Replies to This Discussion

DeveloperPermalink Reply by Randy on July 12, 2012 at 9:27pm

Florian,

    Ok, well..sort of successful it sounds like.

    Yes, indeed you seem to have a very powerful quad there.  Perhaps a bit more powerful than you really need.  Be careful and good luck with the next steps.

-Randy

• ▶ Reply

Permalink Reply by Florian G on July 12, 2012 at 11:18pm

Hello Randy,

hello everyone,

here are two videos - actually took place during the same "battery load".

http://www.youtube.com/watch?v=m6vHeKIBSR8&feature=plcp

In the first one, i fly for a bit in stabilize, then switch to loiter - nothing really happens.

http://www.youtube.com/watch?v=OSIp9zY4dxM&feature=plcp

On the second video, at 1:45, i push the gas quite heavily and it shoots off.
At 2:32, when i try to take the craft down, it starts to shake heavily left and right, almost goes out of control and catches itself at last again.
Now, at 3:03, i try to land, Assumption (hard to tell for me from video): Here it somehow turns one engine a bit and as one motor is tilted, it delivers a lot of yaw - the craft tries to correct this and then after turning like crazy crashes upside down.

Could this COG maybe due to my lipo being positioned with the long side on one axis and no lipo load along the second axis ? - Attached are also the log files.

Cheers

Florian

Attachments:

• logs.zip, 219 KB

• ▶ Reply

DeveloperPermalink Reply by Randy on July 13, 2012 at 12:55am

Florian,

     Well, you certainly have a nice location to fly in.

     Loiter does seems to be doing something.  Check out the attached screenshot which shows the "nav lon" and "nav lat".  This is the lean (in the lon and lat directions) that the autopilot is feeding to the engines to try and get rid of the "Long Err" and "Lat Err" (which are the distances from the target point.  One thing you may not have realised is that while in Loiter, if you move the roll or pitch sticks off from center you will reset the target location.  So if you keep touching the sticks, Loiter will appear to do nothing.  between rows 1300 ~ 2600 in your logs it is trying to keep the copter in one place however so just give it some more time.

     I had some troubles getting the 2nd video to play all the way to the end despite several attempts ... in any case, coming down quickly often causes wobbles.  It's because the copter is going through it's own prop wash (i.e. disturbed air from the downward force of the props).  The solution -  come down slower!  Going up is normally smooth 'cuz you're going through undisturbed air.

     To me, it looks like you're copter is mostly a-ok.

-Randy

• ▶ Reply

DeveloperPermalink Reply by Randy on July 13, 2012 at 1:01am

Sorry, one more thing..in your logs, you can occasionally see a row called "MOD".  This is short for flight mode and it appears when you change flight modes.  It looks like you're changing your flight modes and sometimes you're spending a very short time in some modes like "RTL" and "AUTO" which makes me think you're not intentionally going into these modes.  You might want to check the mission planner's APM Setup / Flight Modes screen and ensure that as you change the position of your switch, it's going to the appropriate mode.

Also it looks like you landed in loiter mode?  You'll have better control of the throttle in stabilize mode which is important for landing so I think you should switch back to stabilize mode before landing.

Row
19    MOD STABILIZE
1272 MOD ALT_HOLD
1290 MOD LOITER
5576 MOD RTL
5603 MOD AUTO
5620 MOD STABILIZE
6296 MOD LOITER

• ▶ Reply

Permalink Reply by Florian G on July 13, 2012 at 2:04am

Hello Randy,

the location is in fact quite nice.

Unfortunately in 500m away, there is a industrial hazardous waste burning place - so it has a strange feeling to be so close to this - in any case, i could not find a better place to fly searching for a long time.

I have checked again the video and the frame and the wobble is along the axis which does not have too much weigth on in terms of battery - so i guess when it regulates this axis, it should have a different P or so. - The wobble in the video is really heavy. Maybe it check again, if you can see it now ?

- I will re-build the copter and then re-try.

On the mode switching:

I am using GV9x on Turnigy 9x and this has a rolling switch configured.

What you see is exactly the modes and their arrangement. For every push on the switch, it goes up one mode - so i will always cycle through modes.

Stabilize, Alt-Hold, Loiter, RTL, Auto is exactly the sequence. One other possibility could be to install this turning switch:

http://www.ebay.de/itm/Mode-switch-6-POSITION-9X-radio-ArduCopter-A...

However, there i also cycle through modes when changing e.g. from Loiter to Stabilize.

I definitely need to gain some more experience in flying - which i plan to get during the next weeks !

Cheers

Florian

• ▶ Reply

DeveloperPermalink Reply by Randy on July 13, 2012 at 3:26am

Ok, i managed to see the end of the video (it was probably just a corrupted download in my browser history) and indeed the end (where you land) does not look good.

I'm pretty sure that's not a software problem though.  You hit the ground pretty hard.  It's very possible that some wires jiggled loose and caused a brown out or power surge or something like that.

Permalink Reply by Eddie Weeks on July 13, 2012 at 10:54am

If your craft is over powered, just add another battery to weigh it down... !!...

The motors really need to spin up to be stable and to have things work right.. 

I had that problem on many tri / quads...  I used 4S and 900 kv with 12x3.8 = BAD...

Now  4S 600kv 12x3.8.... much much better

Eddie Weeks

• ▶ Reply

Permalink Reply by Florian G on July 13, 2012 at 11:45am

Hi Eddie !

I have done a few calculations - the idea to to add another battery is not too bad - this would take the lift-off throttle from 31% to 47% - just about right.

Here are some further calculations on your craft:

4S / 900 kV / 12x3.8 -> Draws approx. 410 W per engine and provides enough thrust for a copter of 3 kg - max. carry in addition is 1,5 kg (total 4,5 kg)
4S / 900 kv / 12x3.8 -> Draws approx. 185 W per engine and provides enough thrust for a copter of 1,7 kg - max. carry in addition is 650 g (total 2,35 kg)

How heavy is your copter and which types of engines do you use ?

It i am not too far off, then your craft actually should be in the range of 1,5 - 2 kg ?


Just as a general idea - what do you think about the following configuration ?

8S or 9S / 400 kv / 15x4 -> Draws approx. 1275 W per engine and provides enough thrust for a copter of 5 kg (34% throttle) - max. carry in addition is 5 kg (total of 10 kg) ?

I like to use 8S or 9S - just because the Current is lower, easier to measure and cables are thinner.

BTW: If somebody is interested - i am developing a 12S (50V) Fast ESC currently for up to 70A - another smaller 12S 40A version is likely to follow..

Cheers

Florian

• ▶ Reply

Permalink Reply by PACEFE on July 15, 2012 at 6:35am

Hi,

  would it be possible to get a cable about 20cm to connect the APM2 and daughter board? It might be nice to have the possibility to have magnetometer away from magnetic fields and at the same time been able to correct desviation on the mag calibration by rotating the pcb as I can do on my NAZA controller....

I know we can order gps and mag separated but this have an impact in $$.

Pablo

• ▶ Reply

Permalink Reply by Vernon Barry on July 15, 2012 at 7:15am

Anything is possible, the problem here is pratical. If you remember with APM 1.0, the magnetometer was separate and there was a rather nasty IC2 bus bug that if in flight, the signal was interupted, the entire APM hung. While the library used for that function has been updated, it's still not clear exactly what happens when you have a cable issue. It also comes down to cost of manufacturing. There is a reason why APM 2.0 is complete, has everything and is $50 cheaper than APM 1.0. Since the boards are handled with a pick and place machine, they can be rapidly produced, soldered and tested. By having the components on 2 board, they eliminate a lot of previous electrical issues found from all the separate items before. Physical mounting and orientation between the sensors is just a critical. Mounting your own board, some distance away is sure to cause a loss of precision, not a gain as you are assuming. A non directional sensor such as the speed sensor or other devices make sense for remote mounting but not the magnetometer since it is best to be as physicially close to the gyros as possible.

A healthy dose of commons sense to remove wiring and other noise sources from the location of the APM is the far more "correct" choice. Even a half inch is enough to make a difference.

• ▶ Reply

Permalink Reply by PACEFE on July 15, 2012 at 7:55am

HI Vernon,

               Im sorry to disagree.

In my naza the Magnetometer is conected via cable to IMU. This way the MAg is not affected by motor currents, etc. And if you happen to have this nasty circles during Lotier, you just have to rotate the gps module (that includes the mag) until the circle tendency is gone.

I have never succeded to adjust my magnetometer. Look how measured values change when I apply power to my motors in the attached file. At 2103 you can see the motor stopping and the sudden chance in MAG.

Also with no current and with the correct declination it is imposible to set North, at lest here in Argentina. I just have to fly with out compass meaning no Lotier / Auto / rtl is possible.

This is no challenge at all for naza, and I think it has a lot to do with mag position.

Pablo

Attachments:

• compass.jpg, 988 KB

• ▶ Reply

DeveloperPermalink Reply by Randy on July 15, 2012 at 8:34am

By the way, if you find rotating the compass a bit helps getting rid of the circling, it really means that your declination is off.  I don'tknow about the DJI but in ArduCopter, there's a declination field in one of the APMSetup screen in the mission planner.

I forget which way you have to do it..i.e. if you're circling clockwise you should reduce the declination (try starting with 10 degrees to see if that helps).

• ▶ Reply

• ‹ Previous
• 1
• 2
• 3
• 4
• Next ›
• Page