ERROR_BRAKE_RESISTOR_DISARMED on ODrive S1 during Deceleration

[Labrey]

Hi everyone,

We are experiencing ERROR_BRAKE_RESISTOR_DISARMED when trying to decelerate our motor too quickly. Our setup:

Motor setup: A 315g object is suspended vertically from a 4cm diameter spool. We are using a
Dual shaft motor - D5312s 330KV Odrive motor.
Power supply: We tested both a 24V DC power supply and a 24V Li-Po battery.
Brake resistor tests: We tried 1Ω, 0.5Ω, and 0.25Ω resistors, but the maximum deceleration before triggering the error didn’t change much.
Observations:
Going up is fine up from 0 turn/s to 40 turns/s at 85 turns/s².
Going down, if we try to slow from 10 turns/s to 0 turns/s at 15 turns/s², we get ERROR_BRAKE_RESISTOR_DISARMED.
Brake resistor flag: brake_resistor_saturated is turned on.
Configuration: Full settings are available here: ODrive S1 Config : O_drive_configuration - Google Docs

Is there a way to prevent this error while maintaining fast deceleration?

Any insights would be greatly appreciated!

Thanks in advance!

[solomondg]

Hi! Can you take a look at the top of your board, and let me know what the revision number is shown here?

Also, can you take a picture of how the encoder is mounted? Is there any risk of the shaft slipping, do you think?

[Labrey]

Hi !
The version of our board is REV A1.
I don’t think that the encoder slipped, I provided a picture of the encoder mounted to the 5mm shaft.

[Labrey]

Is the current sent to the brake resistor limited by abs(dc_max_negative_current) ?

It seems like we didn’t define max_regen_current in our system, but on GitHub code this value is initialized to 0,0f which is coherent with our power supply not being able to dissipate negative currents.

We will be able to do some test tomorrow, we will try with a 0.1ohm resistance, do you think of other test to solve this issue ?

[Labrey]

Hello,

We noticed that the error was due to overvoltage.
To temporarily fix the issue, we:

• Added a 10 mF capacitor in parallel of the power supply.
• Increased the upper voltage limit to 32V.

This solved the problem, but is it normal for the voltage to rise like this?

After further testing, we found that the ODrive is sending current back to the power supply during deceleration. Here is a picture showing the current flow from the power supply and the additional capacitor to the ODrive.

We measured a minimum of -2.8A (with a scaling factor of -0.02 on the readings).

Given that we configured no negative current, is this behavior normal?
How can we prevent the ODrive from sending so much current back to the power supply ?

Here is our configuration:

odrv.config.dc_bus_overvoltage_trip_level = 32
odrv.config.dc_max_negative_current = -0.01
odrv.config.max_regen_current = 0

Thanks in advance for your feedback!

[solomondg]

This typically will happen if either your brake resistor resistance is too low for the amount of current being regenerated (e.g. Vbus/Rbrake < Iregen), or if your actual brake resistor resistance is higher than the value specified to the ODrive; this is typically caused by poor wiring connections, or the brake resistor being a bit out of tolerance. If you’re confident you’re not regenerating more current than the brake resistor can dissipate (Vbus/Rbrake > Iregen), then you can try increasing the stated resistance value to the S1 by a bit (e.g. specifying value of 2.1 for the brake resistance, if you only have a 2 ohm brake resistor).

[Labrey]

The first issue is unlikely in our case, since our brake resistor should be able to handle the regenerated current. However, the second point about the actual resistance being higher than the configured value makes sense since we use a 0.1 ohm resistor and could explain the issue. We will test this as soon as possible by slightly increasing the configured brake resistance and report back on the results.

[solomondg]

Okay! Note the S1 brake bridge is only rated to about 80A max – a 24V supply with a 0.1 ohm resistor would result in 240A, so I’d recommend using a 24/80 = 0.3 ohm resistor max – I think 0.25 ohm should be fine.

[Labrey]

Okay, thank you for the warning, we used a 0.1ohm for test purposes because it was the minimum resistance value we read online but from now on we will use the 0.5ohm provided resistance.

[Labrey]

Hey Solomon,

Sadly, this didn’t work. We used a 0.5 ohm braking resistor and tried adjusting the stated resistance from 0.45 ohm up to 0.7 ohm, but the power supply voltage still rose to the same peak as before with high variability.

We also experimented with modifying:

• dc_bus_voltage_feedback_ramp_start = nominal bus voltage +2V
• dc_bus_voltage_feedback_ramp_start = nominal bus voltage +6V

But we didn’t notice any change.

Any other ideas on what might be causing this?

Thanks!

[Samuel]

This is still indicative of the true brake current being lower than the calculated brake current, perhaps due to a connection issue or damaged FET.

Here are some more things you can try:

• Plot brake_resistor0.duty in the GUI while decelerating to verify that it’s really trying to use the brake resistor
• Plot brake_resistor0.current_meas in the GUI. But this is only informational, the brake current sensor has some limitations which I don’t remember from the top of my head. So no worries if this shows 0 but still interesting to look.
• Since you have an oscilloscope, you could measure BRAKE+ to DC-, to see if it’s really switching. You could also measure BRAKE- to DC-, which is the voltage across a shunt resistor.

To simplify measurements, you can force non-zero brake duty cycle even when you’re not decelerating and the ODrive is in IDLE. You can do this by setting brake_resistor0.additional_duty = 0.02. This is for low level diagnostics only and must be used with caution. In your case this works out to (24V*0.02)^2 / 0.5Ω = 0.4608W which should be safe.