Quadcopter barely makes it home!

Hello I’m running AC3.4 on a 450 sized quad that has been designed for survey usage.
I should also mention I do not use an RC and everything is done through an android device.

Quad is auto-tuned, precisely balanced and has performed hundreds of missions.
Engine RPM has been checked and all perform the same.

However yesterday I decided to have a test flight on a quite windy day (wind speed ~35km/h),
loiter and hovering was OK, but then I decide climb to 50m and go to a waypoint no further than 150m. I then command RTH and notice the quad really was struggling to come back, often rocking back and forth.

It barely managed to come back, and the dataflash log reveals very peculiar behavior (Engine 4 often maxed out, loss -gain of altitude, loss- gain of speed).

I have been using Pixhawk and AC fw for more than a few years and have never encountered anything similar.

Pls check the included log file and let me know if you can figure something out.
Is it maybe a setting I can change?


Thanks in advance

Got any bullet connectors on there ?

Looks like motor 4 lost sync,but tried hard to get it back,then got it back and recovered.All the othere motors go hard at it when one shuts down hence all the activity and lack of forward progression.Quads don’t fly too steadily with three wheels on their wagon.

I’ve had a few inexplicable motor failures with bullets (as it turned out) to not run them.Any I have fitted are hard soldered together.If you don’t have any then check all the solder points.I just had a dry solder joint catch me out.It looked perfect.

Could be an ESC on the way out,so magnifying glasses and check out the chips for signs of burning,even if it means unwrapping them.You may try to de-sync it manually with only that one motor hooked up.

Thank you for taking the time to answer,

I do use bullet connectors yes, However I tested the esc for desync by rapidly requesting 100% throttle a number of times and saw no abnormalities. Also I think that if a motor would have lost sync that would look more dramatic, like a flip or total loss of control? Also flying hard on manual mode didn’t show any issues.
Oh and what particular aspect of the log makes you suspect a desync?

It just went to max and stayed there.Signs of a failed motor or desync.A flat line.Luckily it recovered.

It was doing some wind fighting and wobbling about a bit which demands fast responses.Three motors did and one got stuck for a moment.It could easily be a bullet connector having a moment.The solid ones are far better than the sprung cages for dealing with vibration

I’ve had desyncs with just a crack in a ground wire solder to an ESC.It only broke completely after being wiggled.The giveaway there was a sharp rap on the shell would reinstate it.Dead cert bad connection.In this case I think it was the hard work it was doing that provoked the fault and I’d suspect the bullets every time.And solder them solid just in case.

Bear in mind you had just done a 50m climb with a flattening battery.It hit battery failsafe during the wobbles just after peaking out.Less available thrust at the bottom of a battery,by quite a bit.

I see, went through all my connections again, regular flying seems ok, I guess I will now have to wait for a windy day in order to replicate the scenario. One more thing though, couldn’t it have been just the strong wind pushing from one direction that could have required the engine to give full power, but still that wasn’t enough thus the behavior seen?
Thanks again

That’s possible.I didn’t check your ANGLE_MAX setting but if it’s up at 45 degrees it must have been pretty windy up there. :grinning: