Stumped - Death Roll

[ElectriSean]

So I am stumped. My FT210 build went pretty smoothly until the maiden... It lifts off and hovers fine, yaws and rolls and pitches fine, until you get on the throttle. Over about 50% it seems like motor 4 dies and it drops. Every time, same direction. I've tried everything I can think of in Betaflight and BLHeli configurator, no change. Super soft mounted the FC, replaced the ESC and still no change.

Video of the behavior: https://youtu.be/VE33nsLEBXY


FC: Lumenier Lux v1
ESC's Sunrise Cicada 30A BLHeli_S 16.6
Betaflight 3.1.6

I'm getting a little frustrated, I've broken more props on this thing than I have on my Wizard, which has many many flights and crashes on it Any suggestions are welcome, thanks

 

[Snarls]

Check your wiring for loose connections. I had a similar issue with my KISS setup where it would death roll on hard punches. People were experiencing the same issue due to a firmware bug in the KISS FC, but for me it turned out that the power connection to the FC was just loose enough that it would disconnect at high accelerations.

Maybe try a bench test where the quad is held still, but you throttle up the motors. That way you can confirm its not because you are pulling too many amps causing ESC shut down or something.

 

[ElectriSean]

I went over my soldering with a microscope, after using 99% IPA to clean all the flux off the board so I'm pretty sure I'm good there. The bench test is a good idea, though the prospect of it terrifies me.. lol The OSD has current sensing, I'm curious what that will say. I haven't even put on the goggles with it armed yet, just for a quick test that the FPV worked at all.

Thanks Snarls, good to have you back

 

[PsyBorg]

Well flying that tiny quad with chainsaw motors probably puts more vibration into the quad then super soft mounting can handle I think.

Seriously though..
I recently read about the death flip somewhere when I was researching things. I think it was in the faq for beta flight or KISS when I was looking into setting esc's on my Gremlin. If I remember correctly it had something to do with min throttle setting and desync from being too low. It looks like it happens when you let off the throttle which kinda makes sense if min throttle is too low.

 

[ElectriSean]

Well flying that tiny quad with chainsaw motors probably puts more vibration into the quad then super soft mounting can handle I think.

LOL they aren't that crazy... I don't think at least, I guess we'll see. TBH the wizard is plenty fast for me at this point, which is why the 210 is taking so long to get figured out What sold me on the RG motors is the durability.

Seriously though..
I recently read about the death flip somewhere when I was researching things. I think it was in the faq for beta flight or KISS when I was looking into setting esc's on my Gremlin. If I remember correctly it had something to do with min throttle setting and desync from being too low. It looks like it happens when you let off the throttle which kinda makes sense if min throttle is too low.

I remember reading about it in the BLHeli_S thread a long time ago, but that thread is massive... And the search over there just sucks. My min_throttle is 1070, they all start around 1025 or so, it's pretty low. I'll try bumping it up to 1100 though, something to try to avoid bench testing

 

[PsyBorg]

Remember if you go over 1100 thats the minimum arm check..It won't arm unless you raise that as well. But 1070 is kinda high already and would make it "Floaty" on free falls I would think.


Also be aware that when you cross that 1100 mark you are close to where that thing will pop into the air as soon as you arm it.

 

[ElectriSean]

Remember if you go over 1100 thats the minimum arm check..It won't arm unless you raise that as well. But 1070 is kinda high already and would make it "Floaty" on free falls I would think.

Also be aware that when you cross that 1100 mark you are close to where that thing will pop into the air as soon as you arm it.

Actually my min_check is set to like 1020. This is just the throttle value that the FC must see to arm, it has no relation to min_throttle. At zero throttle stick the FC see's exactly 1000. I thought that min_throttle at 1070 was pretty high, I think I raised if from my initial setting of 1030 to test against this death roll. One of the benefits of BLHeli_S is supposed to be really low min_throttle.

Looks like I'll be bench testing... Maybe I'll put it on the ground and pin it with my steel toe boot wearing foot... Wish I knew a hokey goalie to borrow his pads lol.

 

[Snarls]

Looks like I'll be bench testing... Maybe I'll put it on the ground and pin it with my steel toe boot wearing foot... Wish I knew a hokey goalie to borrow his pads lol.

Ha be safe! What if you switch the props around so that all the trust is downward. You'll still want to secure it be at least it wont want to fly away.

 

[TazRC]

ElectrSean, I'm curious to know what modes you have programmed in BF and what mode(s) you were in when it flipped?

 

[ElectriSean]

Ha be safe! What if you switch the props around so that all the trust is downward. You'll still want to secure it be at least it wont want to fly away.

I remember seeing a video about this awhile ago, I think it was Joshua Bardwells... Going to look into this, I value my shins

 

[ElectriSean]

ElectrSean, I'm curious to know what modes you have programmed in BF and what mode(s) you were in when it flipped?

No modes set up - acro only. Accelerometer disabled, airmode permanently enabled.

 

[Snarls]

I remember seeing a video about this awhile ago, I think it was Joshua Bardwells... Going to look into this, I value my shins

I remember that for APM/Pixhawk builds if you want to test the electrical interference on the FC from the motors/escs you have to run a compassmot command with your motors running under load on the ground. So you have to switch the props otherwise you have a very scary beast to hold down (see 1:30 in this vid).

 

[ElectriSean]

Thanks for that, looks simple enough.

 

[TazRC]

This issue is probably deeper than my current BF knowledge but perhaps try Angle without Air being programmed, to eliminate any problems with Air mode. I had similar issues with Air mode, which Boris is supposed to be correting in 3.2 according to GitHub. My stepson had similar issues too and cured it by swapping out the F3 FC with an F4. I also read these type of issues don't exist with CleanFlight.

May also try swapping motors and/or ESCs to see if the issue changes corners.

FWIW

 

[ElectriSean]

May also try swapping motors and/or ESCs to see if the issue changes corners.

I replaced the ESC with a new one before, to no good effect... Guess I have a spare. I put the props on upside down as Snarls suggested, but couldn't make the problem present itself enough to see using the Tx, so I went through the motors tab and slowly ramped it up. Sure enough, around 1300 it started stuttering and around 1600 stopped completely. Controlling through the motors tab like that eliminates the PID loop completely, and since the other motors work fine with the same BLHeli_S settings, I'm pretty sure I'm looking at a bad motor. I have a racerstar 2205 I will swap in to test out.

 

[PsyBorg]

Maybe its something in code... I have tried several times now to get my Gremlin to fly using Dshot. Only 3 motors will spin. When using multishot or lower in the *shot category everything works.

 

[ElectriSean]

Maybe its something in code... I have tried several times now to get my Gremlin to fly using Dshot. Only 3 motors will spin. When using multishot or lower in the *shot category everything works.

I've actually been using Oneshot 125, haven't bothered with DShot at all yet. I've heard too many horror stories, and I'm pretty sure any benefit would be lost on my dumb thumbs anyway

 

[PsyBorg]

I love Dshot on my Alien. No calibration, lower motor start up speeds way more smooth in slower rolls and tighter control with yaw as well.

It all depends on how good you are at tuning whether you think Dshot is worth using. If you can tune to like 95% of perfect then Dshot won't seem to change anything. If you cant tune so well (like me) you notice a big change on a current tune you are used to. I think it would help smooth out the Gremlins flight specially since I use these heavy monster batteries atm. Getting it to up side down is work and then the inertia kicks in and I over correct most of the time coning out of maneuvers.

 

[ElectriSean]

I actually have done very little tuning on any of my machines. Most of my flight time last year was spent learning to fly/struggling for control. After a winter on Velocidrone I've come a long way, and the wizard actually flies fairly well so i haven't really delved into it. I still consider myself a terrible pilot though, so I'm spending most of my energy trying to suck less At least most of my landings are intentional now lol.

 

[ElectriSean]

I am happy to report a successful maiden Changed out the motor and did the bench test and it all worked fine. Flipped the props upside up and flew a couple of packs with the mismatched motor at the ball diamond across the street. Quite a bit of video noise, though that is probably from the chain link fences all over the place. I'll be going to the club field tomorrow which has pretty much zero interference, hopefully it cleans itself up.

Thanks everyone for your suggestions