Blogs

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

Join diydrones

Comments

  • Greg October 18, 2011 at 4:17am

    I think after doing everything possible to reduce the sonar and baro noise as I reported here (Sonar-nose-reduction-update) I have to agree something else is needed for better altitude control.

  • Aurelio R. Ramos October 17, 2011 at 8:41pm

    It always bothered me why barometer or sonar noise has to be a problem. I would think that accel Z can be integrated twice for a short term aproximation of altitude using leaky integrators. The result can be combined with low pass filtered altitude sensor data to drive an altitude control loop. Most disturbances that upset altitude occur in the short term, when a very accurate accelerometer reading would be useful. Think of it as dead reckoning altitude control by the accelerometer, or using different sensors to satisfy different bandwidths of the desired signal. Not very critical for the bandwidths to overlap precisely as long as the total bandwidth covers the area of interest.

     

    For a few months I had a local modification on ardupirates which used a similar method to implement altitude damping, so that rate of altitude change would be countered by a D term

     

    leaky integral of Z + G (to eliminate the constant pull of gravity, which manifests as Z = -G when sitting level) gives short term speed, which scaled by a negative constant (experimentally tuned) to create a comfortable level of manual control of altitude by user throttle input. Integrator leak rate is also experimentally tuned.

     

    For the feature to work properly I added some clamping to limit its effect on the throttle and scaled its influence down as the pitch / roll of the craft deviated from horizontal, in order to prevent the damping effect from doing anything during any maneuver other than a straight hover.

     

    I was very happy with my feature, enabled me (a complete beginner) to hover at a very constant altitude just over head height) and follow the craft down a trail near my house.

     

    Precise altitude hovers on AC2 require a lot of constant throttle input for me.

     

    Now I am trying to teach myself better control and move beyond beginner level, but I might port the change to my local AC2 code.

  • Developer
    Olivier ADLER October 17, 2011 at 12:58am

     

    The baro in the Wookong is the same as in MK. (MPXAZ6115A)

     

    it is an analog output Baro with 0 - +5V output.

     

    Obviously the precision here does certainly come mostly from the analog to digital converter precision, and carefull hardware design to avoid noise.

     

    http://cache.freescale.com/files/sensors/doc/data_sheet/MPXAZ6115A....

     

    For those interested in the wookong design it's here :

     

    http://www.rcgroups.com/forums/showthread.php?t=1510587

     

     

    http://cache.freescale.com/files/sensors/doc/data_sheet/MPXAZ6115A.pdf?fpsp=1
  • Denny Rowland October 16, 2011 at 10:51pm

    I believe the pressure sensor in the Wookong M is the same as in MK, both deliver outstanding alt. hold but Wookong is unbelievably good. I think we need one of those. 

  • Greg October 15, 2011 at 4:09pm
    Eduardo I think there is some validity in your comments, but at the same time the arducopter is more an open community development platform and lots of potential moving forward. I agree the Altitude precision is limited at the moment. Many people experience height fluctuations and this seems to be more related to the baro than the sonar provided the correct postion and noise filtering are used. I have the recommended pad on the baro, and the on board sensor filters enabled but this makes little difference. There is no further adjustment on the baro beyond this that I am aware of My feeling is an updated sensor array is needed which would include replacing the baro or having a stand alone baro with better precision. It is interesting MK can get so much precision out of their baro alone. Also the arducopter is designed as an automonous vehicle capable of complex mission profiles as opposed to a professional filming platform. I am very interested now in trying different control electronics now from your post. I certainly would like something that can hold a precice position at a fixed altitude rather than circling and moving up and down. Right now even in a slight breeze I have to fully focus on keeping my copter in the air. It wobbles when there is a slight variation in wind and on decent with the latest code. This is a standard design arducopter not custom. Of course I like a challenge and its fun sorting out issues and I learn more each week. In my case I still need to tune the arducopter obviously but I wonder about the design limits like any new development.
  • Fouad October 15, 2011 at 12:42pm

    @Owen, Well said , couldn't express it better.

    Good luck for all.

  • Owen Booth October 14, 2011 at 6:23pm

    In the 8 months since I bought my AC2, the updates that the community, and especially Jason Short, Michael Oborne, and the rest of the core team etc have uploaded, have taken this craft from a skitterish novelty item, to a comparatively smooth, reliable platform. I wouldn't of mounted my GoPro HD on it if I thought it would let me down. Mikrocopter make some very flash kit, but at five times the price. In addition, the MK flies out of the box, which might suit a customer who needs it to just work, but the AC project gives everybody the opportunity to participate in it's developement, something which has provided me, and thousands of others, with hundreds of hours of imaginative entertainment. Keep up the fantastic work y'all!

  • Denny Rowland October 14, 2011 at 9:41am

    Although I am now using DJI I must say that AC 2.0.x is fast becoming a real player in the heavy lift AP stakes and I have set aside a machine just for the purpose of testing each and every release as it comes along.

  • Denny Rowland October 14, 2011 at 9:31am

    @ I.S.

    37 " Some things are changed since that picture. Battery is now one single 5800 ma. and further back. props. are now 13.5 x3.5 carbon and they are swept forward. Would be too dangerous outside of those shrouds .Weight is now closer to 5 Kilos depending on ancillaries. My multipilot 32 system was a lot heavier than the DJI.

     

     

    A good prop. is the 14x4 xoar wood.

  • Denny Rowland October 14, 2011 at 8:03am

    The high mass beam floats on about 2" of open cell foam rubber. This isolates the mount from vibration and also provides an important base that will not move as a result of opposite reaction from the mount servos.

This reply was deleted.