Unusual fly away without loss of Typhoon H (thankfully!)

slw74 · Nov 3, 2022

Hey guys,
I want to call on your great expertise in helping me understand what happened yesterday! I was flying my favorite Typhoon H which I have been flying for several years. It's my go-to for flying CCC missions around construction sites at least weekly to monitor the construction progress. I have never encountered any problem with this drone; it's never had a crash or even a hard landing. The CCC is a large loop around the perimeter of the property, flown at 125 feet and lasting about 3 minutes and 30 seconds; I do a pass with video recording and a second pass taking 1 photo/sec for details. My batteries are in good shape and I can do both circuits and land before the battery gets to 14.5 volts. I have not had a first battery warning. I had an excellent day for flying yesterday with little wind and clear skies. There was good GPS lock on both the controller and the drone. Telemetry signals were good. The two pass flight completed as usual and the drone returned to the end point and hovered there which is fairly close to where I was standing. I ended the CCC and prepared to land. But, there was no response to the sticks, landing gear toggle, etc. Within 15-20 seconds the drone started a slow lateral flight, maintaining a constant altitude without any turning or variation in it's course. At this point it passed over me, still with no response to the controls. It was about to go out of my direct VLOS due to an intervening building. I could see where it was via the camera image. The voltage at this point was 14.3 volts and I got the first battery warning. As the drone was just out of my direct view I switched it to RTL, doubting whether it would have enough remaining battery to climb and return safely. Much to my surprise, it ascended and came back into view. As it started its descent, I regained controls which allowed me to adjust the landing point and lower the gear. It landed just after the second battery warning without incident. What a relief that was! At this point, I don't know what caused the unexpected behavior. I am concerned that this will happen again and I won't be so fortunate in recovering it. The controller monitor shows all switches, buttons and sticks to function normally. I am attaching the FlightLog zip file for your review of 11-2-2022. Any information that you can provide would be greatly appreciated.
Thanks in advance,
Stephen

WTFDproject · Nov 3, 2022

slw74 said:
....The two pass flight completed as usual and the drone returned to the end point and hovered there which is fairly close to where I was standing. I ended the CCC and prepared to land. But, there was no response to the sticks, landing gear toggle, etc. Within 15-20 seconds the drone started a slow lateral flight, maintaining a constant altitude without any turning or variation in it's course. At this point it passed over me, still with no response to the controls. It was about to go out of my direct VLOS due to an intervening building. .......

I was looking for issues in the data. Couple of things, but nothing any where near "smoking gun" quality. Then I looked at the flight path on Google Maps. The recorded flight path makes it look like the CCC either did not end properly, or it restarted (in reverse). It appears the drone was following the CCC route (backwards ?) until you sent the RTH signal. I don't have much to support this other than the flight path image, but is it possible the drone was simply still following (retracing?) the CCC route?

WTFDproject · Nov 3, 2022

Update. Please understand I don't use CCC, and only set one up once. I'm hardly an expert, so take this with some salt.
But at 13:09:36 the VALUE FOR CH0 (THROTTLE AND ON/OFF BUTTON) went to zero four times within a couple of seconds. That's sort of the same thing as pulling back on the throttle (except the values are greater.) That catches my eye for two reasons.
1.) It was the exact second the drone started flying the route in reverse.
2.) It matches pretty well with the instructions for how to reverse the route.

Starting to feel more confident it was not your run-of-the-mill flyaway. But I have to turn in for the night. Early appointment tomorrow.

I can see where you manipulated both left and right sticks while the drone was stationary over the intended landing point. I see NO response from the drone at that point. I don't know where (or if) the CCC status is recorded in the data. When/If you agree the drone was actually following the CCC route, the next step will be to figure out what should have happened vrs what actually happened while it was over the landing point.

slw74 · Nov 3, 2022

Tom,
I'm not sure what would cause it to do that, or more importantly how to avoid it, but as you note, it did seem to retrace the CCC in reverse. The RTL executed along the CCC path as demonstrated in your photo. Awesome analysis; I really appreciate it. Now perhaps we can see why it happened?!?

WTFDproject · Nov 3, 2022

slw74 said:
Tom,
I'm not sure what would cause it to do that, or more importantly how to avoid it, but as you note, it did seem to retrace the CCC in reverse. The RTL executed along the CCC path as demonstrated in your photo. Awesome analysis; I really appreciate it. Now perhaps we can see why it happened?!?

We were typing at the same time. See my last post.

slw74 · Nov 3, 2022

When I get home next week, I'll fly the course and then intentionally try to initiate a reverse CCC. If it responds in the same way, it would explain things. I'll give an update then.
Thanks,
Stephen

Vaklin · Nov 4, 2022

Pressing the red button is the same as a full-down throttle stick. Check the health of the button if you didn't press it.

I'm not sure is possible to go out from the CCC when the path is in execution.

Never used CCC, everything said by me is just speculation.

Pier2162 · Nov 4, 2022

Vaklin said:
Never used CCC, everything said by me is just speculation.

I don't even know what it is.

Vaklin · Nov 4, 2022

RTFM!

DoomMeister · Nov 4, 2022

Vaklin said:
I'm not sure is possible to go out from the CCC when the path is in execution.

It is possible by pressing “Exit” on the task menu or by switching flight mode (such as selecting Home).

Instructions for CCC start on page 35 of the enhanced user manual.

We must having been replying at the same time @Vaklin

phoonflyer · Nov 4, 2022

just be glad it didnt take off (been there) My go-to in situations like that, just to regain control, is to instantly switch off GPS. if its gonna fly away its gonna fly away but, if its got a false sense of positioning for whatever reason, Its worth a try. P.S. If you've been flying a H for years without something weird happening you were overdue

DoomMeister · Nov 4, 2022

I took a look at the flightlog and concur with the assessment of @WTFDproject that CCC was still active and the left stick down action reversed the direction the CCC was running. Most likely is that the touchscreen did not activate exiting the CCC when the "Exit" soft button was pressed. One of the ways to exit CCC is to change the position of the Mode switch and that happened when you changed it to Home (RTH).

The following is a chart compositing data from @h-elsner program Q500log2kml showing flight mode, errors, and positions of the control sticks during the flight.

The red line on the left is motors arming. Note that the value CH0 throttle drops to 0 just prior to that indicating the motor arm button is pressed. F_mode changes from 16 to 8 (arming) to 3 (Angle mode). Ch0 throttle then increases to launch and gain altitude.

Next the blue line indicates activation of the CCC indicated by f_mode changing to 26.

The magenta line indicates Ch0 increasing to start motion and speed of the CCC.

The green line shows completion and exiting of CCC after the first circuit is complete. Note the change in f_mode back to 3 (angle).

There is about a 35 second delay then the next blue and magenta lines mark the change to CCC and start of forward motion and speed.

The cyan line marks the point at which CCC was assumed to be exited followed by stick movements that resulted in perceived lack of control. The large Ch0 down at about 09:36 is likely where the CCC started running in reverse.

Lastly the purple line marks the change to Home mode causing an exit of CCC as shown by f_mode changing to 13 (RTH going home) then 14 (RTH landing) then to 3 (angle).

I'm also attaching the kml file for this flight and if you watch it in Google Earth Pro you can play it as a time lapse to see it in motion. Be sure to rename it dropping the .txt extension first (added so it could be uploaded here).

I hope this helps Steve, and if you experience this issue again a quick flip of the mode switch to Home or Smart then back to Angle will cancel the CCC and the sticks should work normally.

slw74 · Nov 6, 2022

Well, as usual, you guys are very knowledgeable and great detectives. In very little time you had this sorted out; amazing! I look forward to getting home to try the sequence of events again to confirm what seems to be the correct explanation of what actually took place (but this time with plenty of battery reserve). I guess many times folks lose a drone because of these unforeseen responses due to pilot error. "Thankfully" as it says in the title, mine made it home. I'll update once I get the test flight completed.

DoomMeister · Nov 6, 2022

We can not be positive in our conclusion, but seems to be the most plausible. My suggestion to you if this starts to occur again is to make the quick change on the Mode switch to Smart or Home and then back to Angle. That will exit CCC and allow you to regain control.

We are anxious to hear about your future results at this site.

h-elsner · Nov 6, 2022

Pitch and yaw sticks are not centered any more. I think they need some cleaning and stick calibration.

br HE

slw74 · Nov 13, 2022

OK...once I got home and the results of the hurricane subsided, I was able to fly the drone in order to best duplicate my past experience. The CCC completed as usual without any problems. I intentionally did not close the CCC program and depressed the left stick fully for about 4 seconds. The drone proceeded to fly the CCC in the reverse direction and was identical to my prior experience. So as usual, you guys were exactly right. Somehow on the original "flyaway" flight, I did not get the CCC closed using the screen tap on "Exit." And per h-elsner's note, I cleared the sticks with a little locally applied deoxit and stirring.
Thanks again to all of you,
Stephen