Skip to content

Conversation

sfuhrer
Copy link
Contributor

@sfuhrer sfuhrer commented Aug 13, 2025

Solved Problem

We have false positive motor failure detection due to wrong, tiny current measurement.
image

Solution

Not declare the current measurement valid if never above 1A

Changelog Entry

For release notes:

Improvement: Reduce chance of false-positive motor failure detection by discarding any current measurement below 1A

Alternatives

The 1A threshold is very arbitrary, can be changed. We're also investigating the ESC side, where the wrong measurement comes from.

Test coverage

Untested

@sfuhrer sfuhrer requested a review from dawr68 August 13, 2025 09:46
Signed-off-by: Silvan <silvan@auterion.com>
@sfuhrer sfuhrer force-pushed the pr-failure-detector-harden-small-current-measurement-main branch from 736d0d1 to f5bc059 Compare August 13, 2025 09:51
@sfuhrer sfuhrer changed the title FailureDetector: increase threshold for valid current to 0.1A FailureDetector: increase threshold for valid current to 1A Aug 13, 2025
@dawr68
Copy link
Contributor

dawr68 commented Aug 13, 2025

ESC used in this flight doesn't support current measurement for individual motors.
Value of 0.1A sounds more resonable for me - it will allow it to pass on a bench (motors off) - when current is usually below 1A.

@dakejahl
Copy link
Contributor

dakejahl commented Aug 13, 2025

It might make more sense to use the esc_status.esc_armed_flags instead of esc_status.esc[x].esc_current. Because I think what we really want is "only trigger failure detector if an ESC goes offline (telem timeout) after it's already been in use".

If that's not the case, and instead we simply want to detect when an ESC goes offline, we can just use the esc_online_flags and telem timeout.

@github-actions github-actions bot added the stale label Sep 13, 2025
Copy link

This pull request has been closed after being marked as stale with no further activity. Thank you for the time and effort you put into this contribution. If you’d like to continue the discussion or update the work, please feel free to reopen it or submit a new PR.

@github-actions github-actions bot closed this Oct 13, 2025
@dakejahl dakejahl self-assigned this Oct 13, 2025
@dakejahl dakejahl reopened this Oct 13, 2025
@github-actions github-actions bot removed the stale label Oct 14, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants