Disappearing Airplane - Where did it go?

Where did this airplane end up? Not where the log says, so where is it?

Thursday morning, we launched an Airfield Giant on a simple mission. The goal was to test downlink video after the VTX antenna had been moved back from the nose and on top of the wing attach saddle. The video transmitter is on 1280mhz, 1 watt output and using an IBCrazy 3-loop omni. Video is from a GoPro through a MinimOSD v1.1. We were just testing to see if it worked as well or better than up on the nose.

We took off and climbed to about 350 feet while just flying randomly over the field. GPS shows 9 satellites and HDOP at 1.2. We switched into AUTO to execute a small square pattern. GPS (orange) locked onto WP1 and the plane started towards that point. In a few seconds it was apparent the plane was diverting in a shallow right turn. It continued that turn and flew away to the south (about 90 deg from the intended course. All the time, the APM stayed on WP1. We went to RTL from the console and backed that up with the RC transmitter. MP showed this, the orange line swung back to HOME. I also gave an MP command to climb to 1000 feet so that we wouldn’t lose 915mkhz data or DragonLink control. The plane started a gradual turn back to the field from about 2 miles out. RTL was in effect and we used small turns to help the plane stay close to a return course. Another Change ALT was sent for descent to 400ft. When the screen showed the plane making a circle overhead, I called that out to our safety pilot, and he went to manual despite not having visual contact.

At that point, the throttle was at idle, and the APM telemetry data shows the plane gliding to a stop about 550 feet to the east of “HOME”. But, there was no plane. Two people were looking overhead for it and neither saw it at any time. We went to the landing area and found nothing. No plane, no parts, not a thing. This is an active field and we are well practiced at recovering planes. Despite an extensive search using feet, cars and a helicopter, nothing has been found.

So, my request and challenge to this group is to see if there is anything in this telemetry log file that shows what really happened to the plane. The GPS shows good 3D data with an HDOP at about 1.23. The wreck should be right at the spot or close by. Did the APM record bad data? Was the GPS system being spoofed? Sunspots? Martians?

I guess my question is… what did you see or record on video? Did you lose video feed?

secondly, I would guess someone stole your plane… if the landing site was out of your view

Was this firmware pre 2.77 by chance? We had a GPS failure that was not reported to mission planner. It looked much like yours. Tridge patched this in 2.77 however.
If you search my posts you’ll find the X8 crash in referring to.

I have had a look at the log and it looks like something goes very wrong around 19:55:19 (25%). At this point airspeed and ground speed information is lost, HDOP, Sats and groundcourse all freeze in their current status and altasl goes to 0 for the remainder of the flight. The key for me is groundcourse, there is no way that should remain exactly identical (128.89) for such a long period. Mission Planner may not have reported a GPS failure as HDOP and Sats were showing normal (although frozen data).

Scenario 1
Without knowing the major details of the dead reckoning code I think you would find that the aircraft may have been attempting to calculate location based on this and not GPS inputs. If this was the case I would pay very close attention to the wind that day which in the early flight looks to be about 2-4m/s from the north. This could have caused the plane to come down south of where it was reporting if dead reckoning was in use. One other consideration is the climb to higher altitude - this could have exposed it to stronger/different wind - is there data from your local weather service available on conditions that day?

Scenario 2
Alternatively if you review the off-course flight, the roll values are very low which suggests that the aircraft may not have been turning to the right as seen on the map (did anyone actually observe a clear right turn due south visually?). If the apparent right turn was due to drifting gyro/mag data it is possible that it continued on to the southeast in a relatively straight line possibly drifting to the south also with the wind as per above. If you use a stopwatch or analyse the track information to work out how far it traveled along the ground after loosing GPS data this should give you the approximate range at which to start the search. In the absence of someone who can give you a real calculated search pattern I would project a straight line on the last GPS track (128 degrees) the distance you think it would have flown at those throttle settings and start from there southwards to account for wind drift.

I would also suggest perhaps reviewing the log using the gyro yaw data with a stopwatch and manually plotting the off-course flight (cover up the map view and ignore it) The throttle seems relatively constant for most of the off-course flight (except in climb) and if you know how fast the aircraft typically travels at this throttle you should be able to roughly re create the flight path on a paper map to try and work out where it might have ended.

I hope some of that is helpful to you. One of the developers may be able to give you better insight into how the software handles these situations that might help identify scenario 1 or 2 or others as the most likely candidate?