Hi all!

I have this quad that I have built a few weeks ago, that has a problem I haven't encountered before. And instead of spending more of my time making random adjustments of my PIDs, I figured I would ask you smart people here for tips.

This is the fourth quad I have built, and I may have been lucky so far, but the other three flies somewhat decently on pretty much any tune. This one, however, is almost scary to fly. It flies OK as long as I don't do any pitch flips, but if I do, it sometimes goes into a crazy, uncontrollable death-flip. Sometimes, not every time. It will continue flipping for a second or two, until it either recovers, crashes into the ground, or I chicken out and disarm.

My rates are 0,80, super-rate 0,84 and expo 0,30, giving 1000 °/second. It's the rate I use on all my quads. Is this simply too much for this quad to handle? I see from the betaflight logs that it reaches rotations of well beyond 2000 °/second at times during the craziness.

I've tried moving the sliders around, watched tutorials and tried turning off the FeedForward, the Dmin, and the Anti Gravity. But I feel like I'm just changing things at random, and have no idea what is actually causing this.

Could somebody please give me some tips as to what could be causing this? I've tried to include a video of a couple of the last flights where I did flips, so you can see my problem. It doesn't happen on every flip, as you can see from this.

The setup is:
Armattan Badger 5' DJI Edition
iFlight SucceX-D F7 TwinG + 4in1 50A ESC Stack
iFlight XING X2208 2450KV NextGen Race Motor
DJI Air unit and camera
FrSky X-Rsr receiver
GemFan WinDancer Durable 3 Blade 5042



Any help is appreciated!

a_pihl

Can you post logs?

(01-Mar-2020, 01:16 AM)voodoo614 Wrote: Can you post logs?

Thank you for your reply. I'm new here, not quite sure how to add files, but can you access this?

https://drive.google.com/file/d/1Jt1bfSs...sp=sharing

If you look at log 7/9 at 0:20, this is the flip where the prop came loose. Do you need more of the older blackbox logs, let me know. 

And at the 0:16 mark of log 4/9 I believe you see the second to last flip in the video, where I got lucky and managed to recover.

a_pihl

Haven't forgotten about you. I haven't had time to look at the log yet.

Not a problem! The weather here on the west coast of Norway isn't all that drone-friendly at the moment anyway...

I'm happy for any advice you can give me! I figured people here had more experience than me with stuff like this. All my previous drones have been very easy to set up, and behaves nicely on almost any tune. At least with the stock Betaflight tune. So in a way, it is maybe a good thing to get a more problematic drone, so that I'll have to learn more about tuning...

Hello again!

Had a few minutes of decent weather this morning, so I thought I would try again. I managed to do a left roll, a right roll, and tried to flip aft, but immediately crazyflipped into the ground. This happened again, but on the second crash, the battery fell out and the DJI air unit did not like that, and the video file got corrupted. So I didn't think I would have any movie to show. But then I remembered the googles also have a built in recorder... Technology is great! 



I don't know much about Blackbox and tuning, but I noticed that during the extremes of the rotation, all motors seem to stopp for a fraction of a second. As if the FC just gives up! And this happens both times, and I noticed it again on earlier flights. On one of the earlier flights, for as much as half a second. What is happening here?



I only managed two flights, then I had broken to many props, and realized I only brought spare CW propellers, and of course needed a CCW one... Apparently, I do most of my crashes with the forward right arm...   I included the blackbox logs if anyone could have a look and give me a tip on what is wrong...

https://drive.google.com/open?id=1HDZmq2...-SJrKGiION

Have a nice day!
a_pihl

Why have you upped the P in Pitch axis? It does look like the P term is overshooting really badly. Try getting it down and see how it behaves.
Other than that I can only advise to up your filters (Sliders down) to see if it helps.



See how the P term (yellow line) behaves on pith vs roll?

Thank you very much for your input!

I was just moving the P up and down at random, to see if there were any improvement. And as you can see in the end of the first video, I had the P term as low as 24, and I started to think it was getting TOO low. I believe the defaults were 46.

But I will try again, with lower P on pitch axis next time. And with more filters. And see how it goes. 

Thank you again for taking the time to take a look at my log! I really appreciate that! 

a_pihl

I looked at the first link log. unfortunately, there is only one log and not multiple logs as you alluded to. There was not much to see.

The second link just look like you hit the ground on a flip, but the quad did not spin out of control.

Hi again!

Really? Maybe I've messed something up, there should be one log about 6 MB in size, that contains 9 flights, and one log that is about 1 MB in size and contains two. Maybe try this link?

https://drive.google.com/open?id=1Sj1vXI...kubcUwXbz1

Like I said, the first log, should have 9 flights. The first one is me just hovering, but things are more interesting in the other ones


and the second one has 2 flights, the first one i disarmed in the air, but in the second one I hit the ground.



I have gotten the advice from brnmd to reduce P and increase the filters, which I have. I will try this at the next opportunity.

Thank you very much for your time, and I apologize if I messed up the links!

a_pihl

I see, all your flights are on one log. I have to go a back and check. I thought you had different logs for each flight.

So I looked through the logs again. Check motor 1. I would reflow all the connection to motor 1 and check any wire plugs for motor 1. It seems to be the motor that desynced the most often.

Hi again!

Thank you for your suggestions. I flew again today, with a lower P and more filtering, and all was OK. No crazy flips. I did notice a pretty big post-flip-bounce, but nothing that would send me crashing to the ground. I tried lowering the P further, but that only made the bounce worse. And increasing it, made the quad more stressed, so I felt it was about to do the crazy flip again. I will try again later, but for now I'm happy. At least I have a much better starting point to do further tuning from.

You can see for yourself on this video



I will look into the desync of Motor 1 as well. By the way, how do you see the desync in the logs? What do they look like?

Thank you both for your input. It is really appreciated!

a_pihl

Normally you look for one motor to goes to 100% just right before your crash. The FC is trying to tell the motor to push harder, but since the motor desynced, it is doing nothing. The FC then max out at 100%.

Send in a new log after that so we can help a bit much with your tuning