#Random Flip after a second of flying
445 messages · Page 1 of 1 (latest)
That's super odd
does your FC have a blackbox?
if so, can you provide a log from it?
also, some specs pls
I think so, let me check. I have a speedybee v3 f4505 stack
Okay one issue, I have a blackbox but I didn't realize that it was missing the microssd
Also eco 2 series 2306 2400kv motors
receiver?
did you get anything in the OSD when it happened?
No not from what I was seeing
ill do another test to get some of that info though, I found an microSSD
I cant seem to get the blackbox to record onto the ssd
Yeah I cant figure out why the blackbox is not recording anything
I have tried setting it to a switch and the only file it ever adds is FREESPAC.E
Any suggestions for settings I could try to see what the problem is?
for the flipping issue
Right now I can't even begin to guess what the issue is
make sure that nothing is hitting the FC, especially the gyro on the FC @amber vector
I'll give that a try, I think one of my wires might be against the board and I didn't realize the gyro could be messed with like that
the gyro is suuuper sensitive
intentionally, but also to its own detriment
that's why FC's tend to be mounted with some sort of flexible isolation
any sharp movements (FC moving around on screw threads, knocking about, etc) will massively screw with the gyro
can't tell.
after adding in a spare nut and securing the FC it now does unwanted backflips
I would send a video but it is dark here
Yeah, if anyone else see this post, I am still i need of help, the drone will not fly and I have tried everything I can possibly think of
@amber vector still need the blackbox log...
Do you know why a blackbox may be not recording?
my blackbox downloads a file to the SSD but the file is just named this:
can you show a screenshot of the BF configurator's blackbox tab?
ideally the whole BF configurator window pls
that part is especially weird because the SSD has zero files in it other than the freespace thing
it looks like it should be recording, odd.
is there a max ssd size maybe?
just FYI, SD xD
ah you right, you right
as far as I know, 32GB is the hard limit
Here is a picture of the exact sd
but it was supposed to be fixed in 4.3.2...
What issue?
with SD cards
oh I thought you were saying that the fix didnt work or something
yeah im not sure whats up
so, first
the FREESPAC.E file is what BF creates
it basically provisions some space like that
when recording, it shrinks this file in order to free up space for logs
got it
did you already fly it with the SD card in?
yes
and was it set in Blackbox tab to "SD Card"?
yes
then I am super confused
me too lol
would you happen to have a DVR recording?
make sure that you have warnings enabled in OSD, and some OSD elements that update frequently too
no, but I could get one
any specific warnings?
wait nevermind
just make sure that you have the warnings element in OSD
and that all of the warnings are enabled
got it
ill do a flight with the dvr and send it
its a really terrible dvr but its all I got in terms of that
I did two test flights but it didnt start recording until about half way through on the first one, but the second one was there all the way
i suggest zero audio for the best viewing expierence
If you like I could convert it to a different format or something
@amber vector rough guess - it's either inputs on the radio/receiver being screwy or the gyro being faulty.
I wanted the BB log to tell you which of the two
Blackbox: fault is so odd tho
Damn that sounds like I'm gonna have to order new stuff
Is there a way I can find out which one it is without the blackbox?
Also I reflashed the FC and I'm gonna try flying again tomorrow to see if anything changes
check how it behaves in Setup or Sensors tab
the 3D view / gyro graph
It seems to be moving fine during that
Like all the controls correspond to the right stuff
Use sharp movements of the drone to test if it behaves reliably
also, are you flying angle or acro?
I was wondering about angle because sometimes gyros fail in a way that only shows in angle
sometimes banging the drone on the desk can show issues
Do you want me to set it in angle then bang it on my desk?
Also where would I monitor what that would be messing up
I'm not telling you to arm it lol
go into either Setup tab or Sensors tab
in both you can visually see some of the quad's behaviour
in Sensors you want to be watching the gyro plot during normal movements (maybe shake it a bit)
in Setup tab you want to be watching what the drone does when it's shaked around or hit, if the 3D model does some wacky movement
I gotta question
So I just noticed that 2 of the drone arms wiggle ever so slightly, would the vibrations from that cause the flips?
also got some news on the gyro
This first video shows it WITHOUT my controller being connected
Second video shows it with my radiomaster zorro connrcted
idk what to make of it but it seems like a major problem
@amber vector go to CLI tab, send this command: diff all
press Save to file and post it here
also, send this command: status
and paste the result here
With or without the controller connected?
doesn't matter
I want to know the config and some other specs, those don't change
does it unf*k itself when the radio is turned off afterwards?
yes but also now the gyro is randomly working with the controller plugged in
You seem to have blackbox set on a switch?
correct, because originall I thought the auto blackbox startup was not working
ah ok
all of that looks okay.. dammit
I was wondering if you had one of the broken ICM gyro chips, but you have a BMI
Damnn
which is actually one of the more reliable ones lol
I have no clue why the random flips lol
You think the arms moving ever so slightly would be the issue?
Cause I noticed a screw was a little loose and they wiggle ever so slightly
it's not good, but doubt that that's the problem
I'll have to refer you to Betaflight discord I think
idk how to fix your blackbox
and without it, we won't move on
Right, I can probably find that discord
But tomorrow I'll do the test flight with the recently flashed FC, idk maybe that will work a miracle
Yeah idk, I have flown a few quadcopters and drones and gotta say, this is the strangest one yet
<@&797250070567059487> sorry for the tag, but I need other ideas here
- SD blackbox shows the SD card working when on USB, but in OSD when flying it shows "FAULT", WTF?
- The whole "drone suddenly rolls/pitches 90° and crashes" thing
- See clips of BF configurator shown just above with Owen's description, WTAF.
I'm attaching a converted DVR file so you don't have to download it
For an LOS video see the message I am replying to
That motor twitch sound tells me bad gyro
That's one of my guesses, but I wanted to confirm that with BB...
it's a BMI270, those rarely fail tho
That's just me though because I had a drone make that sound but it yaw death rolled
have you touched PIDs or filters?
@amber vector
the diff all is here @scarlet gust
4.3.2, which should've fixed the SD issue, and it's pretty much stock config
what is confusing to me is that it flies so nicely for a couple seconds before going apeshit
blackbox would help sure
what is the cpu usage like in that video ?
it's a 3.2k gyro with DSHOT300... what are the odds of it running out of CPU?
low, but never 0 :))
the odds are nearly 0 xD
@amber vector when you get around to it, can you send some nice pictures of the build wiring?
oh what could be another problem - wire with some current and or vibrations near the gyro
do you have physical access to the drone?
no, see this msg ^ @scarlet gust
I think we should dig deeper on the fact that the bf preview freaks out when the receiver is connected
yeah that is just.. new
I've seen BF preview freak out, but not conditionally like this
is it?
@scarlet gust
okay what to do:
desolder EVERYTHING, just do ESC
check everything etc.
solder on RX and connect
and lets fix it at that point
no VTX or anything else messing with stuff
this is a 100% indication that the gyro cant update correct
or bf isnt picking it up
but why only when the radio is connected?
because the RX is sending noise somewhere

check RX connection, what board is in there?
betafpv ELRS lite 2.4ghz
inb4 receiver powered from 3.3V pad causing the gyro to crap out when sending telem

why???!?!?
you don't know what inb4 means do you
okay hear me out: 3v3 rail is overloaded when RX connects
I'm just predicting
in before - but i cant translate that to something in german i understand
it means that I'm writing a prediction, a guess of what's going to happen
oh ok
but IF thats indeed the case - and the gyro isnt getting enough juice anymore
that would explain it
it just made me think that maybe the flight time before glitch is somehow tied with the receiver
which made me remember the telemetry ratio on ELRS
the receiver draws the most current when sending telemetry..
that just turns off parallel PWM receiver functionality
the - at the beginning means off
first check the wiring
ofc
remove everything and start from the beginning 😄
and when the error accurs, remove the last step
That video seems like esc/motor issue to me
Reason I say that is because motor is dropping on one side
it doesn't look like it's dropping to me
it moves on an axis, not in a motor direction (diagonal)
yes, it looks to me as well as if the quad is actually "correcting" a gyro signal
so radio -> fly straight. Gyro lags super hard behind and FC got reported from the gyro, that quad is sideways
if you see the lag in the RX connected video
lol I was rebuilding a friend's PC like that once
it always crapped out at a random step and required several steps back to start working again
eventually found out that the CPU was damaged (flaky PSU) and would crap out when the hardware reserved RAM or something like that got too high, and that would save in the "CMOS" for a while
damn - thank god we dont have to deal with complicated things like a CPU or DDR ram here xD
yeah that one had me banging my head against a wall for a good week
eventually got my hands on another CPU and that fixed it
that was back in the time when CPU failures were completely unheard of, so warrantying it was yet another story lol
Do you have dvr audio?
yeah xD CPU dead? NO WAY cant be
lol mistag
lol yes
dont write so fast 😄 my slow hands cant react in time
btt: start from scratch, one component at a time
I removed the audio from my re-encoding to preserve your hearing 😄
you can fly LOS without VTX
to eliminate that
and show every solder step
show FC, ESC from both sides
Run tasks in cli
It'll show you if it's running out of cpu or not
It smells like cpu overload to me
No
a F4 on DShot300 should not have a cpu bottleneck
at what do you have the loop rate?
This is where most of the main configuration of your flight controller will be done. This includes things
what does that show you?
Im sorry ill have to answer this later, am in school rn
But I'll answer all questions when I am back, I am eager to fix this thing
Also an update I realized on the second video I sent
The reason the model and gyro seemed to be acting weird for that was because I was dumb and connecting the drone mid air, which was consistently causing that issue
kek
But when I connected it on the ground, the way its supposed to be as your guys know, it worked fine
FYI, the old name for gyro_calib_limit was moron_threshold
That's an accurate name
Is it f411? Are there additional filters running because it's a 3.2k gyro?
isnt the f411 to f405 differente not performance but IO capabilities?
It is the speedybee v3 f405
Tho I'm not sure what the difference is honestly
no, f405 has a higher cpu clock, don't remember the exact numbers tho
doesn't 3.2k gyro mean the filtering is done by the gyro, hence the lower max update rate
That sounds right, something like that 
https://github.com/betaflight/betaflight/pull/11400 if you wanna read up on it drama
2 problems I see, the ESC looks to be touching the frame. that can cause other issues, like shorting, including the esc signaling, but also just letting the smoke out.. The 3 wires touching the top of the flight controller are going to give the gyro a hard time. That doesn't explain the 90 degree sharp bank, but it's going to mask it in a gyro log, and could knock it out of the sky all by it's self. https://discordapp.com/channels/536623076796923924/1062857895361921024/1062877072449605702
the wire is fixed. We have a idea: BMI gyro with a HUGE noise floor
perfect, thats correct for the BMI gyro
so we are back to where we dont know whats happening
please read above and give us the info we need
xD yeah that's the story of this quad
everything points back to WTAF
Is it possibly a loose prop?
watch the first video, a loose props has some pretty good audio queues
we cant hear that
loose prop would cause a diagonal movement
That's true, though I have had some loose props act in a similar way before
Still, worth a shout fixing all the mechanical problems right now
Like the stack not even being squooshed
And missing a bolt 
I mean sure, just give the props a couple uga dugas
it's just a nut, and I mentioned it before I think..
yeah squish would be useful
no no, a very fine 4.31 Chugga buggas
where do you see the ESC touching the frame? me>blind
Yeah I saw just mentioning again
It doesn't look like it's touching the frame to be hondst
It's just really down low, which isn't necessarily bad, but can cause problems if the bottom plate cracks
if the bottom plate cracks
then you have worse problems than the ESC being too close >.>
If I were him this what I would do :
-
Take stack out so you're left with screws, and add a nut in-between esc and bottom plate, should give better resonance, and lift the stack slightly
-
Desolder the motors and inspect the traces - likely nothing there but if for some reason somehow something shorted a motor pad that might be your issue, and clean the flux
-
Resolder motor pads and re assemble stack, this time adding some squish onto the stack and ESC
-
Try and replicate the issue with black box logging
I get the feeling that it may be a combination of the stack being so loose that it may have shifted enough from vibrations to touch the bottom plate - however unlikely it may be it's still possible with how low it is and untightened
if hard stuff is hitting hard stuff then the drone will be hardly flyable
which is what doesn't make sense
those 4 steps there aren't a fix necessarily, but more of a proper build setup
I'm still wondering a bit about a possible input corruption or something like that
but for that we need to see a bb log 😦
If we can eliminate all mechanical factors first, then we know it's software and potentially radio
Man this is a fun one, there's a lot of things it could be but there's no direct clues
This how high the board is from the frame
And currently I'm having the betaflight people see if they can fix the blackbox issue, so I will keep you guys updated with the info I have
I mean the betaflight discord btw, I have not sent it into a repair place but I'm really considering trying to find someone to look at this
I don't know what country you are in, but the cost of the bare flight controller is $41, and most repair places would charge that much just to look at it. Do you have another quad you can swap flight controllers with?
I am located in the US, but unfortunately this is my only FC
unless if a tiny hawk 2 freestyle counts
As Lorin stated, not much point
Also did you do what I suggested? With the nut under the esc
we have bigger issues, this wont fix the underlying problem. It can help, if you are at a tuning stage where you cant eliminate some resonances, but now - this wont do anything
I know, but this will eliminate any possibility of it being mechanical
I'm leaning to it being it being mechanical with how suddenly it flipped that way
have you watched the videos?
if the mechanical problem you want to eliminate, it would twitch and freak out right away
not fly 3 seconds perfect and than just turn right
Unless it's at higher frequencies , which is actually still very low
But we should try to eliminate all variables possible first
It's most likely software but on the miniscule chance it's mechanical, it's still a good idea
if you would have read the BMI gyro link, you would have seen a much more possible cause of this problem
right that part kind of went over my head, I'll take a look at that now before anything else
but to confirm this, we need black box - and blackbox is broken
I actually may know how to fix his configurator
@amber vector when you go to the presets tab and try to apply a preset, does it just do nothing and kind of freeze?
If so, clean re install, and maybe get the new configurator
Clear your SD card and make sure it's the correct format, and to be sure just leave black box always on
Do a quick arm and hover, see if it still does it, and if it does plug it in and check to see if you can enter in Mass storage mode
If you can't do that, I'm more leaning to FC having an issue somewhere on the actual board rather than it being some obscure setting on bf
Okay I will try this tomorrow as its super late and I cannot fly
but i will let you know asap tomorrow
I got a question for the battery experts
I just did another test flight and the whole random-movenment that usually results in a flip, still happened
but not to the same degree as earlier
like I was actually able to recover it, still very uncontrollable, but I was able to keep it in the air for a bit
The question is would using only the external battery slot on the radiomaster zorro cause weird transmitter related issues due to not receiving enough power or something?
The battery im using seems to have enough voltage with it being a 7.4v lipo battery while the zorro normally uses 2 3.7v 18350 batteries that are located internally
Sorry for the super late test, work got in the way
Nope. that should make no difference whatsoever. You can test that by opening up the receiver tab in betaflight with your radio on, and see exactly what signals your drone is receiving from the radio.
ah then idk, but maybe what that joshua person said will work
good news!
updated betaflight
and now the blackbox works
will test tomorrow with external video and will send here
thank you for the patience guys, you have all been extremely helpful
I thought you already had 4.3.2? me>confused
i didnt have the latest configurator, so I downloaded the latest, then got told that it was buggy by the betaflight discord, and then got one of the nightlies and now blackbox works.
Called it
Im no expert, but it seems like input delay. I am not sure though
I feel like theres something wrong with this log.
I've only got 4 seconds, and yeah it looks like there may be issues.
Oh, I see there were only 4 seconds of flight.
Not running OSD...
Is this blackbox the same as from the video posted?
Just checking if the maximum crank on pitch right before disarming is consistent with the inputs. Although the quad reacted incorrectly nonetheless.
Also I don't want to look like I'm always blaming feedforward, but every time your quad maxes out motors in the wrong direction it is associated with a feedforward spike. So let's turn that off first
what is feedforward?
also I think it should be the right log, but I could check
It should be the latest one as I turned off the drone right after
also how do I turn off feedforward in betaflight?
do you know how to get to the pid tuning page?
yes
yeah, turn that boy on
and then some more sliders should pop up
one of them is stick tracking (and should also be labeled feedforward) slide that baby to 0
Here are the rest of the logs, I will look through em, but in case I miss something then here they are
some are just test arms, so they dont have any actual data tho
gotcha
should I do a test flight like that to see if it fixes anything?
probably won't, but it would give me data that i'd be better at reading
so i'd like a flight with FF at 0 🙏
which one?
Stick response (FF Gains)
got it
it's my thought that it's overreacting to small twitches in your sticks and making the quad confused
hmm, is that something I will need to replace the FC in order to fix?
I hope not
I wish you luck deciphering this
I don't see that motor maxing out stuff anymroe
Is that good?
helps me read it

Okay, so by the blackbox, you commanded full right stick until the drone was almost completely upside down and then stopped
It's weird and the gyro isn't following that closely but it looks like it's following
Maybe? I'm not sure how
I have the radiomaster zorro
Also when I move the sticks in betaflight, it seems to correspond to the right stuff
this is what your quad is seeing
it's seeing you pull the stick all the way to the right and is doing the same
System settings, hardware tab (should be all the way to the right), calibrate. Move everything through the full range of motion gently.
I see, it seems like even moving it a tiny bit rn it causes the stick to shoot to the extremes
Hey I got some really good news
It works!
Now it Flys well!!
Also should I turn this back up Now that it works?
Thank you guys for the help! You all have been incredibly helpful and this experience has been a wonderful introduction into this hobby, I cannot express how thankful I am, I have been wanting to fly this drone since I ordered the parts
:)
You can!
Issue wasn't with the quad at all, I see.
Yes, luckily, as I really didn't wanna have to order a new FC or ESC
another idea: can you switch your RX to a other UART?
this could be a timer issue, where the RX cant talk fast enough with the FC
just a wild thought i had when you came up with the "lag problem"
and turning off FF doesnt solve the issue, it freaking around is more of a symptom - it can make things better, but the underlying issue is still there
as i am stupid and didnt read all - the problem was, that your radio wasnt calibrated? @amber vector
@ember dock
ffs - thats such a easy to miss thing -.- i read all now and yes this was the case - ffs nobody though about that holy smokes
that's why I wanted the blackbox the whole time @scarlet gust 😄
you can see RC traces in the blackbox log xD
i think if we would have had bb from the start, this wouldnt have taken so long
yeah
you are right, if we would had the indication, that his sticks were all over the place
Hey, I only worked it out after reading 2 whole blackboxes that you guys helped him fix, no big deal!
It was present in the first one, but I saw the motors maxing out and had to know if it was him or the quad
My mental process was this: The drone is flying fine for a second before flipping out on its own, odds of it being sticks: 30%
tbh i NEVER thought about the radio
but yes, it looked like the quad was reacting to a movement command - that was the strange part
Has anyone run tasks in cli yet to comfirm
I know it's really unlikely but it'd only take like 5 seconds to do & I've had a tinyhawk do this same crap before when cpu was too overloaded
It very well might just be a defective gyro too though I've also seen this as well
What stuff would you need me to run?