I was thinking today what would happen during a flight mission if my airplane with a nitro engine quit. It would seem the APM would just keep giving it full up elevator leading to a nasty stall thus leading to a not so nice ending. So wouldn't it be possible for the APM to detect engine vibrations? If so it could be set with a number of parameters leading to a better chance of a better ending. To confirm it's really a dead stick it could do full throttle for 2 seconds if no vibration then a dead stick situation would be confirmed. After that it could close the throttle to do the best job possible of keeping dirt out of the engine on a hard landing. But more importantly it could be set to simply fly a gradual circle nose slightly down until the last 10 feet of altitude then level flight until landed. Better than nothing. But I don't know if that would be something the gyros could detect? Engine vibration that is.
You need to be a member of diydrones to add comments!
Replies
Hi Richard,
I fly a lot of nitro and gas planes so I had the same idea last year. I was aiming to detect the engine cutoff fast enough that the plane could glide home, plus give a warning on the ground station fast enough that the pilot could land it on the runway.
I found it was a bit harder problem than I expected - I could detect it, but not quickly, and the plane had already lost a lot of energy before it would trigger. If I set the trigger level too fine then it would give false positives when the throttle changed a lot.
The problem is the amount of filtering we're doing of the accels and gyros. On the APM2 we filter with a cutoff of 10Hz, and that filter is done inside the MPU6k. That means we've lost a lot of the info we need to detect engine cutoff before the data gets to the APM code. I think it will be much more feasible on the PX4 where we do the filtering in software, and we could ask for the filter to also give a measure of the high frequency energy. For example, we could use a high-pass filter above 1KHz and measure the total energy at that frequency and above.
I haven't had another go at it since moving to the PX4, plus the latest engine we've bought has a tachometer builtin (which we haven't hooked up yet), so I probably won't work on this. If you have a go at it then please post your results!
btw, with TECS the plane should try to keep the speed above the stall speed, so it hopefully won't stall anyway, but it certainly won't do as good a landing as it could if it knew the engine was off.
Cheers, Tridge
The plane would be prevented from stall based on airspeed sensor or groundspeed. The plane will start to lose altitude to maintain speed in this case.
Wouldn't it be easier to use an optical sensor on one of the unused inputs to measure engine RPM? It would only take a very simple and inexpensive light sensor right behind the propeller, detecting the light-dark-light as the prop revolves. Engine stalls, tachometer detects engine out, flight controller handles appropriately.
Just my 2 cents.