[solved] Strange Crash and Prearm IOMCU error on cube?

vosair · April 24, 2019, 8:53pm

Hey friends,

First off, I’m aware of the grounding notice for 2019 cubes. This is a 2019 cube running the latest master which from my understanding contains a possible fix to switch IMU in flight if a problem is detected. I’m also running the latest MP beta which should detect if there’s an inconsistency with the IMU and throw a warning upon connection to the cube. I did not get a warning when connecting. I understand that there’s a chance that this crash was caused by the this issue, even though it doesn’t seem like it according to what Phillip has posted.

I was doing a test flight with a new battery configuration a meter off the ground. It was flying great and holding position very well, when at about 8 minutes into the flight it turned on the yaw axis and crashed. It was a pretty slow crash and didn’t damage anything. The only thing strange I noticed is after the crash is an error “PREARM - IOMCU ERROR”. I’ve never seen this error before.

Any idea what happened? I’ve posted the log from the flight below.

peterbarker · April 24, 2019, 11:34pm

First off, I’m aware of the grounding notice for 2019 cubes. This is a 2019
cube running the latest master which from my understanding contains a
possible fix to switch IMU in flight if a problem is detected. I’m also

I see no evidence in this log of estimation failures, and the vehicle was
quite happily alive through the entire event.

I was doing a test flight with a new battery configuration a meter off the
ground. It was flying great and holding position very well, when at about 8
minutes into the flight it turned on the yaw axis and crashed. It was a

Check the motor / ESC / whatever on channel 10; it looks like it failed in
flight. Not suddenly - possibly a mechanical problem with the motor?

pretty slow crash and didn’t damage anything. The only thing strange I
noticed is after the crash is an error “PREARM - IOMCU ERROR”. I’ve never
seen this error before.

That’s a new one, and I believe it is unrelated to your problem here.

vosair · April 25, 2019, 5:26pm

Thanks for taking a look! From my understanding, the IOMCU controls the motor outputs? That’s what made me think it might be related.

proficnc · April 25, 2019, 5:39pm

I agree with peters assessment here, but how did it arm with the prearm failure?

vosair · April 25, 2019, 6:45pm

The prearm failure showed up on MP after the crash.

peterbarker · April 26, 2019, 1:34am

The iomcu errors that led to the prearm failures hadn’t occurred when the
vehicle armed. Those iomcu errors actually turn out to be bogus/expected,
and tridge has a PR to correct that issue.

As master currently stands, you can not arm with any “internal error”
having occurred (776c111d3cc094dc73172acf89a01705b3183d87) - and this
iomcu error is considered to be an internal error.

Incidentally to this (@meee1) , MAVProxy monitors the SYS_STATUS message
to see if any internal errors have occured in the software:

github.com

ardupilot/MAVProxy/blob/master/MAVProxy/modules/mavproxy_console.py#L353


    self.console.set_status(s, s, fg=fg)
for s in announce:
    bits = sensors[s]
    enabled = ((msg.onboard_control_sensors_enabled & bits) == bits)
    healthy = ((msg.onboard_control_sensors_health & bits) == bits)
    was_healthy = ((self.last_sys_status_health & bits) == bits)
    if enabled and not healthy and was_healthy:
        self.say("%s fail" % s)
self.last_sys_status_health = msg.onboard_control_sensors_health


# check for any error bits being set:
now = time.time()
if now - self.last_sys_status_errors_announce > self.mpstate.settings.sys_status_error_warn_interval:
    for field_num in range(1, 5):
        field = "errors_count%u" % field_num
        x = getattr(msg, field, None)
        if x is None:
            self.console.writeln("Failed to get field %s" % field)
            self.last_sys_status_errors_announce = now
            break
        if x != 0:

The message initially said “land now” - because if those bits are ever set
then something “not right” is definitely going on and your vehicle should
not be flying (and may not be for long). tridge is averse to long speech
output, so now it is just “critical failure”

rbachtell · April 26, 2019, 9:16pm

@peterbarker Loading latest master on three of my early cubes now gives me an IOMCU unhealthily pre-arm. So the question is, am I dealing with a real or “bogus/expected” problem. All three have been flying well for the past few years and if I go back a few days to a previous master all is well.

Cheers, RB

vosair · April 26, 2019, 11:34pm

I installed an older master and was able to complete a full 27 minute flight without any issues or the IUMCU error. That leads me to believe that the crash was related to a bug in the latest firmware? @tridge this might be of interest to you.

tridge · May 1, 2019, 8:03am

I’ve fixed that now, please re-test.
Cheers, Tridge

rbachtell · May 1, 2019, 12:37pm

Confirmed, all is good now. Thanks

Cheers, RB

vosair · May 1, 2019, 7:51pm

Just to follow up, the crash was caused by my ESC hitting it’s temperature protection limit. I added more ventilation for the ESC and haven’t had the issue since.