The AUTO flight mode brought my plane down :-(

Hello everyone!

I have been setting up a pixhawk-based plane.

The whole process was going well and the plane was flying. However, when i first tested the AUTO mode, the plane came immediately down.

Let me recapitulate the chronology of events.

I’m using the PIxhawk with Plane 3.7.1. I setup the parameters using data I had from a similar plane.
My first test flight went well. I used the modes FBWA, RTL and AUTOTUNE.

In my second flight I did use specifically the AUTOTUNE function to calibrate the flight parameters, as explained in manual. The behavior of the plane improved respectively, the curves in Mission planner looked liked they should.

In my third flight, I did use AUTOTUNE again, since during my previous flight the wind was to strong and I had to interrupt the flying. So, the first part of the flight was dedicated to do some additional autotune flying hopping to improve the flight parameters. Then I landed the plane and substituted the mode AUTOTUNE with AUTO in the settings (without disconnecting the plane from the battery once since the landing) , since I wanted to test mapping and other AUTO flight functions.

I started the flight again, went from FBWA to RTL to FBWA and then finally to AUTO. With this last mode switching the plane just went into descent immediately. The 1 to 2 seconds I had to react to go back to RTL where not really enough to save the plane, the velocity was to big and parts of the (foam) plane, I suppose the elevator, broke apart. There was nothing to do. It crashed deadly. :frowning:

Well, I did have a look at the log files, but I still miss the experience to identify the problem. The only thing that calls my attention (but I might be misinterpreting the thing) are the ATT.(Des)Pitch values:

.

As far as I see it, although the ATT.DesPitch wanted to bring the plane up, it just went down (ATT.Pitch). It almost looks symmetrically.

I would be very glad if someone has the time to help me on this one. Maybe the error is so obvious, I just do not see it due to lack of experience.

You might download the log file here: https://drive.google.com/a/openforests.com/file/d/0B4Hds_6TKh84RTNpMXFUZlFLeTA/view?usp=sharing

Please let me know if you need something else to helpe me.

Best regards and thanks in advance for any help

Hi Patribus,

If your plane was flying well in RTL, it sounds like it was an issue with the flight plan you had programmed in AUTO mode.

Can you upload the flight plan that you programmed before switching to Auto mode?

Its imperative to always use Relative altitude, If you use absolute altitude and do not set it up properly, the plane will dive to reach the absolute altitude above sea level, causing a dive into the ground. Mistakes in flight planning can be fatal in AUTO mode if not done properly.

Dear Jman841,
thank you for your response. As banal your suggestion seems, that might really be the point, although I really did not think about it. First reason, I never had problems with the planning, also not with altitude issue, so I did not think, that I could have caused the error myself when planning the mission. Secondly, and I think this is the most important: I really did not think that the plane would go into a kamikaze dive just because of the (extreme) height difference between its actual height and the destination height. The behavior I always observed until now, was that there is a smooth approach to the next (first) WP height. So, observing the vertical flight towards the ground really did not make me think, it could be the planning. I mean, if this is the reason, then it was a really really stupid mistake, on the other hand, it would mean, that I would not have to search for ages some bad parameter in the settings.

Well, to finalize: I did not store the mission plan, but I suppose, that when I open the mission planer again, the altitude setting should be the same as when I closed the software. Is this correct? I will check this as soon as possible

And a last question: So, should I be flying at 300 meters relative to TakeOffAltitude (TOA) and go over to AUTO, having my first WP at 100 m relative to TOA, will the plane also dive as crazy or go smoothly to that point? Because the problem of having lost my plane was the abrupt change in flight direction towards ground, which I could not emend again, since the steering correction made the ailerons break, so harsh was the maneuver.

Cheers

Dear Jman841,

I checked again the standard configuration of mission planner as I used it, and the altitude is set to be relative, so I fear the problem was not related with the mission plan. If only I new what I should be searching for… Some hints?
Cheers

Just to clarify one point:
Isn’t irrelevant if you set relative or absolute in the altitude interpretation as long as you plan your mission using the “Auto WP > Survey (Grid)” tool? As far as I see, he will calculate the relative or absolute flight altitude based on your initial settings relative or absolute.

Hi Patribus,

Trying to explain a recent crash, I have found your post and I am interested in your graph with Pitch and DesPitch values. They are similar to mine. And context also : mode “auto”, pixhawk 3.7.1.
But in my case, the crash happened in the midle of a mission, just when switching to another waypoint.

Did you find an explanation ?

Best regards.

I’m no expert, and I just had a look at the .bin, but at the moment of going into auto there is a large DesPitch spike, and from there the DesPitch and Pitch go their separate ways.
I am leaning more to a mechanical failure or a stall, and as the control surfaces don’t max out, more likely a stall.
If you study your RCout at the time of going into Auto, you can see the surfaces trying to compensate but not achieving the desired goal.
What was the air frame?
What flight speed did you have set for auto?
Could the surfaces have been over stressed by that spike at Auto sswitch time?