Alright!! So, learned some lessons this time around (props to unseen for helping me on the path to enlightenment!). The issue is, ~50% throttle on punchouts, and one ESC is not producing. What I mean, is on punchout, all are giving me around 50% power, except #4 which is pegged at 100% as the quad lists over. Dumping the data and the blackbox. This was a live flight, props on  . Quad was battery powered on level ground, so the gyro should not be an issue. ESCs had been calibrated, but now trying to calibrate them and they do not respond. 

Happens the same as this guys video
 



Quad:
Armattan Chameleon
DYS XSD30AV2
Emax 2205 2300KV
CL Racing F4 fc
TBS Unify Pro HV
X4RSB Frsky
ZOP 1500mAH battery (45C)

Log of flight

One motor only being at 100% means there's something wrong with that motor or esc or maybe the solder joints... The other motors are held back in order to keep attitude.

No clue about the calibration issue though.

Pretty sure it is the ESC. I just replaced this one and will see. I have noticed that the same ESC burns up (same arm) each time. I have swapped motors to see if it could be the motor. Is there any way a flight controller can toast an ESC?

Also, on the calibration issue, I fixed that. MY ESCs are set to DShot300, and DShot is already calibrated. However you can force a calibration on them by putting them in MultiShot/OneShot from the config tab, doing a normal calibration, and then setting them back to DShot.

Hm maybe because of timer issues... On some versions of betaflight there was a cli variable called "sdcard_dma" or something similar which was often recommended to turn off when such issues came up.

fftunes Wrote:Hm maybe because of timer issues... On some versions of betaflight there was a cli variable called "sdcard_dma" or something similar which was often recommended to turn off when such issues came up.

Pretty sure I found the issue. So I have noticed that when I use the artificial horizon, even after starting it up on level ground, it seems to yaw left. No idea why. Pulled up the sensors tab and found this gem (see attachment). 
   

It starts off when plugged in on zero and slowly rises up to around 10. Pretty sure that gyro ate it at some point. I would assume that, coupled with a bad esc is causing these problems.

It certainly looks like there's a problem with the ESC that drives motor number 4. As you can see in your log, the craft actually starts to tilt towards motor 4 when the motor is being asked to deliver maximum power. If changing the ESC does not solve the problem, check carefully that none of the mounting screws for motor 4 are too long.

If any of your mounting screws are too long they can short motor windings to each other or to the frame.

The artificial horizon on the OSD is derived from the accelerometer data, not the gyro data. You should calibrate the accelerometer in the configurator while your quad is level.

The gyro measures the speed of rotation around each axis. What you show happening in your last post is definitely not right. Unless the quad was actually spinning clockwise around the X axis at 8 degrees per second, I'd suspect there is something wrong with the gyro. However, I don't actually see that behaviour in your black box log. Even right at the end of the flight, when you return the sticks to centre, the gyros return to zero on all three axes.

unseen Wrote:It certainly looks like there's a problem with the ESC that drives motor number 4. As you can see in your log, the craft actually starts to tilt towards motor 4 when the motor is being asked to deliver maximum power. If changing the ESC does not solve the problem, check carefully that none of the mounting screws for motor 4 are too long.

If any of your mounting screws are too long they can short motor windings to each other or to the frame.

The artificial horizon on the OSD is derived from the accelerometer data, not the gyro data. You should calibrate the accelerometer in the configurator while your quad is level.

The gyro measures the speed of rotation around each axis. What you show happening in your last post is definitely not right. Unless the quad was actually spinning clockwise around the X axis at 8 degrees per second, I'd suspect there is something wrong with the gyro. However, I don't actually see that behaviour in your black box log. Even right at the end of the flight, when you return the sticks to centre, the gyros return to zero on all three axes.

Yea, checked the screws. The things are almost too short to screw in enough. 

Pretty sure the horizon issue was because I was arming on the hood of my car, which is tilted. However, even when doing it on the ground, the acc. would be good, but the gyro would still drift.

Not sure entirely on the gyro issue. I did email Mr. Cheng, and he responded almost immediately. Sent him a pic of the gyro sensor data and he said I can send it to him for repair, if possible. That guy is AWESOME! I plugged in a battery a few minutes ago and one of the motors that were on that same failing side put on a smoke show. So, it just keeps getting more bizarre. However, considering it was on the same side as the ESCs that kept failing, I am pretty sure it was bad in the first place. So, now to the corner to get enough money for a new set of motors. I have lost trust in these used ones.

From what I understand from your posts, you've replaced the same ESC several times and you've also swapped motors around without the problem following the motor.

If that's correct, then either you are unlucky enough to have two faulty ESCs, two faulty motors or there is something you are doing wrong.

The flight controller itself can't cause an ESC to smoke a motor. The only thing that applies power to the motor is the ESC and there is no command that the flight controller can send which would cause the ESC to smoke the motor. But, you say that the motors you have are used? As in, second hand? If so, maybe the previous owner used screws that were too long and damaged the windings. An intermittent short circuit on the motor windings would certainly cause problems driving the motor and even cause it to burn up when the ESC uses the motor to make the beep sounds you get on power up.

While you are waiting for new parts, build yourself a smoke stopper and save yourself from smoking motors and burning ESCs.

unseen Wrote:While you are waiting for new parts, build yourself a smoke stopper and save yourself from smoking motors and burning ESCs.

Most definitely will! I have already got two of the techs I work with to build quads as well, so we will make it a community effort, lol!

Setup: Motors were second hand from RCGroups. ESCs were second hand as well.

As far as the ESCs, this is the order it has gone in.

Start:
35a BL Heli Bullet ESCs with current motors
 - Noticed the quad wanted to drift, half power only, and listing
 - No issue of ESCs failing
 - Figured issue was front left ESC
 - Ordered new DYS ESCs
 - Put on new ESCs and tested. Quad performed superbly, despite drift issue.
 - Second flight, next day, quad started listing again with power cuts
 - Replaced front left ESC and same thing. Half power and listing.

I have swapped the motors around (1 to 4, 2 to 3), but have noticed no issue there. The only thing I have noticed that is similar in all of the failures, is that these are always going out.

I have checked the ESCs, checking motor tab to ground, usually gives a reading of ~10kOhms. Except this side, whenever the ESC starts acting up, the reading goes to crap. Not even a solid Ohm. Not sure what is up with that. Would not think these motors would kill 30a ESCs. I think I am going nuts. I do have two more motors on order to replace the one that smoked, and one that gave me high  amperage leakage during the test. So we shall see. Couple that all with a new FC, and if this thing does not fly well, I believe someone will be getting a great deal on a barely used Armattan frame, lol.

I'd never buy second hand motors or ESCs. You've just no idea how the previous owner has treated them. I don't really understand what you mean when you talk about 'checking motor tab to ground'. Can you describe what you are measuring? From where to where? Under what conditions?

Measuring the resistance from a motor output on the ESC to ground with no motor connected but the ESC connected to the flight controller can give you very misleading results. Just the residual charge on the gate of the output FET is enough to turn the FET on. This is why your motors can feel harder to turn just after flying, even though the battery is disconnected. An hour or so later, when all residual charge has leaked away, they will be easier to turn.

When you get your new parts, post some pictures of how you have put everything together.

unseen Wrote:I'd never buy second hand motors or ESCs. You've just no idea how the previous owner has treated them. I don't really understand what you mean when you talk about 'checking motor tab to ground'. Can you describe what you are measuring? From where to where? Under what conditions?

Measuring the resistance from a motor output on the ESC to ground with no motor connected but the ESC connected to the flight controller can give you very misleading results. Just the residual charge on the gate of the output FET is enough to turn the FET on. This is why your motors can feel harder to turn just after flying, even though the battery is disconnected. An hour or so later, when all residual charge has leaked away, they will be easier to turn.

When you get your new parts, post some pictures of how you have put everything together.

Yea, will not be buying much else from there. Kind of a let down.

Anyways, I am measuring here

The ESCs have been off the quad for a bit, and it has not been powered up in days at that point. I was trying to find something common between these ESCs that seem to keep going out, and that is all I could come up with. I had noticed at one point that a motor was just a tad slower than the others, and was curious if it had an effect.

Measuring on a Fluke 87 V, resistance setting.

I also measured the motors using the fluke on AC setting, and spinning the stator with a drill. Gave me 2 motors with a consistent reading (across leads 1-2,1-3,2-3) of around 409mV, and two motors (including the smoked one) with odd readings. One read a little high, around 420mV, and the other burned motor, read two phases around 409, and one phase at 390.

I will be shelving it for a bit until I get everything in. Sorry for all of the information dumping. I hope at least a little of this madness made sense! And thank you kindly for all of the wisdom!

Hmm...I'm not quite sure how useful measuring the outputs of the ESC to ground is. Testing the motors as generators is a help for sure but I don't know if the variations you saw indicate a fault or not. Ideally, you'd want an inductance meter to test each phase and make sure they're within a few uH of each other, plus a 4-wire resistance meter to check the resistance of each phase.

I'll look forward to the next instalment!

(16-Oct-2017, 08:17 PM)unseen Wrote: Hmm...I'm not quite sure how useful measuring the outputs of the ESC to ground is. Testing the motors as generators is a help for sur,e but I don't know if the variations you saw indicate a fault or not. Ideally, you'd want an inductance meter to test each phase and make sure they're within a few uH of each other, plus a 4-wire resistance meter to check the resistance of each phase.

I'll look forward to the next instalment!

Haha, Christ. I'd rather it not be looked forward to, lol. Almost foreboding. Anyways, to clarify a little on the ESC test I was doing. I found that the ESCs that we're performing optimally, would have uniform readings of 10.98kOhms. This was both the DYS and surprisingly, the Emax bullets as well. The ones that had issues, on that first side I highlighted would be off. Just found it odd. Not exactly sure how it applies, but it's data none the less. And yes, inductance meter would be sweet, but my shop (electronics shop) doesn't seem to have one

Out of idle curiosity, I just dug a bunch of random brand new ESCs out of my spares box and made the same measurement.

V1 RacerStar 20A BLHeli_S : 11K
DYS XM20A BLHeli : 11K
FVT Littlebee 20A BLHeli : 44K
FVT Spring 40A BLHeli_S: 54K

If I reverse the leads on my Fluke and put the positive probe on the ground wire, the resistance rises with each second I keep the probes on the device.

Maybe if there's a significant difference in the measurement between the three motor outputs and ground it might tell you something - there was no significant difference between terminals on the ESCs I tested.

If you look at the output stages of any modern ESC, there are three of these circuits:



A dedicated high and low side driver chip drives banks of high power MOSFETs to generate the output signal to the motors.

There's also the simpler transistor based way of doing it, but the result is pretty much the same:




Each phase output is connected to ground via a high resistance voltage divider and at the mid point to the MCU via a current limiting resistor. This circuit is used to observe the back EMF from a phase when it's not being driven.

What we're measuring with the positive on the phase output and the negative on ESC ground is the voltage divider and all the parallel resistances created by the rest of the devices on the phase output. There are a bunch of floating FETs and a driver in that circuit as well, so I'm surprised how consistent the resistances were, varying by at most 200 ohms.

I test ESCs by connecting them to a motor on my thrust stand and seeing if they can spin a motor and prop. If you don't have a thrust stand, screw a motor onto your bench! I power up the ESC for the first time on a current limited bench supply. The cheaper alternative is a smoke stopper.

Testing and checking stuff should be simple and quick.

Building should be careful, methodical and thorough.

After each step that involves soldering, I inspect every aspect with a microscope or loupe. I've picked off so many tiny little strands of wire and solder balls you wouldn't believe it and not all of them are mine! If you double check everything you do as you work, it's no more time and saves tears later. Once I'm finished with a build, I'm confident that unless I have a bad component, everything should work when I power up for the first time.

It's all too easy in this hobby to get bogged down in details, particularly if you're interested in electronics, robotics and computing. When I'm finished building a new quad, my checklist is short and simple:

1. Check resistance on battery input.
2. Power up flight controller with USB.
3. Power up quad from bench supply with max current set to about 750mA.

If these first three steps pass, I'm confident to connect a battery for the rest of the steps.

4. Check motor direction and correct as required.
5. Calibrate ESCs (if required).
6. Find and set min_throttle.
7. Bind and check receiver.
8. Set favourite rates and mode switches.
9. Check flight controller orientation and calibrate accelerometer.

Then I add props, go outside and fly!

Configuring the OSD, the FPV camera and fine tuning can be done after the line of sight maiden flight. 

I don't buy used stuff. Not ESCs or flight controllers, soldered to or not. Never motors, unless in unopened new packaging. You never know what people might knowingly or unknowingly pass off and it's hard to be sure you didn't buy a lemon before your quad falls out of the sky.


We're here if you need help!