Copter: parachute release is separate from crash-check allows possibility of parachute not being deployed in crash

[rmackay9]

The code for the regular crash-check is quite separate from the parachute release check.

They use almost exactly the same logic except that the parachute triggers after 1sec of loss of control while the crash-check triggers after 2 seconds. Also the parachute-check checks that the barometer says we're falling.

The separation of the logic means we're open to the possibility of the crash-check triggering without the chute deploying. Maybe that should be possible because we could imagine a simple flip on take-off (we might not want the parachute to deploy in this case).. but perhaps we should at least combine the logic and possibly delay the crash-check timeout if a parachute is installed (to give time to see if the vehicle is falling).

[rmackay9]

This has been partially resolved in that the parachute should deploy more easily now reducing the chance that the crash detector will fire first.
1f2d17e.
I'm going to move this issue out of the AC-3.4 milestone because I think the above fix is sufficient.

[Naterater]

Is the fix sufficient enough to close this issue?

[rmackay9]

@Naterater, i'd forgotten about this.. i think we should probably leave it open as a reminder..

[peterbarker]

Wow. Talk about timing.

I hit this problem just now when writing
#10208

[Naterater]

@peterbarker, since that PR is merged, does that solve everything here?

[rmackay9]

@Naterater, the autotest is there now and highlights the issue.. but the underlying issue itself has not been fixed

[peterbarker]

./Tools/autotest/autotest.py build.ArduCopter fly.ArduCopter.Parachute --gdb --debug

[tridge]

related to this, should we fire chute on armed watchdog?

[peterbarker]

STABILIZE> param fetch SIM_ENGINE_FAIL
STABILIZE> Requested parameter SIM_ENGINE_FAIL
SIM_ENGINE_FAIL = 1.0000000
height 30
height 20
APM: Crash: Disarming
APM: Disarming motors
DISARMED
height 10
APM: PreArm: DCM Roll/Pitch inconsistent by 26 deg
Timed out looking for BANG
AUTOTEST: Exception caught: Traceback (most recent call last):
  File "/home/pbarker/rc/ardupilot/Tools/autotest/common.py", line 2232, in run_one_test
    test_function()
  File "/home/pbarker/rc/ardupilot/Tools/autotest/arducopter.py", line 2275, in test_parachute
    self.mavproxy.expect('BANG')
  File "/home/pbarker/rc/ardupilot/Tools/autotest/pysim/util.py", line 406, in _expect_callback
    raise pexpect.TIMEOUT(timeout)
TIMEOUT: 60

[ddomit]

I think this issue is still present! Im doing the tests by arming in stabilize, then rotating the quad more than 20 degrees and i get a crash and disarming motors but not releasing parachute!

Im in beta 4.1 and manual tests with the remote do release the parachute.

Here is the log https://drive.google.com/file/d/1Eha9vKNj-sIjxJ-pdNBw6bcIsmEnjLZX/view?usp=sharing

[ddomit]

Maybe Im missing something or the documentation (here https://ardupilot.org/copter/docs/parachute.html) is outdated, how can i test the automatic deployment of chute?