Hi all,
Any ideas what might cause something like this? I have a 250 quad that I've been flying around for a bunch of months at the same field. I haven't made any changes (software or hardware) and suddenly the quad crashes every time I fly it. I took a look at the log and it looks like the commanded roll, pitch, yaw, and throttle all go to 0 and it just falls, as you'd expect. I've attached the log, you can see the commands drop around 2153.0. Screenshot also attached.
I don't see a radio failsafe, but it seems like there's a GPS error that happens every time (but according to the plot I see it after the crash - is there a time delay in displaying those red flags?). Could be due to the crash though as it lays sideways or whatever on the ground.
I've repeated this 4 times in a row already. The strange thing is I can fly inside just fine. Could it be 2.4 GHz interference? I'm using an HK Orange DSM RX and TX.
Thanks for any ideas
Replies
The most likely cause, given the description, is age. If a copter falls out of the sky, pay attention to how it falls (like a rock or in a bad glide).
- Falls Like a Rock - use may have caused an intermittent electrical issue possibly exasperated by heat. The difference from indoors to outdoors might be how the ship is flown (if it's flown harder outside, then there is more heat and vibration). Pull and move all connections while on but w/o rotors. Sometimes this will ID the issue. Look hard at any solder connection. They can fatigue.
- Falls Like a Bad Glide - the battery is getting older and overtaxed beyond it's C rating. The chemistry collapses causing loss of power. This can happen without heavy throttle like during a fast-sharp turn.
Thanks guys. It falls like a rock - it's responding to the 0 throttle command that it's seeing, so all the motors turn off. I'm going to replace the wiring and receiver tomorrow and see if that does anything.
Good detective work on the zero throttle command.
You might only have to replace the main connection to the power module.
- if it flips and drops then it's a connection to one motor
- if all four motor stop and it drops like a rock, then it's most likely one of the main power supply connections (probably the one going to/from the 3DR power module).
It is common for shrink wrap around a solder joint to hold the wires together until heat or vibration separate the wire ends that have become broken from fatigue. It is unlikely to be the solder connection on the power module. It is more likely the solder joint on the ends of one of the wires going to/from the power module or power distribution board (if you have one).
Don't feel bad. During the setting of the world record for flight (about 15km of flight over 1.5 hours), near the end when the ship was brought into a hover about 3' off the ground to round out the trip, it prematurely dropped out of the sky upright onto a cushion grass. A main connection had come loose. That would have ended with a different result if it has occurred just moments before when racing around at 50 meters over an orchard of trees.
Thanks Forrest. Yep, all 4 motors stop. Here's more detail on my system:
APM is powered by a regulator on one of the ESCs. I'm not using a power module. I think if there was a bad solder joint, that would take down the ESC regulator and then the APM, and it wouldn't have kept recording data - right? The data kept recording past the time of ground contact.
That leads me to believe it's somewhere between the TX -> RX -> APM wiring or hardware instead of the power system. I've been jiggling wires around on my bench trying to reproduce the issue, but no luck. I can disconnect the RX with hover throttle and it goes to radio failsafe, as I'd expect. Pretty stumped here.
I'm going to try to move the experiment outside where it's colder - maybe that's related?
Are these clues correct?
- APM keeps recording in a single file (not two). [so power remains at the APM]
- APM responds to motor disarm after the fall. [so connection remains or reconnects between the APM-Rx]
- All four motors fail at once. [so a gross power failure or a command to cut power]
If you still can't find it, then change the ship response to a Tx Loss. Make Tx Loss do something like hold in place (do nothing) and if the ESC are the plane programable type versus multicopter type, then program the ESC to maintain the last command. If it is all ready set this way, then we know it's not a Tx loss. You can test this by momentarily turning off the Tx (but be careful). This test is critical as it greatly decreases the possibility of causes. It will then be narrowed down to major loss of power to all four ESCs at once (i'm assuming that the APM is powered off of the PDB by wires not via an ESC BEC.
I assume that if you do not have the power module then you have a power distribution board. I'd also look there. If the board had taken some hits, there might be hairline fractures or chips with bad connections (I don't use PDBs so don't know if this is unlikely or common).
- Epoxy contracts more in cold than silver or gold
- silver and gold contract more than glass and carbon fiber
I think I found the cause:
Inside the building (warmer, no GPS, less interference) when I turn off the TX, the throttle PWM is 1003. Doing the exact same outside (colder, yes GPS, more RF interference), the throttle PWM is 1018. I change nothing else between the two tests. The failsafe PWM in mission planner was set to 1010, so in one case it triggered, in the other case it didnt. Maybe a sensitive RX? Or RF sensitive wiring?
well done!
so what was the setting for the failsafe mode of the copter? it sounds like this needs to be checked in the APM and possibly ESC. then test it as mentioned before to ensure that your ship doesn't just fall out of the sky.
An example of this is when i fly inside i need to change the failsafe from go to 100m and return home to do nothing (or do it very slowly). otherwise it will immediately try to fly through the ceiling if something unusual happens.
Sorry, I was away and didn't see your reply.
The failsafe was set to RTL. Yes, agreed, maybe at least some check with an override to make sure you can't set the command to 0 while in the air? Maybe.
Yeah, I just set mine to LAND when I fly inside.
Check for intermittent damage caused but crashing would be my first guess.