Replies to This Discussion

ModeratorPermalink Reply by Graham Dyer on February 21, 2012 at 11:30am

Thanks, will try removing the Rate_D. Props are balanced on a magnetic balancer, haven't got any softer wire/cables but will see what I can do

• ▶ Reply

Permalink Reply by John Hanson on February 21, 2012 at 9:14am

@Graham - I also have a X525 frame with the APM2 mounted with orings -

http://diydrones.com/forum/topics/share-your-drone-info-thread?id=7...

 at first, I had a lot of vibration from flimsy motor mounts but I upgraded them. The unit sounds very smooth now, holding it I can only feel relatively slight vibration. It helped, although it still seems subject to leaning. The board is so light I don't see how rubber can isolate it, maybe a weight on the board might help. Unfortunately, Denver weather is not letting me test.

• ▶ Reply

Permalink Reply by Dave on February 21, 2012 at 9:33am

Just had a test flight. Was ok but I find now that when travelling with a constant attitude (ie nose down to go forward) for any length of time the internal state get's a bit funky and loses its sense of what is level.

What happens is when I let go of my pitch stick it does not level, it overcorrect and pitches the other way beyond level. It corrects itself after some seconds (time to correct proportional to the size of the accumulated error) and becomes normally flyable again.

Now the problem I had today was that I was trying some aggressive flying in stabilise mode (full pitch and roll in various directions and high throttle) to see how the new code was and the errors accumulated were huge. So huge that the copter became totally unflyable and smashed into the ground as it was leaning so far over that I could not actually level it even with stick at full right roll.

The frame behaved admirably and I only wrote off the four props and few of the mounting legs/screws so that's fine.

This problem of accumulating errors in stabilise mode was present in 2.2 but to a much lesser degree and it wasnt in the previous versions at all (or at least a tiny amount) as far as I can tell.

Why is this happening and why has it got worse ? My quad is able to hover perfectly in low wind and the motors etc are all in good condition - I'm convinced this is a controller logic problem.

• ▶ Reply

DeveloperPermalink Reply by R_Lefebvre on February 21, 2012 at 10:10am

Do you have logs?

What you describe could be because of I-term not being tuned properly.

• ▶ Reply

DeveloperPermalink Reply by Marco Robustini on February 21, 2012 at 10:13am

I agree with Robert, post here your parameters.
With the 2.4 should not make mistakes, otherwise those are the results.

• ▶ Reply

3D RoboticsPermalink Reply by Chris Anderson on February 21, 2012 at 11:21am

Dave. Aggressive flying should be done in Acro mode. Stabilize is designed for stability and does a lot of damping, which can account for your results.  

• ▶ Reply

ModeratorPermalink Reply by Graham Dyer on February 21, 2012 at 10:50am

Also trying to download my logs but am getting this below and it's taking a really LONG time...

Is this normal?

 

• ▶ Reply

Permalink Reply by Cliff-E on February 21, 2012 at 8:43pm

Are you downloading over an Xbee?

(If so, don't do it, down load over USB--you can verify by going into the CLI, type 'logs' and should print out log sections--then you can download them in MP's CLI download option).

Fighting loiter/alt hold bouncing currently here... 

• ▶ Reply

Permalink Reply by Cliff-E on February 22, 2012 at 8:25pm

the characters look odd. when I check my DOS screen I see this:

Log] exit while

dump 39

exit while
exit while
exit while
exit while
exit while
exit while
exit while
exit while
exit while
exit while
exit while

exit while
logs enabled: ATTITUDE_MED GPS PM CTUN NTUN RAW CMD

5 logs
Log 38, start 1089, end 1233
Log 39, start 1234, end 999
Log 40, start 1000, end 1061
Log 41, start 1062, end 1077
Log 42, start 1078, end 1088

exit while
Log] exit while
dump 42
logs enabled: ATTITUDE_MED GPS PM CTUN NTUN RAW CMD

5 logs
Log 38, start 1089, end 1233
Log 39, start 1234, end 999

May want to clear your log space (run erase) and log a flight and see if it happens again..

• ▶ Reply

ModeratorPermalink Reply by Graham Dyer on February 22, 2012 at 9:50pm

related to this thread ?

• ▶ Reply

ModeratorPermalink Reply by Graham Dyer on February 22, 2012 at 9:51pm

USB, see here

• ▶ Reply

DeveloperPermalink Reply by Marco Robustini on February 21, 2012 at 11:34am

ArduCopter V2.4 - Full auto mission and Loiter test with my little X-Quad X525:

Hi all! Today I focused my tests on all automated systems, especially the "Auto" and "Loiter" with my little "lab rat", X-Quad 525.
I found some anomalies which have already been reported and I hope will be corrected, including the critical parameters of the Loiter, unfortunately, still linked to other parameters, including "Nav" and "Stab" section.
All this of course does not in any way compromise the stability of this release, good for me, but some things they want improved to make it "perfect".
Here the video of the tests performed, waiting for your test.

Bests

Marco

• ▶ Reply

• ‹ Previous
• 1
• …
• 7
• 8
• 9
• 10
• 11
• …
• 90
• Next ›
• Page