Loosing control in loiter mode

Hello,

Had crash twice when operated in loiter mode recently and I am not able to understand the actual reason for it. From the log I am seeing the current position shown is not as expected from what copter was at. Could you please help me to figure out the actual problem here ?

attaching the log files for reference
2020-02-01 22-24-51.bin (878.1 KB) this is with two gps in blend mode both ublox neo 8m

2020-01-22 18-21-12.bin (416 KB) This had only 1 gps

While I try to use in stabilize mode this problem is not seen

there where crashes before also in loiter mode and at times loiter mode was not getting engaged itself because of gps hdop criteria not met.

Thanks

Hi @xloving,

I had a quick look at the 2020-02-01 log (the other one is Copter-3.6.11 so I’ll ignore that one). It looks like maybe the vehicle didn’t actually takeoff? It seems to be bouncing around at a very low throttle level.

The number of satellites is quite low at only 11 but the hdop is not too bad at 0.8 to 0.9.

Maybe you’re just being a little too cautious taking off?

I guess the problem looked similar to me in both cases where i was losing control.

Yes I was very cautious to see if anything is wrong at very low altitude, but the moment i increased the throttle it went off control as you can see from the barometer reading

Wondering what does the circled portion mean ? the gps module was not sending any data during that time ? May be because of landing ?
I see a ramp down instead of a step wave. How better to interpret this ?

By the way this is on the stabilize flight for which I said things were fine :frowning:

@xloving,

Ah, so according to the baro the vehicle climbed to 15m and I guess this is correct?

If “yes” then it climbed with a very low throttle output. Have you calibrated the ESCs and done the motor test?

I believe yes since the other flight which was in stabilize had everything same except a different battery pack and the flight mode. One thing that is surprising is the position estimate shown in the map and the actual position shown by GPS. calculated actual position was wandering around as seen in the map view for

Even in case of 2020-01-22 18-21-12.bin it happened something similar even though its a different firmware version

How big is this copter?
It is grossly overpowered which is the start of your issues.
Have you been through the Tuning page in the Wiki?
Your desired and actual look OK until it hits height, and maybe they separated due to its inability to reduce power enough for controlled decent.


Once you have been through the tuning Wiki page I would suggest adding weight (maybe a bigger battery or a second battery) and try again.

I did notice you were caught out by having the throttle down when armed in Loiter, and not realising it has a delay (it thinks its below ground) when raised above neutral because of this.
I strongly recommend getting it flying well in Stabilise, followed by AltHold and then Loiter.
That way you check all the basics first before expecting a miracle in Loiter.

Thanks for the info, but I am not fully convinced here.
Its a typical 450 frame with 1400kv a2212 motor with 30a esc and 1038 prop. I had successful flights with 1045 prop in loiter without any problem (the gps module had some problem recently - magnetometer was not giving orientation- because of the crashes from my bad piloting :slight_smile: and changed with the new one). This flight had a higher capacity battery with 1038 prop. So there is something more to it I guess than overpowering. Since I had some problem in all recent Loiter irrespective of firmware version and basic configuration, a quick look again if possible will be of great help.

should have also had the same problem if it was because of overpowering since setup is same apart from the mode, battery and some rc sensitivity(again, I guess).

Should I suspect my GPS module since I see some gps related error messages in log

Yeah, I had not gone to the extend of tuning since I was happy with the performance before

BTW which is this wave plotting tool ? I am having difficulty when the values are not having comparable magnitude in mission planner for a combined analysis

Any further suggestions here will be of great help since I am reconstructing the quad after the last crash. Earlier pitfalls if any, are to be eliminated this time for a better confidence.

Thanks,

Other than a quickly sagging battery there isn’t much to be learned from this log. It didn’t fly long enough to capture a MOT_THST_HOVER value which is still at default. In fact much of what I see is at default. Follow the guidelines in the tuning Wiki as suggested before flying again.