I wanted to report a potential problem and maybe see what can be done to fix it, or if it solely my fault, what I can do to avoid it in the future.
I just installed a new receiver (Hitec Optima 9) and calibrated my radio. Next up was calibrating the ESCs. I proceeded to plug the signal wires (going to the ESCs) into the APM, plug in USB, flipped the APM into CLI mode and hit reset. Immediately the channel 1 motor began to shake violently and smoke poured out.
I recognized this symptom from my reading of the manual, so I went to read up on what went wrong. (See http://code.google.com/p/arducopter/wiki/AC2_Troubleshooting , Section - One of my motors started shaking and then burned out!)
I had heeded the warning and only plugged in the wires for essential testing (I don't see how this warning would prevent this problem, as I followed it and the problem occurred). The description on the troubleshooting page clearly states that this problem only occurs with APM1280, certain ESC types, and the older Arduino bootloader. I am running an APM 2560 provided by jDrones along with jDrones 30A ESCs and jDrones 880kv motors. So the scope of the problem is clearly not limited to just that hardware.
Does anyone have any insight as to what happened? Can someone provide any technical detail on what exactly causes this to happen? If this is a common problem with only certain ESCs, should the community not adopt different official ESCs?
Can anyone tell me what the likely extent of the damage is? Do I need to replace both the ESC and the motor?
I only have three APMs and I am not nearly as active as Jason, but I maybe have a couple thousand USB/battery connects already, more than your average bear because I also use a HAM power rig, so "battery" or servo power is trivial to me when testing at home (props removed or tethered.)
I have castle creations ESCs, but by now at least 1/2 of my power cycles/connects are on jDrones, since I have two quads w/ jDrones 20A, and I retired my quad that uses the CC ESCs.
I have not once (so far, knock on wood) experienced this burn out, and I am quite/too cavalier w/ USB vs Servo order, and reconnects. But I can easily report that I have used the reset button on an/any APM only once, and it was curiosity, not need, when I did it.
I'm thinking that avoiding the reset button may just well be a good practice. It might not prevent this issue, but it certainly seems to happen more frequently to people who use the reset than to those who do not.
I regular run code that I just wrote so I'm religious about the absolute safest route. Just sayin....
Be careful about plugging the USB in, it will reboot the APM.
And use care with the mini USB; I've "lost" one IMU due to lifting the connector off the board. I plan to make a second attempt to reflow it to the board, but in my first attempt, I did not remove the "broken" solder, and it worked electrically, but failed again mechanically.
I've reinforced one of my other IMUs, but left one untouched. I'm not certain if I helped or compounded the issue, and did not want to risk damaging a nearby SMD weld.
I'm curious, Jason, if you do anything special to avoid this issue? I expect you connect via USB more often than nearly everyone, but expect that you might mainly leave the mini-USB connected to the IMU you use for most code-reload, and only remove from the computer? I tend to keep use whichever aircraft is seeing the least amount of work connected to the miniUSB, and perform any experiments on it (for example, last night I was using ArduPlaneHIL on one of my quad-mounted APMs because I do not have to unplug/attach the miniUSB needlessly - and I do not plan to use the quad in the next few days.)
I have been down this path like you recently and started with the same ESC's (30A) and 880Kv Motors. I had an unfortunate brown out on my maiden outside flight, after weeks of bench testing. i have recently spoken about the various motor issues, but I had a few interesting exchanges with other members including Olivier here on ESC's. There is some good motor information and also some good ESC's choices to consider. For me I am looking at a few motor choices at the moment I like the Axi's models made for multicopter use but they come at a price. Pulso's are a good Axi clone but for precision bearings (they have 3) its hard to go past Axi.
I like CC and use their UBEC's. I have no experience with their ESC's, but the posts here have spiked my interest. Also have a look a using Opto ESC's if you do not mind paying more. There is a lot to be said about isolating the noise from the ESC from the APM. I have also recently been looking at YGE who make an opto cable which you can use with any ESC and this is cheaper than say the 40A Platinum Pro ESC's discussed in the forum post link I gave you earlier.
Regarding booting the APM as Jason said there is nothing we can do in software about the CH1. However given I have the same ESC's and motors as you I do take the following precautions which works for me.
I never power the system using the copter lipo unless I am flying. If I have to do any reset and radio calibration then I use a Castle UBEC to provide the 5+ regulated power to the APM and Receiver. I have a separate small lipo or flight pack I use to power the UBEC. I have the signal wires detached from the ESC's anyway unless I am about to fly with no dramas. I did this because the constant regular beep from the motors without the Signal from the ESC is really annoying after a while. Whats more to the point if your doing any calibration the slight twitch from the motor intime with the beep registers on the IMU sensors. For example have a look at the MP flight view and realtime gauge readings while the motors are beeping away and doing there tiny twitch. There is no need to send power to the ESC because the external UBEC avoids having to use the internal ESC bec which normally supplies 5+v to the APM and Receiver.
I do all my programing work from the USB and anytime I need to re-calibrate an power the receiver etc. I just connect the +5V to the APM as I described. Also calibrate your ESC's manually so again no need to have the risk of the signal noise from the APM affecting your motors.
Of course I agree with Greg, a good ESC will not cause these motor issues.
I hope this helps
That's why it is better to use opto ESCs if you have an expensive load, they will never send a voltage at the input thanks to the opto isolator. So if you loose an ESC / motor on a hexa or octo setup, you have better chances to recover.
Next, the opto permitt to stay away from ground loops problems and is a very good RFI filter.
Unfortunately, the smaller opto ESCs available are in the 40A range, or 30A but almost nobody has that in stock.
It is possible to add a 8 channels opto isolator between the APM and the ESCs. Some manufacturers do have this kind of module or it is even possible to choose the do it yourself way, as it is not a complicated circuit.
The only drawback i see with opto ESCs, is that you need a separate BEC. But is it really a drawback ? I've always found safer to have a separated BEC. (i have two BECs on my hexa frame for added reliability).
Normally you can't do that, except if the BEC manufacturer do permitt to do it.
A solution is to use a dual shottky diode or two shottky diodes.
You will loose 0.2 or 0.3V through those diodes.
I recommend Castle Creations and use their UBEC's and ESC's too.
UBEC is the best I have found, you can set output voltage with CastleLink up to +12v if needed.
Great for Video Tx power supply!
Dimension Engineering has great stuff too.
I connect / disconnect AP USB many times daily and on all AP units sold here and several others.
I follow these guide lines all the time...so far NO blood or component smoke! (I did melt wire harness on a FPV once, but these rules would not have helped, I had a pinched wire causing a shot to +12v battery. (Replaced harness to fix, no harm done to camera or video TX)
APM Testing and Operation Safety Procedures. Protects you and the UAS.
1. Do NOT connect Motor lipo's unless you are testing the motors or flying!
When bench testing, first test with NO props! if all motors spin OK, then add props to continue your testing.
Additional method for ensured safety with props ON.
Remove one of the 3 power lines to all motors, (when power is applied motors will twitch and try to spin but at very low power, you will know it is getting a signal, as it twitches but won't spin up.
Note: ONLY run motors for a second or two, I don't know if prolonged test with 2 power lines will damage motors or your ESC. ( I have done this many times with many motors an ESCs.)
2. Do NOT test FailSafe function with props ON! Motors can go to cruise speed 30% or more when FailSafe is activated and mode automatically goes to RTL. If radio is turned OFF, or signal fails this will happen, if FailSafe is setup correctly.
3. I usually test with 2 or 3 small servos only and use USB supplied +5v
If all servos are used,
Power up APM with a 2 cell lipo battery or a regulated power supply, Set to +6 volts (APM will regulate it down to +5v) with 3 amps, if using more than 4 digital servos, 5 amps or more may be required.
4. Since I have two computers connected to APM when developing OSD integration code, I also connect power to OSD board last, add disconnect it first.
Do you have a schematic of how to set that up? I've never felt comfortable with only one BEC. :-/ I use a 10A Castle which should be robust enough but I like the idea of having a backup :)
Just take the +5V output of each BEC, take two shottky diodes (strong enough) and connect each diode anode to each BEC output.
Then connect the diodes cathode together and you have a redondant power supply.
If you are not confortable doing this, you can buy a ready made shottky module.
How do you match the 2 BECs?
Puting together straight away? or any diode in between?