Replies to This Discussion

Permalink Reply by Georg F. on July 18, 2012 at 12:53pm

Has anyone an idea?

• ▶ Reply

DeveloperPermalink Reply by Adam Rivera on July 18, 2012 at 12:55pm

I am assuming you are using the Mediatek GPS... what kind of accuracy are you currently getting?

• ▶ Reply

Permalink Reply by Georg F. on July 23, 2012 at 10:32am

Here is a Link to one of my last flights. Here you could see, my GPS is not so exactly as i hoped. It fly in Loiter an radius of 10 - 15 meter.

Download Log File

 

Hope someone could help me with my Settings:

• ▶ Reply

Permalink Reply by shayan on July 23, 2012 at 1:02pm

hi georg
could you please give us some information about your arducopter such as what motor,prop and Also wight...?
is alt hold works with these parameters?

• ▶ Reply

Permalink Reply by Georg F. on July 26, 2012 at 7:48am

motors Suppo 2212/13
 turnigy 18a plush
accu turnigy 2.200 mHa 30C
wight ca. 1.2000 Gramm

quadcopter X

Alt hold works fine. But the copter hold the position not so exatctly as I hoped...

• ▶ Reply

3D RoboticsPermalink Reply by Chris Anderson on July 26, 2012 at 10:20am

Georg: Please upgrade to 2.7 via the MP. 

• ▶ Reply

Permalink Reply by Jani on July 11, 2012 at 6:06am

Hi,

If i unsolder compass module from APM1, place it like DJI style with cable and change orientation, do i need to comple code manually with edited module position like manual says:

http://code.google.com/p/arducopter/wiki/AC2_Magnetos

or is the new position handled with calibration in MP??

Regards Jani

• ▶ Reply

Permalink Reply by palo sestak on July 11, 2012 at 6:15am

Hi all,

just got back from field with broken tricopter ;o(

loaded v2.6 (slightly modified for LEDs, others untouched) and tried only basics (STAB, LOITER, RTL, AUTO, RTL ...)

When switched from RTL to stab copter seemed to not respond and gone away. I switched back to RTL but didn't help and copter fall down in the field. Luckily I found it ;o) (only broken arms and props). From logs I can see the roll (yellow curve) got crazy right before the copter fall down (around 7250).

Can somebody explain whats going on from log? (log attached)

Attachments:

• 2012-07-11 13-36 5.log, 1.3 MB

• ▶ Reply

Permalink Reply by palo sestak on July 11, 2012 at 9:10am

When I dug bit in log file I found mag heading - went too high. Can somebody explain me what mag heading is (what units it is)?

• ▶ Reply

DeveloperPermalink Reply by Randy on July 12, 2012 at 10:15pm

Palo,

     Those are in degrees * 100.    So it's not terribly unusual for it to have big jumps like.  Now, if it wasn't really turning that much then something went wrong with the compass readings..looking at the rapidly changing roll, I'd guess that the compass offsets are bad..you can check this by leaning the copter left and right in your hand 10 or 20 degrees..does your compass heading change by more than 10 or 15 degrees?

-Randy

• ▶ Reply

DeveloperPermalink Reply by Randy on July 12, 2012 at 10:28pm

Palo,

     I had a look at your logs a bit.  CTUN is another useful log type to enable as it shows you your altitude and throttle info.  In any case..I'd say your roll and pitch PID values are not great.  Looks like you're getting familiar with checking the logs so you've probably already seen this but attached is your Pitch PID values with the pilot input being in red, the tri's actual attitude in green.  If you check out the time from 25~30, the copter is consistently 30 degrees off of what you're commanding it to do.  I'd say your PID values are far too low.  I'd suggest increasing the pitch rate controller's P value for sure.

     By the way, you can actually turn on PID logging for the controllers them selves.  "logs", "enable PID" then use channel 6 to tune their the stabilize controller or the inner rate controller and it should dump out the P, I and D values into the dataflash and that will give some more details about how well the controller is responding.

-Randy

• ▶ Reply

Permalink Reply by palo sestak on July 13, 2012 at 1:49am

Randy,

thanks a lot for your interest. When I stared flight I was almost sure the NTUN and CTUN are enabled too  but weren't ;o(

PIDs are taken from somebody from here with tricopter. These PIDs are the first I can flight the loiter with without fear.  When the PIDs are higher my tri in loiter is going crazy and aggressively hunting the position with pitch and roll over 50 or 70 degrees :-o.

Therefore I was glad with such lazy PIDs. But seems they should be really bit upgraded ;o) I have to learn inflight PID tuning, didn't try it yet ;o). But have no much time to going to field, just once per month or so :o(

Anyway the crash is what happened to me more times - tri stopped respond to my commands in STABILIZE and even RTL didn't help. This is the point I must solve otherwise I lose confidence in APM at all. Perhaps it has nothing to do with APM, just battery was too low and APM hadn't enough power to stabilize it, who knows... I can't read it from logs.

With v2.6 and lazy PIDs I was almost happy when the stabilize appeared very good now and loiter wasn't that stressful. AUTO seemed to be OK too but after that I lose the control and scenario repeated.

Anyway since 2.0.49 it is most reliable version of ACM for me.

I wish developers many successes with the code. I'm going to like it since I do small modifications in it and it seems to be more transparent for me now.

Yet another question - how can I achieve faster climbing to next WP?

Scenario : I have tree WPs basically on the same place but middle one has far higher altitude.

I need the tricopter to climb fast to say 2000 meters and then safely dive back to home (just measure humidity and pressure with onboard device in altitude profile).

With my PIDs the climbing and diving to next WP is very slow and with that long flying I'm taking a chance going out of mAh in battery ...

• ▶ Reply

• ‹ Previous
• 1
• …
• 58
• 59
• 60
• 61
• 62
• …
• 68
• Next ›
• Page