Sticks of Centre after fllying

Hi Guys, I've got a problem with my Ardu-Mega that I just can't find.

 

In rate mode it flys fine in a gentle hover and its very stable.

Stable mode it comes off the ground without any cyclic input and its absolutely rock solid 

 

until..

 

If I fly around a bit the sticks go off centre, it can get to the point that I'm holding full forward elevator just to hold a hover.  

 

Switch off/on and its fine again until I fly around.

 

I soldered the filter bridges last night but it's made no difference.

 

Just can;t find anything wrong - could someone give me some ideas to look at ?

 

 

PS I'm fairly new to Ardu.

You need to be a member of diydrones to add comments!

Join diydrones

Email me when people reply –

Replies

  • Same here. If I move the roll stick to forward or backward the quad goes always forward. If I move the pitch to left/right the quad goes always to right.

    This happens ONLY when the quad is in air. At the ground, I tested the radio and the motos spin in the right order, I mean forward=front motors spin lesser than back motors; rigth= right motors spin lesser than left motors.

    Any help appreciated

     

  • So what's the deal with this?  Is this a real issue or not?  It sounds serious and I'm fairly concerned about trying any forward flight, but I don't see much discussion?  Is this only happening to a few people, or have only a few people attempted forward flight for long distances?

     

    Does the Arducopter lose track of the "level" setting or not?

  • Wow, this sounds very similar to the problem I had with Flymentor.  After flying forward for a while, it "forgets" where down is.  That system only has 3 gyros, no accelerometers, so it can never actually tell where "down" is, it just tries to keep track of how much you have rotated.  But over time (10 seconds), the rotation correction decays so that whatever attitude you've been flying at is the new level. 

     

    I thought this couldn't happen with Arducopter.  I thought the "level" is locked in via the accelerometers when you do the setup?  Why would it ever accept any other orientation?

     

    Or is it the case that *accute* vibration at the time of switching to Stabilize mode is causing the problem?  If that is what is being said, why does it seem like the problem occurs after flying around for a while?

  • Ryan, after you get confident with Arduino, you can start downloading the soure code from the download section in http://code.google.com/p/arducopter/downloads/list

    and have a look at the basic instructions to compile your code here: http://code.google.com/p/arducopter/wiki/Programming

    One you have downloaded it, you should always copy the libraries folder to the arduino folder. This way you can compile the code.

    APM_config.h is the place where you can configure your copter for different configurations (eg. QUAD_FRAME or TRI_FRAME etc.)

    Once you have succesfully tested the compiling, you can upload it through your USB connection.

    Be sure to select proper board and COM port from the Tools menu.

     

    Have fun!

    Emile

  • If you guys are confident with compiling, you can change in motors.pde the kp_roll_pitch value.

     

    dcm.kp_roll_pitch(0.030000);

     

    If you raise it you will have higher drift correction.

    This seems to have corrected a similar problem I had.

    In my case I ended up in a tree backwards after 5 minutes of FPV at high forward speed.

    It seems to happen a when you fly for a long time at a decent speed (using a lot your elevator) .

    But I'm using a diffrent processor at different speeds, so I can't guarantee for you.

    Only thing I can say is that previously (2.0.39 and before) this value was set to 0.12, but DCM speeds were also different.

    As usual a good test is mandatory before you fly. Personally I think that if you set it at 0.08 or 0.1 you will have some benefits without compromising your flight behaviour.

     

    Hope it will help you.

    Cheers,

    Emile

     

    P.S. Jason's suggest is also mandatory before putting your hands in the code. A good vibration dampening is essential!

  • Developer

    The accelerometers are receiving too much noise and are influencing the orientation too much with bad data.

    The filter pads were an unfortunate design choice. They are not optional.

    Please  do your best to balance your props and isolate the APM from vibrations with foam.

    Jason

This reply was deleted.

Activity