Arducopter Heli V3.5.0-rc8 motor starts with interlock switch in off position. PH2. ( Dangerous )

Today I had a very strange experience. I flew my 450 heli ( PixHawk2, Trad Heli V3.5.0-rc8) and then landed it. Immediately after landing I switch off the interlock switch ( CH8) to shutdown the motor. It did shutdown. However it started to move intermittently until I disarmed it using the Yaw stick. In the image of log below you can see that RCIN CH8 is low and the RCOUT CH8 also goes low. However as you can see it gets some high values causing the motor to move. I have been flying this heli with this setup for a while and I never had this before. However today this happened twice. Does anybody has any idea why this might be happening. I am little injured so I only fly small heli in a small field nearby, but had this been a big heli, it might injure someone. Hence this is also a safety issue. I will appreciate any inputs.

I updated to 3.5.0 and had 3 flights. This happened again but only once. Since I was hoping it would happen and I was ready for it, I did not panic to disarm using the Yaw stick. Usually if I just shut the motor the heli disarms on it’s own in few seconds. However when this intermittent throttle happens then the heli does not disarm on it’s own. Today I waited for good 2 minutes ( as you can see in the log below) the motor kept starting in intermittent manner. As you can see input CH8 is low. I am starting to wonder if my PixHawk2 has gone bad since nobody has reported this issue. I have a couple of other helis and I have never seen this problem before.

The official report on the Pixhawk 2.1 the first batch of products, because the design of PCB and shell factors, there is the possibility of CH8 failure, you look at, I remember a special report on Facebook! Can be mapped through the channel to CH6, CH7 channel to solve!

That is a known issue with channel 1. The trad heli motor is controlled using channel 8. This seems to be something else. The motor does not run if the heli is disarmed using the rudder, this also cannot be explained by CH1 issue.

Here is a link where the issue is discussed.

https://discuss.ardupilot.org/t/sb-0000001-critical-service-bulletin-for-beta-cube-2-1/14711/58

@Sunit_Pal do you have any further reports on this? Has it continued to happen? I cannot see how this can be a software issue. The only thing I see that MIGHT be a possibility is that your Channel 8 low value appears to be maybe like 1100 instead of 1000. It might be just on the edge of where the software triggers the throttle active.

Hi Chris,
I do not have a solution for it yet. It is still happening ( it happens about 30% of times.) I updated to 3.5.0/3.5.1/3.5.2 and the problem seems to persist. I even build my own firmware for 3.5.2 ( just to test) and I still get this error. I think my RC values are OK. I am attaching the params file. I am also attaching a sample log when this happened. As you can see in the log even when I have switched off the motor ( using input CH8) the motor intermittently received signal (by output CH8). The motor moves quiet slowly and not at full RPM, since the output is not consistently high as you can see in the log. Also it does not disarm on it’s own when this happens. However I can force it to disarm using the rudder stick. Then the motor stops completely.

I must mention that one day I was flying on a very windy day on CA coast. The flight was very rough. After that flight I saw this happen for the first time. Perhaps too much shaking did something bad in the hardware but that is just my guess. After that day I occasionally get this issue. What is making me more worried is my other heli is a giant one running Pixhawk2 Copter 3.4.6. If something this like happens there, it can be very dangerous. But as of now, this has not happened in that heli.

I will keep exploring the reason for this and update this forum. If anyone want me to do some experiment, then I will be happy to do so.

Thanks
sunit_heli_params.param (13.7 KB)
no_interlock_mot_run.zip (965.9 KB)

Hi sunit,
You have H_FLYBAR_MODE set to 1. Is this actually a flybar heli, or have a downstream FBL unit? If so, why are you using Rate P gains? Or was that just an oversight in the settings?

Your RC cals and trim certainly looks normal. I would think if it was a hardware problem that it wouldn’t be logged showing the intermittent throttle out with the RC low on channel 8. It looks to me like the software is doing that. But don’t know why. The only thing I can think of is that your channel 8 RC limit is 1055 min. It looks like the actual signal out is 1100 or maybe a tiny bit more when you go governor low on the switch. Does your radio have a display outputs function where you can see what each channel is actually putting out for PWM value? Maybe verify that you don’t have a channel endpoint set wrong, or an old mix that’s still active causing channel 8 to not actually output the minimum value when you go governor low on the switch.

Edit:
I graphed it in APM Planner2 where I can see point to point values of the RCIN to the PIxhawk. It is at 1056 or 1055 when Pixhawk fires the motor back up, then shuts it off intermittently. The radio is not the problem.

Hi Chris,
Yes it is a 450 size flybar. I was playing around with the tuning values to see how it behaves in extreme windy conditions, that is why you see rate P values. I think you have a valid point about the channel 8 RC out. Perhaps that is the reason. I am using spectrum DX8 and not some cheap radio so this is surprising to me. I will go back and check the logs carefully for values of channel 8. However I have notices that sometimes the heli would not arm and mission planner would say “PreArm: Motor Interlock Enabled”, even though the switch is not engaged. This is further indication that the channel 8 does emit bad values.
Now say the heli is armed and flying. We land it and switch the motor off. Channel 8 starts sending bad values before the heli is disarmed which will make the motor run. I think that is what is happening here. It seems to be a problem with my radio. I will check the logs for all the times when this has happened. Looks like I will have to spend some $$$ in a new radio.

Sunit

Hi Chris,

I missed you EDIT before posting my reply. I see that RCIN is logged at 10Hz, however RC values are updated at higher rate. It is possible that the radio is misbehaving but since all the RC values read are not logged so we do not see them. This will explain the arming error due to Interlock. I will try to change the firmware to log at higher rate and see if I can get to log the bad radio values.

Sunit

Yes, that is why I graphed it in APM Planner2, then blew the timeline up so I could see each log point, using the popup that shows a box by the mouse pointer for the actual logged value. I did not see any misbehaving by the radio at the 10x per second logging interval.

One thought I did have, though - a normal endpoint value would be 1000 pwm. Your Spektrum DX8 is putting out 1055. What does the software use for the minimum value before it engages the motor interlock? Maybe the endpoint of your radio channel is right on the edge of triggering it in the software?