I’ve been having some issues with connecting to the USB ODrive v3.6 56V. Last weekend, I wired up two ODrives to a RPi, which seemed to work great. The motors were able to successfully calibrate and run to positions on command. This translated to this morning, where the project continued to exhibit the same behavior. However, this didn’t seem to persist: after attempting to connect the USB directly to a workstation, the ODrives no longer seemed to want to connect to anything – running lsusb on the RPi, a workstation, and a laptop all seemed to not detect the device whatsoever. The power light is on, the only issue seems to be the connection. All the cables were verified to work as expected.
To solve the issue, I attempted to reflash the boards with new firmware. The flashing was successful, but didn’t seem to change anything connection-wise. However, after repeated flashes and power-cycles, the one of the boards started smoking, with the USB housing melting and fusing the cable to the board. This is shown below:
I don’t know how plausible this is given the actual electrical design of the ODrive, but any time you melt a low-voltage connector like that, it implies that you’re putting a ton of current through the shield/ground of that connector unintentionally.
My guess is this: the motor ground and USB shield are probably shared on the ODrive. When you had the odrive connected to a Pi, ground and power on the pi are fully isolated due to being plugged into a wall wart, which maintains full isolation between the AC and DC sides.
When you plugged USB into a workstation, USB ground became earth ground due to the workstation shielding design.
That then interacted with however you’re powering the motor-side of the ODrive. Perhaps motor supply (-) is connected to a power supply or battery or something such that it’s actually a voltage below ground? In which case you’ve set up a loop where the motor supply is shorted through the USB connector through the chassis ground of the workstation through earth ground.
I’d go looking for some electrical loop of that sort.
Wonder if you connected the break resistor. What power is your motor?
As motor speed ramps down, it generates electricity and need to go to the resistor or else push up voltage on motor power input line or flow into unexpected path, may be?
Thanks for replying! I also suspected a ground loop initially, but the Pi and the workstation were grounded from the same 24V source, so I don’t think that’s an issue.
We are using a brake resistor. However, I don’t think the brake resistor would make any difference: The issue arose when trying to just power on the board with a USB hooked up rather than when running any motors.
