Issues with IMU data and tuning FBWA

Archive
1

Hi all,

Finally managed to get a maiden flight with my X8 yesterday.

APM 2.6 running 2.76, external GPS and compass.

The aircraft flies well in manual mode, the CoG is correct and everything works as it should. It is appropriately powered with a thrust to weight ratio of about 0.8.

In manual it requires about 30% more elevon throw than one would expect with a flying wing, but I have found no cause for this and it may well be airframe specific.

Issue 1: Both the flash and the telemetry logs show exaggerated attitude values. As a visual estimate, I would say that the aircraft never exceeded ± 45-50deg in the roll and ±30deg in pitch. The values shown in the logs do not diverge from reality in a linear way. The indicated attitude seems to be roughly correct until about 20deg and then starts diverging as roll increases. I can guarantee the aircraft never reached the 80deg roll maximum that is indicated.

Other logged data also seems suspect, but it is difficult for me to say, as observing other aspects of flight may not be as accurate as an observation of attitude.

I was not aware of any of this behavior until I reviewed the logs.

Issue 2:

When attempting to tune FBWA, the aircraft was completely unresponsive to roll inputs at the baseline RLL2SRV_P value of 0.4. This was raised in increments of 0.1 at first with minimal changes observed. Increasing by increments of 0.2 brought observable results and the aircraft was responding in a satisfactory way at a RLL2SRV_P value of 1.6, Both D and I values were left at the baseline of 0 for the time being.

Pitch tuning was not successful. I got up to a PTCH2SRV_P of 1.4 with only a very small observed response to input and decided not to go any higher until I have advice on this.

I guess my questions are as follows:

What are the possible causes and remedies of issue 1?

Is it possible that issue 1 is related to the consistently high GPS HDOP values during the flight (cca 2.0)

Is it possible that issue 2 is related to issue 1?

What are the maximum “healthy” values for RLL2SRV_P and PTCH2SRV_P?

If my P values are too high already, is there some sort of scaling parameter that can be applied?

P.S. Please note that the actual flight only starts around line 105800 of the flash log. It seems I’m having trouble uploading the flash file. Alternative link http://depositfiles.com/files/pjwr3g755

Many thanks,

Hans

2

Hans,

Its been a while since your post - have you made progress?

Are you using Elevon Mixing in the AP?
I have found with Elevon Mixing that in FBW-A with a Max Pitch setting of 20 there is not enough elevator response to even tune as we have insufficient elevator control to even keep the plane in the air in FBW-A. In Manual and Stabilize everything is fine. Not sure if this is the way but fixed that by increasing Max Pitch to 30. This immediately gives more elevator throw, not quite sure why as a limitation to the attitude should not directly limit the available throws?

Also curious to know high the RLL2SRV_P and PTCH2SRV_P can be set safely. I also find values of 1.6 upwards for roll and approx 2.1 for pitch working well on the X8.

PD

3

I had the same issue trying to use FBW with Elevon Mixing. It’s been this way for at least 6 months now, but I haven’t seen any resolution other than using extremely high max/min pitch numbers and higher P’s than one needs for auto.

4

Hi all,

Yes, I am using the APM for elevon mixing. I seem to have got the roll loop tuned ok in FBW-A, the pitch loop is still not tuned correctly.

The problem with increasing the max pitch values is that I can’t really do that because of airframe limitations. The X8 has a fairly thin margin between overspeeding the aircraft and stalling it. To make matters worse, the aircraft is slippery and will accelerate at anything more than about -10 degrees.

I’m still trying to figure out if there is some sort of scaling parameter that mighy be able to help with this.

I was kind of hoping Tridge might be able to provide an answer, but he seems to have missed this thread and I’m wondering if it’s bad manners to PM him with a request for help.

Getting frustrated here.

Hans

5

Trent from MyGeekShow recommended to me a few things which were recommended to him by Tridge, one of the things was to mechanically increase the elevator throws then use expo or rates to tame manual flight.
I haven’t had noticeable results yet but I haven’t increased them by a large amount, will try again tomorrow.

6

I noticed that there is an Elevon Gain parameter, has anyone tried that yet?

7

[quote=“Graham Dyer”]Trent from MyGeekShow recommended to me a few things which were recommended to him by Tridge, one of the things was to mechanically increase the elevator throws then use expo or rates to tame manual flight.
I haven’t had noticeable results yet but I haven’t increased them by a large amount, will try again tomorrow.[/quote]

Knowing my particular build, I don’t believe I should be increasing my throws any further. I did a fair bit of flight testing before I even installed the APM, just to make sure I had a good baseline for future troubleshooting. The control deflections on my X8 are already more than I have seen on any other wing and increasing them further would probably make the plane aerodynamically unsound. The plane has been logged in a 3G pull up at 22m/s so I don’t think I could do with more pitch authority.

It just seems to me we should be applying a software fix to a what seems to be a software issue.

Interestingly enough, when in FBWA on the ground, the AP will apply full elevon up when commanded do so. This points me further in the direction of a software issue. I think we need a fix rather than a workaround…

Any of the devs reading this thread?

8

Have you tried this:

Mixing Gain (ArduPlane:MIXING_GAIN)
The gain for the Vtail and elevon output mixers. The default is 0.5, which ensures that the mixer doesn’t saturate, allowing both input channels to go to extremes while retaining control over the output. Hardware mixers often have a 1.0 gain, which gives more servo throw, but can saturate. If you don’t have enough throw on your servos with VTAIL_OUTPUT or ELEVON_OUTPUT enabled then you can raise the gain using MIXING_GAIN. The mixer allows outputs in the range 900 to 2100 microseconds.

Range: 0.5 1.2

9

I have noticed that one, but I might be very careful.in playing with that because it affects manual control as well and I’m a bit scared by the possibility of overdoing it with no possibility of reverting to manual.

10

I have noticed that one, but I might be very careful.in playing with that because it affects manual control as well and I’m a bit scared by the possibility of overdoing it with no possibility of reverting to manual.[/quote]

Tridge writes that it is set at a default of .5 although most hardware controllers are set at 1.0. This seems to me that it is set very conservatively, and could safely bd increased incrementally. You should see the effects immediately on the bench as the throws increase.
Please let us know if it helps.

11

The weather’s been horrible lately, so no flying, here’s the last available Tlog:

droneshare.com/view/bkeq2kb