Had a pretty traumatic experience today while I was trying to get an APM 2.0 working on an RTF "trainer," when I turned off the transmitter and the ESC burst into flames. I was hoping the community might be able to point me in the direction of a solution.
I was ground-testing an electric RC (a Hobbico NexStar electric) with an APM 2.0. The APM was completely stock, and had been loaded with firmware from the Mission Planner without any modification. Controls were calibrated on the Mission Planner as instructed in the wiki, with CH5 on the transmitter dedicated to switching between manual and "stabilize." This was using a Futaba 6ch controller.
The jumper was connected on JP1, in the stock configuration on the APM 2.0. I was supplying power to the output rail via 4.8v NiCd power source, which I checked beforehand actually was supplying about 5V. A standard feature of this trainer-style airplane, an external on/off switch was between the power source and the output rail. The 50A ESC used in the RC airplane did not have a BEC. A previous test without APM 2.0 proved the servos, ESC, motor, RX and TX were all working as designed.
With the APM installed in the airframe and connectors attached as per the wiki, I turned on the transmitter and then the power source to the APM. In manual mode, all controls responded to inputs as was expected. I switched CH5 to "stabilize," and servos and control surfaces responded to yaw, pitch and roll, albeit in the reverse, and potentially with not enough input (I chalked this up to a calibration issue I hoped to resolve after shutting everything down).
I switched CH5 on the transmitter (CH8 on the APM 2.0 input rail) back to "manual," and then turned off the transmitter. The motor turned quickly as if responding to input, then stopped, and smoke rolled out of the electronics bay, followed promptly by a fire coming from the ESC. This fire got approximately 6 inches tall from the ESC and began to burn away the aircraft's external coating and lighted the wood frame of the fuselage. I yelled to my wife to run and get the fire extinguisher, which she ripped off the wall and passed to me, and I applied the extinguisher directly to the electronics bay. This put out the flame and saved the house, and I carried out the airplane to the back yard for safety. By this time, there was no more smoke.
An inspection after the fire showed that a section of external coating was burned, as was some of the balsa frame. Several wires leading to the ESC (two motor wires and battery positive wire) were melted right through. Although the LiPo batteries were mounted next to where the fire took place, they appeared at least on visual inspection to be unharmed by the fire. APM and receiver seemed unharmed from a visual inspection, but I'm not certain at this time whether the chemical flame retardant from the extinguisher has altered the function of those electronics.
At this point, I'm real hesitant to test this particular APM 2.0 on a new ESC to investigate the problem, and I'm really concerned about this as a safety issue. Does anyone know of a condition whereby the autopilot would supply max current to the ESC if the connection between TX and RX is severed? Thank you.
Specs:
TX: Futaba 6EXA 6ch
RX: Futaba R168DF
ESC: Hobbico SS-50D 50A (w/o BEC)
Motor: RimFireTM 42-50-800 Outrunner Brushless Motor
More pics:
Left side of aircraft, showing damage to external coating.
Electronics bay shortly after the event. In front are dual LiPo batteries, wrapped in protective foam. The 4.8v RX/APM power source can be seen hanging over the edge of the aircraft on the right of the photo. Yellow residue is flame supressant.
A shot of the APM and RX upon extraction. Both were coated in a thick layer of frame suppressing powder, until some of it was vacuumed off.
Comments
It defiantly looks like your electrolytic capacitors failed. By design the are meant to handle loots of current for a short time. An internal short cause a runaway short, like a fire until some thing failed, looks like the red lead to the cap.
Capacitors are like battery cells . They are two long strips of metal with an insulator (paper) layered in between. This long sandwich is then rolled up into the shape you see as a cylinder. When the paper between the sheets fails it just perpetuates until the "fuel" is consumed, or the circuit broken. The energy is coming from the battery which has quite a lot, and if that ignites you really have a problem. Electrolytics spew out a lot of smoke and flames until they burn clear.
The fuse should be more than your motor can normally draw. Maybe 150 % of max. What ever it takes, fuses prevent fires and also cut power, its a trade off, but if you have a short then the fuse will blow, and you glide in. Also handy in the remote instance of a catastrophic crash:)
I he'd the same, on a walkera Heli, just changed the battery for another flight,
just after a few sec in to the flight the heli caught fire. The ESC was completely burnd.
The APM is surely not at fault! It’s the ESC or motor or wiring.
@Bjorn - Thanks for the video and the info about the capacitor. I think you're going to love this picture:
Also, thanks for the note about the ammonium phosphate. The powder was indeed yellow, and the fire extinguisher was at least from 1989. I'm going to pay very close attention to how the APM will behave from now on.
@Vernon - thank you for the info -- that raw sensor screen is very handy. I checked my system and sure enough, when the TX drops out, so too does the throttle out signal from the APM. I'm now 100 percent this was exclusively an ESC issue. I'm going to find a new ESC with a BEC with an appropriate voltage now.
@Greg - That fuse sounds like a great idea, one that I wouldn't mind integrating into my drones. Would this fuse be soldered right after the connection to the battery (for example, on this setup, after the deans plug on the positive [red] wire, leading to the ESC)? How would I go about choosing the correct amperage for the fuse?
A ESC bug, it switch mofset a bad time and create a short. I fried a motor with buggy ESC (caused by to low power on signal wire). So it's highly unlikely in this situation, I think 2 wire were a little exposed or maybe a soldering of the three wire on the ESC may have a defect.
Oh,, and if the extinguisher dust is pale yellowish it is probably ammonium phosphate, which is corrosive. Also the white powders used in extinguishers (Sodium or potassium carbonate) are rather corrosive and may easily ruin your equipment. Don't take my word for it as I am not an expert on this, but I think the electronic components that have been smothered by it should be considered done for. Even if they seemed to work, corrosion can progress and they could fail later.
Obviously you have had a short from either motor or ESC, The ESC is fried. Wires with melted plastic was the path. Seems to have burned until the fault cleared it's self(burned though). The weakest link was the one that burned. It could have been an automotive style 30 amp fuse. This was an equipment malfunction on the power end. It is impossible to send a signal to an ESC and have it fry. If it is possible it's junk and you don't want it. I use a power fuse for the ESC on the outside of the plane and take power before this for the electronics. When the fuse is pulled the motor can't run, and if things fail, well that's why God invented fuses ;)
Sorry, here is a better failsafe page. It was linked but here is the direct link. http://diydrones.com/profiles/blogs/spektrum-dx8-and-ar8000-failsaf...
Sorry, I should have been more clear here. The APM does send a signal, that signal is equal to the range that it saw on the input. Further, you can view raw data in mission planner in the lower left corner and "see" exactly what the APM is outputing. To do so, connect with mission planner, and choose in the lower left hand corner "raw sensor view" and a new window will open. That window has a tab for "radio" and you will see the inputs on the left and the ouputs on the right. I fly multirotors but Arduplane is exactly the same in that screen.
Shutting down the receiver can cause the APM to go into failsafe, but again, the normal RC PWM range of signal is sent. Further info on APM 2.0 failsafe http://code.google.com/p/ardupilot-mega/wiki/Failsafe
Most likely, it was a wiring issue, bullet connectors or even a tiny scrape in the insulation. ANy movement could have caused the wires to touch. Testing the plane before using RC before the APM was installed shouldn't make you question the APM, the fault was in the high voltage and current wiring of the ESC.
The ESC has a microprossor and firmware that reads the input signal. Any signal outside that range is ignored by the ESC. The APM can only send out a digital signal, either 5 volts source or ground, and PWM it within a certain range. Even if the APM sent a signal out of range, then the ESC will ignore it as a non-valid signal. That's just the basics of how it works.
Further, your plan to power with a 4.8 volt nicad pack is bad. The APM will brown out, there is a reason why the pack is rated at 4.8 volts. You need a proper BEC and a battery higher than 4.8 volts. This alone would cause you to crash later on. You cannot depend on the open circuit voltage of the battery. Please follow the directions to the letter for powering the APM. You cannot roll your own here, it needs to be exact. Please re-read the directions here http://code.google.com/p/ardupilot-mega/wiki/APM2board Specifically powering the APM.
There's a sizeable electrolytic capacitor in ESC's. It looks like a small drum, usually sitting between the feeding wires. From the pictures it looks like this capacitor burnt up and torched the hole in the fuselage side.
Here's a video that shows how these things can fire if overloaded:
https://www.youtube.com/watch?v=M8nT97CrgcU
As suggested, you should check if you fed the ESC with too high tension. I agree that it is extremely unlikely that the APM can have caused this.
@Harry - A short in the motor wasn't something I hadn't considered, but I need to inspect the motor anyway. The bullet connectors actually appeared fine - this RTF came with the connectors already connected, and the connection was shrink-wrapped. It's actually a pretty nice package. But again, the motor was still at the time the TX was turned off, and I don't think the motor would turn at all unless it had a current coming from the ESC.
Oh, and to clarify my last comment, the APM actually appears to be working after all of this -- that's how I was able to connect it to the PC and the RX and check the calibration again. I suppose what I really need to do from here is connect it to a new ESC and a different motor and see if I can get it to replicate this behavior. I might just be able to find some time tonight to do that. Outside, of course.
-
1
-
2
-
3
of 3 Next