Trouble with brush less motors and Power.

[vortix]

Bummer yah working on electronics can be infuriating easy to short something or otherwise make a trivial mistake that sets you back days, it's why I've been collecting more diagnostic and repair stuff over time too. Personally learning about STM32 tools and development at the moment and it's super hit or miss on if an imported project will compile or not and flash or not depending on tool chain and seemingly a million configuration/build variables, incredible anyone is able to make this stuff work

Yeah, i think im just gonna use some online service that creates a board for you. that would save me alot of headache.

 

[vortix]

Bummer yah working on electronics can be infuriating easy to short something or otherwise make a trivial mistake that sets you back days, it's why I've been collecting more diagnostic and repair stuff over time too. Personally learning about STM32 tools and development at the moment and it's super hit or miss on if an imported project will compile or not and flash or not depending on tool chain and seemingly a million configuration/build variables, incredible anyone is able to make this stuff work

ok everything works great except one thing, It drifts left. I tried adjusting the PID many times but i haven't been able to figure out which one should i change. im trying to make an angle controller now it works except this drift.

 

[wafflejock]

Drift or accelerometer drift is a really common thing especially among cheaper flight controllers (off the shelf stuff). My KingKong 90GT has some pretty bad progressive drift that just gets worse over time as the accelerometer value drifts. I know there are ways to mitigate that but not sure exactly what you should do, in general look up accelerometer drift and see if that's your issue and any proposed solutions.

Personally I just fly that one in acro mode most of the time and ignore the problem but I'm sure there are some ways to correct for it over time if it's consistent.

 

[vortix]

Drift or accelerometer drift is a really common thing especially among cheaper flight controllers (off the shelf stuff). My KingKong 90GT has some pretty bad progressive drift that just gets worse over time as the accelerometer value drifts. I know there are ways to mitigate that but not sure exactly what you should do, in general look up accelerometer drift and see if that's your issue and any proposed solutions.

Personally I just fly that one in acro mode most of the time and ignore the problem but I'm sure there are some ways to correct for it over time if it's consistent.

but i think there is a problem because the drift feels significant. significant enough to feel like my room is too small for it. i expected a drift but not that much. I suspect that its the build im doing. i have two layers above the motors. One floor takes the battery and another takes the flight controller. its weird looking but thats the only build that seem to work. if i move to the standard: flight controller right on center between motor and battery in the bottom, it doesn't work. it wont even fly. i have yet to figure out why that is the case.

 

[wafflejock]

Just a quick glance at the code it looks like there are some parts related to initializing the accelerometer perhaps you could insert some artificial delay(5000) call during setup or something so it waits 5 seconds before initializing the accelerometer so you have plenty of time to plug in the battery and get away from it and let it settle into a flat position before the accelerometer gets calibrated. If it happens right on take off it sounds like the initial calibration is just off, for mine it happens a bit initially but progressively gets worse over time (pretty quickly)

 

[vortix]

Just a quick glance at the code it looks like there are some parts related to initializing the accelerometer perhaps you could insert some artificial delay(5000) call during setup or something so it waits 5 seconds before initializing the accelerometer so you have plenty of time to plug in the battery and get away from it and let it settle into a flat position before the accelerometer gets calibrated. If it happens right on take off it sounds like the initial calibration is just off, for mine it happens a bit initially but progressively gets worse over time (pretty quickly)

I already do that . i give it about 15 seconds or so before reading the initial. it happens pretty quickly. i would says about a second of stable in place hover and then without touching the direction stick, it starts to drift significantly.

 

[wafflejock]

Yah that's still regular behavior for accelerometer drift, honestly not sure if betaflight just handles that issue with filters/corrections or what exactly (might just depend on the particular IMU used but I believe 6050 was a popular one at least for a while)

MPU6050 accelerometer drift – DIY Drones

^^ maybe link this thread somewhere on that forum see if anyone with experience solving the issue can help you out (since I'm straight guessing now )

 

[vortix]

Yah that's still regular behavior for accelerometer drift, honestly not sure if betaflight just handles that issue with filters/corrections or what exactly (might just depend on the particular IMU used but I believe 6050 was a popular one at least for a while)

MPU6050 accelerometer drift – DIY Drones

^^ maybe link this thread somewhere on that forum see if anyone with experience solving the issue can help you out (since I'm straight guessing now )

i ran into this dilemma. if i increase the I_gain, the quad tends to try to stay where it is but starts to oscillate heavily, if i decrease it, it becomes more stable but starts to drift so much i cant really call it drifting, its more like moving to that direction. I followed every guide to reaching the appropriate PID and it ain't working. Whats even more annoying is that I KNOW ITS POSSIBLE. i find a fantastic sweet spot that drifted just a little bit with a bit of yaw. So i changed it to try and fix the yaw and forgot to write it down. now i lost it i feel like a goof.

 

[wafflejock]

Yeah I'm honestly super surprised you got it to fly (no offense, it just sounds pretty difficult). Like you said it must be possible because ardupilot does it and sounds like you saw it close to working already, thing is little tweaks to how things are mounted (softer or harder mounting more or less vibrations from frame/motors getting into gyro/accelerometer data) can have a pretty big effect in how things perform and I know filtering is used largely to help alleviate the issues of all that extra erroneous data getting into the PID loop basically, unfortunately you've gotten way beyond my pay grade at this point

 

[vortix]

Yeah I'm honestly super surprised you got it to fly (no offense, it just sounds pretty difficult). Like you said it must be possible because ardupilot does it and sounds like you saw it close to working already, thing is little tweaks to how things are mounted (softer or harder mounting more or less vibrations from frame/motors getting into gyro/accelerometer data) can have a pretty big effect in how things perform and I know filtering is used largely to help alleviate the issues of all that extra erroneous data getting into the PID loop basically, unfortunately you've gotten way beyond my pay grade at this point

thank you for your help. i want to improve on it further but right now, its good enough for a semester project . by the way, i think i have a plan for a tiny quad that might work. excited to give it a try! ill use 4 mosfets one for each motor and a regulator to 5v for the arduino. i hope that works!

 

[vortix]

thank you for your help. i want to improve on it further but right now, its good enough for a semester project . by the way, i think i have a plan for a tiny quad that might work. excited to give it a try! ill use 4 mosfets one for each motor and a regulator to 5v for the arduino. i hope that works!

hmm, i encountered this frustrating issue that i dont know if there is a fix for. The Ground solder plate of the ESC broke off. is this a fixable issue?

 

[wafflejock]

Unfortunately probably not you need a low resistance/high current capable path from the battery power to the ESC. I guess could try soldering the negative onto the GND of the MOSFETs on the ESC but wouldn't put money on that working.

 

[vortix]

Unfortunately probably not you need a low resistance/high current capable path from the battery power to the ESC. I guess could try soldering the negative onto the GND of the MOSFETs on the ESC but wouldn't put money on that working.

its safe to say that im losing my mind slowly .
everything works. all good. than i dropped my transmitter. potentiometer for throttle broke. replaced it with a new one and now its not working. all transmitter output looks the same. code is the same. but now, the quad turns off within 2-3 seconds of flight time. nothing happened to the quad , just the transmitter. i dont understand how the potentiometer change the outcome so much.... i thought maybe transmitter is damaged but the output is all the same and both potentiometers are b100k . i was on my way to show my instructor damn it....

any idea what the problem could be ?

 

[wafflejock]

its safe to say that im losing my mind slowly .
everything works. all good. than i dropped my transmitter. potentiometer for throttle broke. replaced it with a new one and now its not working. all transmitter output looks the same. code is the same. but now, the quad turns off within 2-3 seconds of flight time. nothing happened to the quad , just the transmitter. i dont understand how the potentiometer change the outcome so much.... i thought maybe transmitter is damaged but the output is all the same and both potentiometers are b100k . i was on my way to show my instructor damn it....

any idea what the problem could be ?

Sorry no idea really, would just start probing around with multimeter make sure battery voltage is good double check your solder connections make sure nothing connected that shouldn't be (check for shorts to GND that shouldn't exist vs those that should). Try to determine if the ESCs and all are rebooting or if it's staying powered on but just losing receiver input. Will be hard to test since bench testing the receiver won't have all the extra electrical noise it has to cope with when flying but probably a good place to start ideally just hook one ESC and motor to an Arduino or whatever to send the PWM signals out and read the receiver data (really could just go receiver right into Arduino and read the signal and output to serial for debugging the transmitter/receiver separate from everything else)

 

[wafflejock]

Usually with my own custom transmitter/receiver I made for my esk8 stuff I'll use a scope to probe the output of the potentiometer or hall sensor or whatever is being used to sense trigger position to make sure that first level of sensor data is working (basically probe input pin to transmitter mcu/radio) and if that looks good I'll look at the output pin (again using my $35 DIY handheld scope) and see that the values being received and sent to ESCs looks good (output pin on receiver)

 

[vortix]

Usually with my own custom transmitter/receiver I made for my esk8 stuff I'll use a scope to probe the output of the potentiometer or hall sensor or whatever is being used to sense trigger position to make sure that first level of sensor data is working (basically probe input pin to transmitter mcu/radio) and if that looks good I'll look at the output pin (again using my $35 DIY handheld scope) and see that the values being received and sent to ESCs looks good (output pin on receiver)

thank you. it seems that the ESC was overheating. i checked connections and such and then replaced the metal screws that i used to attached to the frame with nylon .it seems to work. might have been a short that i made happen by accident.