allow the vehicle to slow down naturally, I feel absolutely no regen until the vehicle slows to 5mph, at which point I feel very sudden and very noticeable braking, which brings the vehicle to a near stop.

This suggests to me that regen is absolutely working, but something is inhibiting it at higher speeds. I can't find any obvious explanation in the source code, but I'm wondering if it could be a bug as this version with hill hold has not been released and widely tested yet.

Tomorrow I will repeat this test and obtain a detailed log of the DC current, speed, and potnom. If possible, I will also add fslipspnt and ampnom to my logging.

Just went for a drive to gather the data, and it seems I'm somewhat incorrect about the symptoms. I *do* have regen at high speeds, it's just negligable, and then at 5mph something unpleasant happens...

Here's a graph showing a 0 to 60 to 0 run. It shows that I'm getting around 22A of DC regen current, but as the speed drops (to around 25Hz maybe) there is a huge spike in AC current, which seems to exist only between the inverter and the motor.

I assume the solution here is simply to increase the value of regenrampstr to eliminate that current spike, however I am puzzled as to why the regen is so little. At a potnom of -30 I'd expect it to be quite a bit more noticeable.

fslippnt is -2 as it should be, and unfortunately I messed up the command to log ampnom, so I don't know what it was doing throughout.
I will do another run shortly with offthrotregen increased to -40 and regenrampstr increased to 25. Hopefully this combo will result in noticeable regen without the low frequency current spike.

Fixed. Raising regenrampstr to 40 eliminates the current (and braking) spike at low speed. Increasing offthrotregen to -40 creates a meaningful amount of regen braking (-37A) at all speeds. Interestingly
It doesn't seem to be possible to map ampnom to CAN, at least not at the same time as fslipspnt. Each time I try, it appears to work, but then I end up with fslipspnt mapped twice.

In any case, marking this as resolved, hopefully it helps anyone else with the same misconfiguration!

Yes CAN mapping of these is dodgy. They both have uid=0 as I don't want them stored to flash and that makes it impossible to recover CAN mapping