I'm relatively new to the hobby, been flying for less than a year. Over the winter I took on my first full build, the one that's now causing me trouble.

It's an ImpulseRC Apex frame running the Holybro Kakute F7 v1.5 + Tekko32 4-in-1 F3 40A ESC with Apex Hornet 2306.6 2555kv motors.

The maiden flight went well. It was cold so I only flew half a pack, around 3 minutes. I flip-flopped around, did some punch-outs and snap flips and rolls. All was well although propwash handling could be better.

Made some tune adjustments and upped the PID/gyro loop from 4k/4k to 8k/8k and dshot 300 to dshot 600. Went out and after 30s, just as I levelled off from a snap backflip, the quad flipped out. At first I thought it may have something to do with changing the loop times but after some research it has all the signs of a desync. I didnt have a card on the board for that flight, so no blackbox. This is the HD footage from the first event.

From looking at the footage, I suspected motor #3 wasn't spinning. All my screw lengths are OK and there wasn't any visible damage to the motor or ESC, but one of the joints did look a bit sketchy. I reflowed it with some new solder and put everything back together and went for another flight today. I didn't mount my HD cam as I didn't want to risk breaking it but I did put a card into my FC for blackbox.

Again, after 30s of flight (maybe just a coincidence) as I level from a snap backflip, quad goes nuts. Here is the DVR footage from today.

Here is the blackbox log: 
  LOG00004.BFL.zip (Size: 438.15 KB / Downloads: 118)

Although I'm unsure I've set everything up correctly there is some data in there, notably that the #3 motor is at 100% when the desync occurs. Maybe also worth noting that I intentionally left the props from the previous flight on. Not too badly damaged but not in perfect shape either.

Running latest BLHeli32 firmware and BLHeli Suite says everything looks good, here is my config:



When I spin up #3 without props in BF configurator everything seems fine... except that I may have broken #4 (front-left) in the latest crash, it now has a noticeable vibration and feels like I bent the bell when hand spinning it. Not the issue at hand but frustrating nonetheless.

Some additional notes that may be potential leads:

• When I first armed, I noticed the idle didn't seem smooth. The rpm was jumping slightly, airmode is on.
  
• I have the known known issue where the ESC init chimes are out of sync. The GitHub issue for this seems to indicate the problem is benign.
  
• A solder ball went into that motor when I was building. I could feel it grinding when spinning the motor by hand. Even after removing the bell screw I was never able to get the bell off, no matter how hard I would pull. After hand spinning a bit more I could no longer feel it grinding and figured I would be OK.
  
• I'm a real beginner when it comes to soldering and I spent waaaay too much time soldering the negative XT60 lead at high temps. No pads lifted but I wonder if it may have damaged something I can't see. The pads for #3 are the closest to the negative lipo pad.
  

I'm really unsure where to go from here, any insight will be greatly appreciated. Thanks!

Do motors feel hot at all?
A bent bell or even bad bearing can sometimes cause enough noise to present oscillations and a desync too.

I would go back to dshot300 and 4k/4k to see if the problem goes away. We can always help tune out some of the propwash with your PIDs as long as it’s flying for you

What version of BF are you running?

If #3 motor is constantly going to 100%,then that is your problem. Either ESC/signal or the motor. I would swap motor 3 with another motor. See if you get the same issue and see if the problem moves or stays. If it stays, ESC/signal. If it moves, then motor issue. If the problem goes away, then likely a bad solder (even though you said you reflowed).

(25-Mar-2020, 03:52 AM)Drone0fPrey Wrote: Do motors feel hot at all?
A bent bell or even bad bearing can sometimes cause enough noise to present oscillations and a desync too.

I would go back to dshot300 and 4k/4k to see if the problem goes away. We can always help tune out some of the propwash with your PIDs as long as it’s flying for you

What version of BF are you running?

Totally forgot to mention this: before going back out after the first desync I reverted to 4k/4k dshot300, so that wasn't the issue. BF is latest stable 4.1

Motors were cool after the maiden, but both subsequent flights were no longer than 30 seconds, no enough to get them hot regardless.

Motors (Apex Hornet 2306.6 2555kv) were all still shiny and new the first time it happened (it was only its the second flight). The dented bell on #4 was an unfortunate result of the second desync crash. I should have flown over snow instead, I did take the time to apply conformal coating...

In any case, I'm starting to think I damaged something on the 4-in-1 ESC while soldering. When I say I spent waaaay too much time at high heat, I'm talking like multiple minutes at 450° without realizing the lipo negative pad is connected to all other negative pads (hence the heat soak and why I had trouble tinning the pad, but also sent all that heat elsewhere). #3's pads are the nearest to that negative lipo pad.

I'm slowly starting to think I may just save myself the headaches and replace the 4-in-1 and the motor. I'm going to need to replace the dented motor anyway, I'll order two plus the ESC at the same time to save on shipping.

I still want to figure out why it's happening, if possible, as it's going to add up to a 100$ mistake -_-

Thanks for taking the time to respond!

(25-Mar-2020, 04:44 AM)voodoo614 Wrote: If #3 motor is constantly going to 100%,then that is your problem. Either ESC/signal or the motor. I would swap motor 3 with another motor. See if you get the same issue and see if the problem moves or stays. If it stays, ESC/signal. If it moves, then motor issue. If the problem goes away, then likely a bad solder (even though you said you reflowed).

Sound advice, thanks!

I applied conformal coating, so I'll fly over snow next time (for lack of grass). My concern now is on how safe the motor with the dented bell from the latest crash is to fly? It still spins, but the vibration is noticeable and I can feel it catch on one side when spinning it by hand.

As mentioned in the other reply, I really botched the soldering job on the negative lipo pad. I'm talking minutes at 450° and all the negative pads are connected so I sent that temp elsewhere too. The pads for #3 happen to be the closest ones. How likely is this to be an issue?

(25-Mar-2020, 05:09 AM)udy Wrote: Sound advice, thanks!

I applied conformal coating, so I'll fly over snow next time (for lack of grass). My concern now is on how safe the motor with the dented bell from the latest crash is to fly? It still spins, but the vibration is noticeable and I can feel it catch on one side when spinning it by hand.

As mentioned in the other reply, I really botched the soldering job on the negative lipo pad. I'm talking minutes at 450° and all the negative pads are connected so I sent that temp elsewhere too. The pads for #3 happen to be the closest ones. How likely is this to be an issue?

Dent bell my be an issue. Like DoP mentioned, it can desynced very easily. 450 degrees F? That is too low, kind of. If Celsius, then that is very high. Are you using 60/40 or 63/37 flux core solder? My iron temp is usual 600 deg F. Only on really large gauge wire, I will bump it up to 700 deg F. I know that 60/40 melts at 370F, but a lot of time, the tip cools too fast to keep the solder melted. I find 600F works well.

It is hard to say if you toasted the trace to #3. What I would do is use alligator clips and check the continuity of the ground from lipo to the #3 pad you suspect is the problem. Then wiggle the wire on the #3 pad to see if you can get it to loose continuity.

(25-Mar-2020, 05:43 AM)voodoo614 Wrote: Dent bell my be an issue. Like DoP mentioned, it can desynced very easily. 450 degrees F? That is too low, kind of. If Celsius, then that is very high. Are you using 60/40 or 63/37 flux core solder? My iron temp is usual 600 deg F. Only on really large gauge wire, I will bump it up to 700 deg F.  I know that 60/40 melts at 370F, but a lot of time, the tip cools too fast to keep the solder melted. I find 600F works well.

It is hard to say if you toasted the trace to #3. What I would do is use alligator clips and check the continuity of the ground from lipo to the #3 pad you suspect is the problem. Then wiggle the wire on the #3 pad to see if you can get it to loose continuity.

Yuup that was celcius. 63/37 solder. It was a beginner mistake, I've done some reading and youtubing since and know how to heat a large neg pad properly. Now I work around 375°C most of the time, max 425°C on large gauge. At that time I just kept going up until I was able to flow solder onto the pad.

Again, thanks for the good advice. I'll try that continuity check before anything else. My guess is the motor neg pad is the nearest to the standoff.

it looked like a „motor-idle speed issue“ to me. as you wrote „no smooth spinning on idle“, stay a bit more on throttle while hard flips and/or set a higher idle speed. i would try this first. if you still can with the particular esc.

(25-Mar-2020, 04:54 PM)hugnosed_bat Wrote: it looked like a „motor-idle speed issue“ to me. as you wrote „no smooth spinning on idle“, stay a bit more on throttle while hard flips and/or set a higher idle speed. i would try this first. if you still can with the particular esc.

I thought of this but the fact it happened on the same motor both times makes me wonder if it’s hardware related. At the same time, it happened doing the exact same thing, snap backflip, so software could be an issue. It’s a tough problem to diagnose -_-

I kinda wish there was a good way to test this without flying. I haven’t attempted any fixes yet so I could try to reproduce by securing the quad in a safe setting.

Edit: blackbox seems to indicate I was at least at 15-20% throttle when it happened.

im wrong, 20% should be faraway enough.

(25-Mar-2020, 07:42 PM)hugnosed_bat Wrote: im wrong, 20% should be faraway enough.

Correction, I was at 0 throttle at the time. Currently I have min throttle set to 6% (1060) which is the default.

I made another finding that I'm not sure what or if I should read into. I removed the props and plugged in, armed in turtle mode and used yaw to spin #3 (rear-left) which is where the problem is happening and it's opposite (front-right). After several seconds of giving it about 50% yaw I noticed #3 was noticeably hotter than the other. They sound the same as they spin up...

did you play with you filter settings on 4.1 before you changed from 4k, 4k to 8k, 8?.dshot 600 should not be a problem for those esc's.

(27-Mar-2020, 12:21 AM)Fisherking Wrote: did you play with you filter settings on 4.1 before you changed from 4k, 4k to 8k, 8?.dshot 600 should not be a problem for those esc's.

Yes, I had. However I had reverted back to 4k/4k dshot300 before the second desync. I’ve just finished swapping two motors to see if the problem persists and where. Hopefully I’ll be able to try it out tomorrow.

Update

So I have a feeling I know where this is going, but I figure the steps and findings are worth noting for others. Any feedback is appreciated!

TLDR; did a poor job soldering the negative lipo pad on the 4-in-1 ESC, too much heat for way too long and damaged something related to the #3 motor (nearest to neg lipo pad).

A simple process of elimination, as suggested, was to swap the the motor on the arm that dipped in the two previous desyncs with another and see if:

A) problem solved, bad joint somehwere 
B) problem followed the motor, bad motor 
C) problem persists as before, bad ESC

So I swapped the problem motor #3 (rear-left) with #2 (front-right) and went back out, repeating the same snap flips that had triggered the problem the last two times. I finally had the foresight to fly over a soft crash zone (3ft deep snow, conformal coating had been applied) but didn't want to risk my HD cam so only DVR footage. Also had a card for blackbox. Take note that I flew slightly beat up props and the shaft on #4 was slightly bent in the previous crash and causes vibrations. I went back to 8k/8k dshot600 for this flight but had more filtering than the last two, just in case.

The results:
DVR footage of the event
Blackbox log: 
  LOG00007.BFL.zip (Size: 1.06 MB / Downloads: 96)

Same outcome, problem stayed on #3 and occurred on a snap backflip but this time it seems to start halfway through the flip instead of as I level out.

My conclusion is, as mentioned above, that I damaged the something while soldering the negative lipo pad. So my solution is to replace the ESC. I'm also going to need a replacement motor since in the second crash I seem to have slightly tilted the shaft/bell alignment (bell doesn't sit exactly perpendicular to the shaft) on one motor. Unless anyone has any tips to realign that sort of thing...

My only question is if you guys think the BF min_throttle setting is a probable cause here? It's set to the default 6% (1060) and this is all new high-end gear, if that makes a difference: Holybro Tekko32 F3 ESC + Apex Hornet 2306.6 2555kv motors on ImpulseRC Apex frame.

Thanks!

Min_throttle does not get use if you use Dshot.