3D Robotics

ArduPilotMega Beta code now live!

After more than a 1,000 commits and many thousand lines of code, the public beta of the ArduPilotMega code is now out. You can get it here. Note that the new libraries requires a slightly different installation process, so read the manual so you can put the libraries in the right folder. The code won't compile if you don't. Also, you'll need to be using Arduino 19 or higher.

New features include:

  • An awesome command-line interpreter (CLI) running on APM, which allows you to make most common configurations and setup without having to recompile the code.
  • The CLI also allows you to interactively test all your hardware components--no messing with Arduino code or configuration files if you don't want to!
  • Powerful new scripting commands and flight options.
  • New FastSerial code, replacing the standard Arduino serial code, which makes serial communications lightning fast.
  • The entire code has been re-architected to use libraries, which are shared across the ArduPilot projects. That way we can improve everything from GPS parsers to basic DCM routines and all the projects will immediately inherit the improvements. That not only means better code for everyone, but much easier maintenance for the team.

Speaking of the team, special thanks to the core APM dev team that put in so much work and creativity to get us here. These guys have worked tirelessly on this, with daily coding sessions and weekly dev conference calls, all year. This is a huge project, perhaps the most complex Arduino code ever written, and these are the stars who led it:

  • Jason Short
  • Michael Smith
  • Doug Weibel

With help from some ArduCopter team members:

  • Jose Julio
  • Jani Hirvinen

Config utility wizadry from:

  • Michael Oborne
  • Ted Carancho (ArduCopter Configurator)
And all this made possible by Hardware Ninja Jordi Munoz!

As always, please report all bugs and problems in the Issue Tracker. The developers can't respond to bug reports in blog comments and discussion forums.

E-mail me when people leave their comments –

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

Join diydrones


  • Ok going for another test flight on my new and improved Easystar. I tried to setup the radio values with the CLI but it shows the default values only and regardless of my switching nothing changes even though the varying flaps are jiggling. Remove the servo connections and I can set it up no problem. Any ideas? I'm still using the beta too.
    Its safe to setup without servos then fly with right?

    Also I only have a toggle switch for modes so am using channel 7 (not 8) and having channel 8 as just my failsafe like Doug suggested a while back. I've set the values in the config for my modes but the radio setup doesnt do ch4-8 does this matter also?
  • 3D Robotics
    Steve> No, not yet. Camera stabilization is on the roadmap, but since this is an open source project it can be included tomorrow if someone in the community (maybe you?) wants to add it.
  • Does the code currently have camera stabilization? If not, what is the timeline for adding it?
  • Benjamin,
    I have been wondering about the same. Maybe it would be really nice to have a system similar to the issue list (used for bug reporting), just a system for filing test reports? One would then enter the revision number of code tested, flight modes tested, observations that were made, etc. As far as I know in software projects they use these kinds of things for test reports.

    The obvious problem will be that there are as many plane/quad/system configurations as there are users here. Still, such a thing would be useful IMHO. Can anybody from the admins comment on that?
  • Maiden flight without damage, but some issues:

    Airframe: Multiplex Easystar, rudder, elevon.

    MANUAL mode works well.
    STABILIZE mode, with the default PID gain settings very weak corrections.
    LOITER mode, throttle goes to off, airplane turns to a spiral, could activate it only for some seconds
    AUTO mode, could not test it as I do not trust in STABILIZE mode.

    I tried to increase the PID gains (put them to 5), so corrections are clearly visible, but a bit sluggish. Side effect is that the zeros in the different flight modes are not equal.

    Comments of other maiden flyers welcome.
  • The reason I ask is that a lot of the time new SVN code is not flight-tested, so it's hard to know what is "supposed to" work. For instance, the beta currently does not work (see issue 180; it was resolved in the trunk, but not the beta -- see issue 173).

    It would be nice to get feedback about what versions are supposed to work, and which ones have untested code. It seems like it would be immensely useful to ask people to drop a quick note when they make successful flights including what revision of what branch was tested, and which functions were performed successfully.

    I do and will submit issues when I nail down the problem specifically enough; I usually provide a suggested fix as well.
  • 3D Robotics

    Doug Weibel was the main test pilot for this version of the code, so you might want to check with him. We've all been flying various versions of the code on different aircraft, so you may find that the gains aren't right for yours. That's the point of releasing a public beta, so that more people with more airframes can put the code in the air.

    I recommend that you fly the unmodified code and if you're having problems, please file an issue. We have weekly developer conf calls to review the filed issues of the week, but the developers aren't regularly following blog comments.
  • Has anyone actually performed a successful, full test flight with the beta software on a physical plane including the following?
    * Setting flight modes to non-standard values with the CLI
    * Waypoint following in Auto for at least 5 waypoints and a "loiter N turns"
    I'm running a modified version of the trunk code which has some extra features I need so I'm not totally sure my additions aren't causing the problems, but I've been unable to get Auto mode to work since before the beta release. Stabilize and FBWA work, and Auto kind of generally gets the plane going in the right direction, but throttle and pitch control are poor, and the plane kind of does S turns towards the points (as if nav gains were too low, but I haven't changed the default gains).

    As far as I know, there haven't been any explicit reports of, "I downloaded the beta, put it on my APM, flew my real plane with it, and it went great." If you've done this, could you let everyone know to give confidence that the beta is fully operational and that I'm messing something up? :)
  • Beta code works perfect!

    All the issues with the alpha versions are solved.

    Maiden flight may be this evening! In any case this week!
  • lol silly me. back to the alpa
This reply was deleted.