Hi Arduplane Users,
Seeking some help if anyone has experienced this problem before with their APM.
We are currently using APM 2.5 on a Skywalker 1880. Its flashed with 2.40 firmware as we had some custom code changes to trigger camera and also log headings.
Th problem seems to happen randomly..we can fly the plane for like a total of 100km distance traveled in total at a crash will happen on the 101km..or it can happen on the very first mission. Initially we thought it was a mechanical failure but since most of the time we are like 2-3km away from where the crashed happened we couldnt really determine. But recently we are closer to where the plane is flying and this is what it will do:
1) As its banking or on a straight path..throttles up to almost full
2) Dives straight to the ground.
3)Any attempts to switch to stabilize and manual is not possible
All this happened only in Auto mode except one occasion in stabilize it dived but we had the airspeed sensor on for that incident so that may have caused it. Other than that the airspeed was not enabled.
I have some logs but im not sure it will be helpful as we only log certain parameters and at 4 secs apart.
Some of my setup info:
Plane: Skywalker 1880
APM: APM 2.5 with UBlox GPS Lea 6 & 3DR 433MHZ Telem Set
Motor: OS Motor OMA3820-1200/Turnigy G10 1200 KV
ESC : Tahmazo 60A/Turnigy 60 A (Red wire is disconnected as we use seperate BEC)
Servos: 4 X Emax ES08MA
Tx & RX: Futaba 12FG with 7 chn RX / JR 9303 with AR7000
UBEC: Hobbywing 5a, MAX 7.5A switching at 5V
Battery: 2 x 14.8v 3300mAH parralled to ESC and Balancer connected to BEC to supply 5v to APM
Others: PRISM IR Trigerring device, Airspeed Sensor(disabled after one of the crashes as we suspected it may have contributed)
We are migrating to the 2.67 code after successfully porting our custom codes as we felt that all the crashes exhibit the same behaviour and may be down to some bugs in the 2.4 firmware. But we just want to make sure that its not any of of the other electronics .
I could post the TLogs or onboard logs when i get the data from my work laptop tommorrow. These crashes are driving us nuts as we thought we solve the problem and happens again after a while.
Any comments and suggestions would be much appreciated.
Cheers and have a good weekend,
Khairul
Replies
Hi Khairul,
I have same problem with my SW1880, bank and dive with full throttle. My SW1880 a bit nose heavy, AUW 2,7kg. My guess SW1880 hate to bank more than 45 degree to turn with a bit nose heavy caused nose drop down. I set Roll Limit 45degree, but sometimes it can do more than 45degree (i dont know why perhaps wind disturbances?) . It caused lost altitude due to nose drops and APM try recovery altitude with adding throttle to 75% (throttle max limit 75%) but elevator not has enough time to recovery due to weight and throttling . So far have never had this problem with SW1680. Just my thought. I will try again with reducing limit roll to 30 degree and re-adjust the CG. Anyway, do you solve your problem?
Hello All,
Had 2 flights today.1st flight was on manual and stabilize just to check balance and trim. plane was perfect.
2nd flight was an Auto Mission. Just to point out, its a brand new airframe, that was its 4th Flight , new Axi motors, Castle Escs, APM 1 with JP1 unsoldered so its receiving its own power and today we had new 3DR Radios on. Plane was properly balanced and there were no trims on the radio. The flight the day before was perfect on Auto but had no logs as only got my 3DR radios today
So what happen today ? Again same prob , if you played back the TLOG attached, after it passed WP1, it started speeding up and we can see it slowly going down..and since we are so sensitive to the sound i quickly switched to stabilize and recovered flew around a while in that mode then tried Auto again...after that it was fine.
Its a very strange occurence..im sure that if i never took control that plane would've plummeted to the ground. VCC in was stable at 5096 throughout.
Could the 3DR cause interference? Ive attached a Graph of Radio and theres 2 lines thats gradually going up..is that fine?
Khairul
Radio Graph.jpg
Skywlker Condor 16 Dec Flight 2.tlog
I would still like to know why there is a massive glitch on the aileron and rudder channels at 74.5 secs, and why the pitch servo output PWM gets to 977ms just as manual is engaged.
977ms is like something is wrong. Also have you noticed these glitches on other crash logs?
analyze of parameters extracted from 20121212 tlog
PTCH2SRV_P is too big.
I'd suggest the HIL with Flightgear, which is a 'relatively' small download and I corrected the errors in the Wiki so it should be straight forward to get to work. PM me if you have any problems.
I do use the HIL quite a bit and it does use the PID's but the only plane I've been able to use in the Flightgear Sim is the Cessna and it's reactions are way different to a RC plane. The major difference was setting the Nav Pitch Alt 'D' value to 5.000 rather than 0.000.
One of my first APM equipped planes was very neutral at full throttle i.e. it didn't climb, which was quite a problem! My current one climbs too much with throttle which brings its own set of problems. All versions of ArduPlane I suspect, assume a gentle climb with throttle.
Khairul, what do you think, it it possibly this?
I am having a bad time with my PIDs, mostly in pitch power and have had some high power dives.
Graham, I think you use HIL? I dont know anything about it. Can you advise if the PIDs are used in HIL or just bypassed. I dont expect the actual figures to be relevant for the SIM plane but it may help to learn how they affect the flying.
Thanks
Yup, flexing bits and not enough elevator authority and/or headway at high speed?
Graham, looking at the first part of the flight the elevator servo signal is fine, with lower PWM producing pitch up and it tries correctly (if a bit limpley) in auto. As you say that is not is not nesscarly control movement, which is my point about flexing bits.
Not sure how significant the aileron/rudder spike is just before it goes pear shaped.
Here's the last few seconds of your log posted on the Bixler crash thread: the altitude (green) starts to decrease at the point you switch out of auto into manual (light blue).
There are a few oddities I see:
1) The yellow line is your elevator stick input which hits a ceiling? (also it's mid value seems very low, it should be ~1500, the ceiling it hit is 1400)
2) The dark blue is the output to the elevator servo which doesn't seem to move when the stick (yellow) is being pulled - strange
3) The red line is your throttle-in which follows throttle-out closely until you switch to manual then it stays level while the pink line (throttle-out) increases (by itself???)
I can't explain these, it's just what I'm seeing.
-
1
-
2
of 2 Next