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?
The motor is junk. I wouldn't trust it even if it still works. Your ESCs are all junk IMHO because of what they did to your motor. The ESC is a small computer with one input, 1 or 0 (PWM). Its output is a group of 6 (maybe 12?) MOSFETS that are either 100% on or 100% off. If a valid PWM is not received by the ESC, it has absolutely no business firing any MOSFET!
Its all in the sequence and timing of the on/offs. If the MOSFETs are not pulsed in the proper sequence, and the moter doesn't spin, then it's just like a dead short, and thus smoke and damage.
It's like this. Look at one of the windings on your motor. It's maybe what a few inches of solid copper wire. If you unwrapped this wire and connected it across your battery, what happens, smoke. It's the spinning and and electromagnetic magic that limits the current.
An ESC's firmware should never let this happen, yet apparently some do.
And don't worry, the APM should be fine.
It is also my impression that the direct cause of hardware is the ESC behavoir given spurious input, and that this input can be expected during board resets (I do not know if the spurious input could be prevented, I'm sure some EEs here would know.)
I have never seen this with Castle Creations ESCs, and although I have 16 jDrones 20A ESCs, I have also never experienced it myself.
Exactly what mike said! And this has never happened to me, but I have seen twitching servos on a plane if I flashed the firmware. I have never had a Castle Creations try to run or twitch when it shouldn't, but they do cost about 3X as much. Is it worth it? I think so.
I got the Castle Creations ESCs because:
1) at the time, jDrones/DIYD/uDrones/etc were out of stock
2) I wanted to do some motor testing, and Castle Creation has awesome data collection
3) They were in stock with another parts vendor where I was already buying gear
They are outstanding. If you get them, I highly recommend getting the field programming card. It will make you happy.
That being said, they are *expensive* and heavy. I use Pheonix ICE Lite 50 in one of my quads, and I have some thunderbird 35s, but until a recent awesome post about ESCs by Olivier ADLER, I was only avoiding using the thunderbird on multicopters out of blind adherence to CC's recommendation that thunderbird not be used on rotor craft.
Even the smallest Pheonix ICE Lite is twice as heavy as the 20 and 30 amp models like the jDrones. When you multiple that by four (or multiply the $80+ cost per unit by four) they get less attractive for everyday flight.
I have "retired them" for use on test benches and higher end designs for now; and since I have not had any problems with the jDrones ESCs, I am only slightly missing the additional programming features, data collection, and pretty LEDs.
They are worth every penny. They are just not always as well matched to my needs for experimental, light copters that may die a watery or high g death... or where every gram counts. And I like them so much that it makes me sad that they are not a better fit in these areas (cheaper/lighter) But I don't think they can they can be as excellent as they are AND be any cheaper or lighter.
I think your on to something with the resets. I know you are. It may be the case that when you hit the reset, the ESC has all ready been "armed" by it's own safety rules. Don't hit the reset button with the ESC's all ready "armed"(valid 0 throttle signal). Cycle the power, thus also resetting the ESCs. Never hit the reset button. Maybe that's the ticket aye ;)
Well, that certainly would describe CC. There are rare stories of issues with them. They aways are presented in the context of the author raving about how responsive, helpful, and generous the guys at CC are in quickly solving the rare issue. Such is my impression. I've never needed to contact them, and I have 12 CC ESCs (three different models) and their field programming card, and two castle link products.
And remember that by default setting attaching the USB also causes an APM reset; so if servo power is applied when you attach.... it can cause the same conditions.
This is an an ufortunate occurance - I burned a perfectly good roxxy (mikrokopter) motor under the same circumstances. I was unfrotunately using a cheap ESC (and still using them!). The problem was reproducable if I plugged in battery first, then the USB. Now I plug in the USB only for most things, and only plug in the battery (after USB) to calibrate radio. The thing I noticed is the motor twitching is much reduced if the APM fully boots up on USB power before plugging in the battery.
It is unfortunate the cheap ESCs repond to what is presumably some sort of signal noice from APM boards during intial start up or reset - So it seems like a combination of poor EScs and some unexpected signals on output channels of APM.
On my planes and standard Helis I only use castle ESCs, but just last week the Castle BEC that I was using to supply 5V to the APM failed and started delivering input voltage to the output - this fried my APM. Surprisingly Castle has not responded to my emails regading this. So I agree that Castle produces great products, but unfortunately they are not exempt from issues either.
Hi Al, I remember you now. I use these BECs. They have always worked for me with 2x9g servos and everything else. They have a 3A one if you need it.
They replaced a bad ESC for me that had cut mid-air. not cool to see your model crash to the ground, but they replaced it promptly and we're gracious over the phone.
I load code and power up the boards probably more than anyone and most combined. I've never had this issue for one reason. I never press that damn reset button. NEVER. And I never have the USB plugged in and Lipos connected at the same time. I don't know that much about the hardware (I'm the polar opposite of an EE,) but I've seen the CH1 output on the scope and nothing can be done about it on the SW side. It's an AVR quirk, I guess.