pathfinder-for-autonomous-navigation/FlightSoftware

Fix Checkout Cases

Closed this issue · 4 comments

knk38 commented

Per Tanishq's notes two checkout cases: Piksi and Gomspace errored out and did not work. We need to fix these cases and also check other cases for correctness and make sure they finish gracefully.

Also update pprocs with current checkoutcase list (remove spin motors case for example)

Files That need to be looked at:

piksi_checkout_case.py (critical issue)
gomspace_checkout_case.py (critical issue)

  • Could be because piksi off, try changing the initial state or having it boot to standby before beginning or something

docking_checkout_case.py (check for issues)
dcdc_checkout_case.py (check for issues)
adcs_checkout_case.py (check for issues)

These cases hopefully wont require intense fsw changes but maybe they will?

Some cases are failing due to EEPROM causing it to not hit target initial state. We added an initial state doesn't matter feature to EmptyCase.

This requires double checking the reset for flight case, as cycle.auto is not being correctly set to false

Piksi Checkout (Failing)

Screen Shot 2021-09-26 at 5 49 56 PM

Looks like the NaN issues are real,

Gomspace Checkout (Failing)

Screen Shot 2021-09-26 at 5 51 58 PM

Initial thought was it is an eeprom issue, but I deleted eeprom and still got this error. Looks like the pre-boot is expecting the state to be 0 (startup) but since we are running a manual test the inital state is 11 (manual) may need to change the conditions in mission manager

Docking Checkout (Passing)

Screen Shot 2021-09-26 at 5 57 54 PM

Docking case looks good

DCDC Checkout (Failing)

Screen Shot 2021-09-26 at 6 00 15 PM

Probably something to do with the prop changes, looks like the prop fields arent false on initialization. (Note: i rebooted the satellite and deleted eeprom.json)

ADCS Checkout (Failing)

Screen Shot 2021-09-26 at 6 06 02 PM

Looks like ADCS Checkout has some issues as well, adcs monitor returning "not functional"?

@GomSpace, its an issue for two reasons, the screenshotted error can be fixed with check_initial_state false, the more nuanced issue is that piksi_off is interfering with the power cycle commands in way the text case isn't expecting.

@adcs, waiving as ADCS on HITL doesn't work

@Piksi, waiving because Piksi checkout case tells us nothing anyway.

knk38 commented

were fixed with PRs