ODrive does not switch states to closed loop control when requested

Support
1

Hi,

I wrote a python script based off of “odrive_demo.py” to test the ODrive with a Harmonic Drive RSF-Mini 11B motor I’m using (Link to the page here: http://www.harmonicdrive.net/products/rotary-actuators/solid-shaft-actuators/rsf-mini). The motor seems to act appropriately to the calibration sequence to find to the encoder index but when I request a change of state to move to closed loop control and send a position command the odrive doesn’t change states.

Here is my script:

#!/usr/bin/env python3

from __future__ import print_function

import odrive
from odrive.enums import *
import time
import math

# Find a connected ODrive (this will block until you connect one)
print("finding an odrive...")
my_drive = odrive.find_any()

print("Bus voltage is " + str(my_drive.vbus_voltage) + "V")
print("Calibration current is " + str(my_drive.axis0.motor.config.calibration_current) + "A")

# Set some hardware parameters temporarily
print("setting some hardware parameters temporarily")
my_drive.config.brake_resistance = 0.5
my_drive.axis0.motor.config.pole_pairs = 4

my_drive.axis0.encoder.config.cpr = 400000

# Calibrate
print("starting calibration")
my_drive.axis0.encoder.config.use_index = True
my_drive.axis0.requested_state = AXIS_STATE_FULL_CALIBRATION_SEQUENCE

print("waiting for calibration to end...")
while my_drive.axis0.current_state != AXIS_STATE_IDLE:
    time.sleep(0.1)

# Closed loop control 
print("Changing state to closed loop control")
my_drive.axis0.requested_state = AXIS_STATE_CLOSED_LOOP_CONTROL

while my_drive.axis0.current_state != AXIS_STATE_CLOSED_LOOP_CONTROL:
    print("axis errors are:")
    print(hex(my_drive.axis0.error))	
    print("motor errors are:")
    print(hex(my_drive.axis0.motor.error))
    print("encoder errors are:")
    print(hex(my_drive.axis0.encoder.error))
    time.sleep(0.1)

print("current state is " + str(my_drive.axis0.current_state))

curr_pos = my_drive.axis0.encoder.pos_estimate
print("current position is " + str(curr_pos) + " ticks")

new_pos = curr_pos + 400000
print("sending to " + str(new_pos) + " ticks") 
my_drive.axis0.controller.pos_setpoint = new_pos

The while loop at the end is designed to halt until the odrive to change states to closed loop control but it ends up stuck here forever. Here is the print out:

finding an odrive...
Bus voltage is 23.866186141967773V
Calibration current is 10.0A
setting some hardware parameters temporarily
starting calibration
waiting for calibration to end...
Changing state to closed loop control
axis errors are:
0x41
motor errors are:
0x11
encoder errors are:
0x0

Upon looking into the error codes I found that they correspond to these flags

axis errors are:
ERROR_MOTOR_FAILED = 0x40
ERROR_INVALID_STATE = 0x01

motor errors are:
ERROR_PHASE_RESISTANCE_OUT_OF_RANGE = 0x0001
ERROR_CONTROL_DEADLINE_MISSED = 0x0010

encoder errors are:
no errors

I’d appreciate any help. Thank you!

2

Are you on firmware v4.6? I’m pretty sure @madcowswe fixed the double motor error thing.

Please check your motor resistance (odrv0.axis0.motor.config.phase_resistance) and let us know what you see. You can also try increasing the calibration voltage (odrv0.axis0.motor.config.resistance_calib_max_voltage)

1 Like
3

I am definitely on firmware v4.6. What do you mean by double motor error?

I’ll be sure to test the phase resistance and try modifying the max calibration voltage in the next day or two and let you know what I find.

4

So I read 0.2364 Ohms from the phase_resistance once I increased the calibration max voltage to 5V. That led me to find an encoder error of 0x20 for an encoder index not found.

I realized that I may have gotten the wrong info for what the CPR of the encoder was so I started trying to change it. In the process of testing this one of the extra wires coming out of the motor fell up against the top of the ODrive. I quickly turned it off and removed it. On the following runs there was no visible damage to the board and the motor still calibrated as it did before, not able to find an encoder index but the measured phase resistance changed to .17 Ohms, the ARM chip got very hot, and the bus voltage read in was around 13 V! So I think I may have fried something.

Is there a good way to check whether the board is actually broken?

5

Hm yes I think the analog section of the microcontroller probably got fried. If you want to try to recover your board, you could swap the microcontroller out, it’s this part.
Of course it’s not possible to know for sure if there are other fried components too. Do you know where on the board it touched?

6

Ok thanks, I’ll purchase another one of those microcontroller chips and do my best to replace it.

I don’t know if this was exactly the way it fell but it had a strand of exposed shielding wire laying on the ‘M0’ marking and possibly another wire or two on the back side of the IC like this.

This didn’t seem to inhibit its ability to attempt calibration and drive the motor some in the process though so I would like to think that none of the other components were affected.

IMG_20181101_174323
IMG_20181101_174323.jpg3024×4032 1.33 MB

7

Hm I guess it’s possible that the DRV chip is also affected, hard to say until you replace the STM32.