Potentially very dangerous throttle bug

So edit the thread title so it’s more appropriate to the “nothing to see here” condition. There is no bug related to what you have reported…

Hm… this is not my thread :smiley:

Sorry about that, replied to the wrong post!

Hey guys I see You enjoying discussion, but can anyone answer my two questions above? :slight_smile:

p.s.

The thing is I am not dumb and I have read all the instructions, have seen many tutorials and so on before flying. I was even flying successfully before incident happened. There is somewhere in the manual that you must not fly lighter than during the tuning process. Maybe there should be warning that it is mandatory to do full re-calibration whenever engine power, propeller, voltage or weight of your quad is changed (weight to power ratio) otherwise a “jump” can happen.

There should be mode when pilot has full control over the aircraft. I remember cc3d board, there was flight mode manual and this allowed relatively safe testing. When FC was freaking out i just went back quickly to manual. There is no such option in ardupilot, even stabilize mode is affected by too many parameters.

I have also discovered that emergency engine stop not always stops and also sometimes quad doesn’t want to disarm :slight_smile:

I also have tested that after arming, when throttle stick is full down engines are in idle stabilization is not working (is off), also no response to pilots input for roll and pitch. This is unwanted behavior as it means, that when pilot drops throttle stick fully down in the flight - suddenly quad will fall off the sky like a rock. I have tested it with betaflight and desired behavior is that quad is always horizontal even with throttle fully down. It is falling off the sky, but leveled and still reacting to pilots roll and pitch. Howto prevent this situation in arducopter?

By the way, thanks for your comments I appreciate to hear that I’m not the only one getting surprises.

Stabilize mode is the most manual mode, it follows your throttle input and only tries to stay level (or follow pitch and roll input)

How to prevent the situation depends totally on what caused it - and that needs a log to examine.
You situation is definitely ab-normal. Even most fly-aways due to vibrations can be brought back safely in Stabilize mode.
There could be numerous configuration errors. For example your MOT_THST_HOVER is 0.5 where normally you’d expect around 0.2 and even less for an “over powered” quad. You’ve got MOT_SPIN_MIN,0.15 , and if MOT_THST_HOVER really should be less than 0.2 then that’s starting to be a stability and control issue. Not to mention that during take off the quad will be expecting to go to 50% throttle or more just to get off the ground and hover :frowning:

It would be interesting to see the MissionPlanner motor test in action with your quad - when do motors start to turn reliably, and when do they just start to make thrust ?
In MP motor test it shows a percentage you select or input your self, like 0.08% for example, and this translates directly to the MOT_SPIN values in parameters. So you might set MOT_SPIN_ARM as 0.08 and MOT_SPIN_MIN as 0.10 .
Even set MOT_THST_HOVER as 0.12 initially and see what it figures out during flight. MOT_THST_HOVER usually is “learnt” and you don’t usually set it except for an initial value if you suspect your quad is over powered at all.

We’ve tested the RC Input Option 31 Motor Emergency Stop on each multirotor we set up and it’s always worked - both when tested and when actually needed. Obviously your transmitter should have a switch dedicated to a channel for emergency stop, and the same RCxx_OPTION,31 set in Arducopter. You should be able to see it operate in the RC Calibration screen of MissionPlanner.

And of course with a small quad you can always hold onto it with relative safety for testing…

I thought the same about stabilize mode, that it will follow throttle input 1:1, but I am not sure about that anymore, that is why I asked if MOT_THST_HOVER is taken into account in stabilize mode.

Your advice to manually check MP motor test and write notes about MOT_SPIN values is very valuable. I will do so. I have stored this in my notes :slight_smile:

When it comes to the motor emergency stop it was longer story. Battery was drained a lot and I think what I have experienced is FC has rebooted because of 5V failure, but ESC didn’t cut the engines. I had low battery setting in arducopter set and I had ESC low voltage protection also. Both didnt go well and i was unable to turn off engines, because FC probably rebooted (low power). Anyway I have turned off battery protection in ESC and now I am flying using timer, an old school method that always works :slight_smile:

Isn’t acro the most manual mode? There is no stabilization applied!
Also there is a feature called air mode that makes stabilize work on idle / minimum throttle

Yes true, I forgot about acro

Airmode is really good option. That is what I was looking for.
Is it possible to arm quad in acro mode or it always has to lift of in stabilized mode? I prefer acro for testing and then switching to automatic modes when I am at least 3 meters high, so I have time to safe model from crashing if auto fails.

No. Acro uses the same rate and attitude controllers with same PID settings as stabilize. The only difference is that acro is rate command with attitude hold (3D Mode) and stabilize is attitude command. All multicopter’s require some level of stabilization for a human to fly them. As for the throttle control, I am not 100% certain but I believe there is no difference in that between the two modes.

1 Like

Thank You guys for attention. I think we can close this topic, as worked around. To summarize:

  1. Configure Emergency throttle cut on your radio.
  2. On every propeller, battery, weight change of your model lower MOT_THST_HOVER to 0,12 and MOT_SPIN_MIN,0.10 especially if quad is powerful.
  3. Set airmode to on, so you have control over quad when throttle stick at lowest position: " In addition, setting bit 0 of ACRO_OPTIONS will activate AIRMODE in those modes all the time."
  4. Configure one switch to be able to quickly change to stabilize mode and/or acro mode as they use minimal stabilization. Acro is only using rate pid stabilizer and stabilize is using rate and atitude pids. So if more complicated mode fails there is a chance that acro will allow to land safely.
  5. MOT_SPIN_MAX can probably limit maximum pwm value send to the engines.

Be ready to quickly press emergency engine stop if sth. goes wrong. And that’s all we can do about safe testing.

That is not true. Acro uses rate and attitude loops. You would need to set bit 1 of ACRO_OPTIONS for acro to only use the rate loops

ok, didn’t know that. Thanks for information.

1 Like

This is default anyway in 4.1

Just my 2 penneth.
Reflash firmware to flight controller (after formatting new sd card and restarting Mission Planner/qgroundcontrol).
Configure all sensors and recalibrate radio.
Calibrate esc’s.
Test without props.
Test fly outside with a short tether rope.
Gain your confidence back and yr good to go.
Good luck.

I don’t think this is ever a wise thing to do.

Thanks for advice. I will definitely go back to arducopter (this is the only software that can do all things), but this weekend is long weekend for me so I won’t mess with my quad just before :slight_smile:
I hope to have 4 days of flying with no worries, no work, and cellphone buried deep in the clothing bag :smiley:
We are going to our family to the countryside, so my quad must be ready :smiley: I can’t wait to get there, already can’t focus at work :stuck_out_tongue:

1 Like

The problem that i see is that you tuned a drone on 3s, then flew it on 4s. Throws off your pids. that can cause instability or even flyaways

You are right that rising power will be equivalent to increasing P and D factor, however in my case throttle was not computed by pid algorithm as i was in stabilize mode. Attitude stabilization was working fine after changing from 3s to 4s.