absolutely brilliant your analisis above, thanks a lot for contribute to this knowledge.
I have experienced many different behaviors thinking that something is wrong in multicopters and plane but almost always the solution turns out to be a summation of conditions and parameters set correctly . This discipline takes time and patience.
373 01-01-1970 05-30-00.bin (2.0 MB)
Hi, Crashed Skywalker Arduplane 3.40 when switched to RTL while ythe plane was in guided mode āGo to Hereā
I can see that there is no Airplane firmware upgrade for APM beyond 3.40 and I can see that the bug was resolved in Arduplane 3.52. Can someone help for a patch for APM Arduplane 3.40 ?
@Rana As you may have seen in more than one announcement, the orginal APM isnāt supported anymore, so unfortunately you shouldnāt really expect firmware updates for your flight controller anymore.
The original APM has been obsolete for more than 2 years now, my recommendation if you care about your airframes is to stick to a pixhawk clone, or better from now on.
Hi,
I have many many APMās and want to use them, so if either Tridge or someone else can share the patch to take care the issue of crash in guided then I could make use of these APMās
I have many pixhawk too but want to use the APM also.
@Rana, I had a look at your log (ā373 01-01-1970 05-30-00.binā) and it doesnāt look anything like the issue that Marc previously saw.
The first thing I notice in your log is that no roll or pitch tuning has been done - it is just using the defaults. It looks like the gains are way too low. Because of the very low gains it takes a long time to build up enough elevator to start to pull up when it is descending. The main thing you need to do is run an autotune, or even just manually increase the P and I gains or both roll and pitch. It takes nearly six seconds to build up enough pitch integrator to start to pull up, and that is much too long. You probably need about 5x the integrator gain, and 3x the P gain at a rough guess.
I am happy to do a new bug fix only release for the APM2 if a really serious bug is found and I have a log that shows it is needed, but this is not an example of that sort of problem.
Cheers, Tridge
If gains are suspected culprit then the same skywalker plane with which I have done more than 200 successful auto mission flights with APM2 only should have not been.
I mean I never initiated RTL in guided mode in any of my flights with APM prior to 3.4.
Prior to this crash, the this skywalker plane with same APM2 has done few successful flights in auto mode without any issue and that is too with default parameters. Below is link to some of very successful long missions with Skywalker with default parameters in APM2.
Below is the telemetry log fileās video clip of small portion of this flight in which crash happened.
That is one of the problems with the ArduPilot controllers - it does fly OK in many situations even with terrible tuning. That is nice for people who donāt want to bother tuning, but it means they get complacent and donāt tune. Then when the vehicle is in a bit different situation and has to recover attitude quickly it will not recover.
I flew a skywalker with an APM2 as my main test plane for a couple of years. My parameters were:
PTCH2SRV_D 0.07
PTCH2SRV_I 0.3
PTCH2SRV_IMAX 3000
PTCH2SRV_P 1.8
PTCH2SRV_RLL 1.1
RLL2SRV_D 0.1
RLL2SRV_I 0.3
RLL2SRV_IMAX 3000
RLL2SRV_P 2.2
note that they are a very long way from your parameters! The control surfaces on a skywalker are really quite small, so it needs a lot of gain to fly well.
Donāt just use my parameters btw, better to run an autotune. The right parameters depend on several factors that are dependent on the specific airframe (eg. which hole in the control horn you use).
Cheers, Tridge
I will only do a new APM2 release if a serious bug is reported that warrants the work to do that release. So far I havenāt seen any bug reports that are sufficiently serious to do a new release. I looked into your log as I thought perhaps that would be one, but it turn out it is just tuning.
Cheers, Tridge
Hi Tridge, once again many many thanks for you time and support.
Also thanks for sharing your parameters.
I would try to use your parameters and then check in in stabilized mode at ground it all control surfaces are behaving properly.
Only then would do a small test flight at around 200mt alt then try to see if the model is doing well in stabilized and auto.
If it does well then at around 200mt altitude, I will put the plane in guided mode preferably not much away from me and then will initiate RTL and see.
Thanks for confirming that no serious bug reported so far in AP3.40 for APM2
Once again thanks a lot for your wonderful support bro.
@marcmerlin Are the crow flaps on a pass through channel? Or does guided mode turn them off?
Iām quite late to this, but the clue seems to be āturn to bleed off altitude with crow flaps enabledā. If your crow flaps are any good and on a passthrough channel, then no way can you gain altitude even at max throttle with crow flaps on. Once the Autopilot takes over it needs to take control of the crow flaps as well and turn them off.