Replies

  • Today my first autotune with Arducopter 3.1.4 on APM 2.6, esc's 40a SimonK, MK3638 motors, props APC 12x3.8, 4s battery and total weight 3.5Kg.

    1,30 min for roll, 1.30 min for pitch. Why it was so fast ?

    I like final result, very smooth flight.

    here video:

    http://youtu.be/4NGAkf2LJhs

    log attached;

    Ciao - Giuseppe

     

    2014-05-24 07-28-18_esamk3638_autotune.log

  • Successful auto tune, I think.

    Just took advantage of a rare windless day and did a successful auto tune.  The quad flys OK after saving the new PIDs but they are very far from the original default PIDs. See attached picture of the new ones.

    Question is, are these OK or did something go wildly wrong?  The quad flies OK in Satb and Loiter but past issues in Loiter only turned up when it was windy, so I can't test for that today.

    3DR 2013 quad, 880 motors, 11 inch props, flying at 6300 feet elevation (not from from Spark Fun), firmware 3.13, APM 2.5.

    We have a 5000ma battery on the bottom and an Action Pro cam on a pitch control  mount.

    20140523_102711.jpg

    • Flew in 15 to 20 MPH winds yesterday and the auto tune derived PIDs were very stable. Flew in Stabilize, Loiter, and Circle.  The quad was much less jerky in Loiter than using the default PIDs.  And while I need some more flights to be sure it seemed I was getting longer time in the air before the battery alarm went off.  Probably because it was working less hard to stay put and stable.

      It seems the 10 minutes it took to do auto tune from the defaults was because the defaults were far from the end result.  I suspect if I did it again to 'tweak' the first set it would go much faster.

      So I'd declare this a success.  Auto tune works fine.  It's a great addition.  Here is what I've concluded:

      - If you change from the default physical config autotune will take longer.  In my case it was 880 motors, 11 inch props, 5Ah battery plus 1Ah battery, flying at 6,300 feet elevation, elevating the APM about 2" above the base and the power distro module.

      - To get the autotune switch to work you must have a wire connection between ch7 or 8 on your controller and the  7 or 8 input to the APM.  I had to add that wire.

      - The less wind you have the faster it will get done. If the PIDs you start out with are far from ideal it's going to take up to 10 minutes or so. The more time you spend driving it back to the test area the less time is available to complete the tune process.  I got through it successfully only when I had a day that was completely still so I had to pull it back from drifting off only three or four times, and for very short distances.

      - Be sure to read and understand how to preserve the PIDs that result. I had one false start because I complete the tune then flipped the wrong switch and those were lost.

      - Assure the GPS has a good fix before you start.  I left power on for about 2 minutes sitting on the ground before arming to assure it had plenty of sats.

      - Assure all the physical parts are secure and not moving around when the copter does the rather violent roll and pitch maneuvers that are part of the auto tune process.  If something moves around it will create noise in the accel data and the process may fail or result in bad PIDS.  I ended up bracing my camera mount laterally to stiffen it, just for the autotune.  Before I did that I could see it flexing a bit side to side even while moving hard left and right in Stabilize.

      These notes are based on my experience. Your mileage may vary.

      • Top notch documentation Jim, thanks a million. Some of these tips should be added to the wiki!

  • Here is a log of one of my excessively long tuning flights..  (I actually went down after 11 minutes because the battery pooped out..   I had my telemetry volume turned down and didn't hear the battery warning).

    My frame is the standard Hex sold by  3DR.   It has 880KV motors and an aging 5000 mA 4S battery.  I had just replaced my APM2.5 with a new APM2.6 controller with an external compass & mag.  I flashed it with 3.1 firmware and was using the default PIDs.   I was hoping autotune would find the 'right' settings.

    2014-05-18 07-15-08.log

  • Hi,  I'm a little stuck with the auto tune.  I get roll corrections for about 6 minutes, then my low voltage alarm goes off and I have to land.  How can I get a good auto tune with 7 minute flight times?  Is there a way to split the roll and pitch functions, or make it faster?  Thanks!

    -Nate

    • Developer

      If you post the logs here I might be able to give you some starting pid values that get you there faster.

      The time autotune takes depends on how calm the conditions are and how well setup your copter is for autotune. By how well setup I am referring to things like movement in the arms, battery, or gimbal and how well chosen the motor/prop/battery is for the frame and weight.

      • hi leonard, how is it like doing the autotune a pixhawk with the EKF on? Is it fine to do that?

    • I have a similar problem..  

      I can only stay aloft for about 12 minutes before I have to start worrying about the battery.    I just did an autotune today for the fist time (I had zero wind).  My 3DR Hex behaved just like in the video at the beginning of this thread.  Alt hold works very well on my Hex and it only took minimal correction to keep it in one place.

      But..  After 12 minutes, it was still doing the fore/aft pitching.  I had to land.

      How long is this supposed to take?

This reply was deleted.

Activity

DIY Robocars via Twitter
RT @a1k0n: @DanielChiaJH @diyrobocars @circuitlaunch Here's my car's view of that race. About 8.4 second lap times for laps 2 and 3... both…
yesterday
DIY Robocars via Twitter
RT @DanielChiaJH: Great racing against @a1k0n today at @diyrobocars! Pretty cool to both break sun-9s at the track today I think I got very…
Sunday
DIY Robocars via Twitter
Broadcasting the @circuitlaunch race live now at https://m.twitch.tv/diyrobocars Races begin around 2:00pm PT
Saturday
DIY Robocars via Twitter
RT @a1k0n: ran a huge number of hyperparameter tuning experiments yesterday; now I can train a new policy, far with better quality, in 15 m…
Saturday
DIY Robocars via Twitter
RT @a1k0n: Did I get rid of hand-tuned parameters? Yes. Am I still hand-tuning more parameters? Also yes. I have a few knobs to address the…
Sep 26
DIY Robocars via Twitter
RT @a1k0n: I'm not going to spoil it, but (after charging the battery) this works way better than it has any right to. The car is now faste…
Sep 26
DIY Robocars via Twitter
RT @a1k0n: Decided to just see what happens if I run the sim-trained neural net on the car, with some safety rails around max throttle slew…
Sep 26
DIY Robocars via Twitter
Sep 24
DIY Robocars via Twitter
RT @SmallpixelCar: @a1k0n @diyrobocars I learned from this. This is my speed profile. Looks like I am too conservative on the right side of…
Sep 24
DIY Robocars via Twitter
RT @a1k0n: @SmallpixelCar @diyrobocars Dot color is speed; brighter is faster. Yeah, it has less room to explore in the tighter part, and t…
Sep 24
DIY Robocars via Twitter
RT @a1k0n: I'm gonna try to do proper offline reinforcement learning for @diyrobocars and throw away all my manual parameter tuning for the…
Sep 23
DIY Robocars via Twitter
RT @circuitlaunch: DIY Robocars & Brazilian BBQ - Sat 10/1. Our track combines hairpin curves with an intersection for max danger. Take tha…
Sep 22
DIY Robocars via Twitter
RT @SmallpixelCar: Had an great test today on @RAMS_RC_Club track. However the car starts to drift at 40mph. Some experts recommended to ch…
Sep 11
DIY Robocars via Twitter
RT @gclue_akira: 世界最速 チームtamiyaのaiカー https://t.co/1Qq2zOeftG
Sep 10
DIY Robocars via Twitter
RT @DanielChiaJH: Always a good time working on my @diyrobocars car at @circuitlaunch. Still got some work to do if I’m to beat @a1k0n howe…
Sep 10
DIY Robocars via Twitter
RT @SmallpixelCar: My new speed profile for @RAMS_RC_Club track https://t.co/RtLb7TcgIJ
Sep 10
More…