I want to combine APM and a 3d FBL unit


my name ist Mathias, I am from Germany and I am a lot into 3d FBL helicopters but I am newb to APM yet.

Some days I want to do other things with my heli than just smack around.
Here come all the features I can do with gps hold and waypoints…

Now, I try to start my questions directly.

first, is it possible to combine a high end FBL system with the APM?
Like in normal mode all signals are processed through the APM with no changes so my fbl has control.
But, lets say I want gps hold, the APM gives corresponding inputs. My FBL could do selflevel but needs just some inputs to stay at gps place.

Or lets say the APM is inactiv but when it realizes an altitude undershoot it activates selflevel plus pitch…like that.

second, I am quite sure this could be done in general. But can it be done with existing software or would I need to program?

I dont want to go further and I hope you get the point of my imagination.

I already thank you for some help.


Currently, this is not possible in the code. Not at all.

I’ve thought about this sort of inter-operability in the past. But it seems like a lot of technical complexity for no good reason. I decided to focus my efforts into making the acrobatic performance of the Arducopter code better, so that you would not even consider using another controller.

So, as is the way with Open Source, you’re free to work on this if you want. But it’s not something I’m interested in doing.



As Mathias I’m also into 3D helicopters since some time. I’m thinking about jumping into the APM world to use a TREX 600E PRO (FBL) for other flight goals than 3D. I want to be able to keep the 3D/Aerobatics capabilities of the TREX with the added value of the position control (with GPS and all related functions) when needed. I have some questions about this before buying anything, and I’m using this old post to keep it’s old followers updated.

The thing is that I read on this page (http://copter.ardupilot.com/wiki/traditional-helicopters/traditional-helicopter-tuning/#aerobatics_in_acro_mode) that the 3D/aerobatic behavior of the controller is not optimal compared to a “normal” flybarless module. The first question is : is this documentation and behavior still actual? Perhaps things evolved thus making the problem solved. In other words : right now, is the ACRO mode comparable to standard flybarless modules in terms of 3D/aerobatics capabilities (no pitching limitation and no commands in earth frame)?

If not, I re-ask the question of the feasibility to use an external FBL controller (MicroBeast in my case) with the following configuration :
[li]Swash Type (H_SWASH_TYPE) = 1 --> No CCPM mixing meaning for eg. roll/pitch/collective commands will be separated on PWM outputs[/li]
[li]Flybar Mode Selector (H_FLYBAR_MODE) = 1 --> As read on the doc page stated above, this will shortcut any flybarless/rate control in the APM and output direct stick inputs in ACRO mode (which is what I want for 3D/Aerobatics flight)[/li]
[li]Tail Type (H_TAIL_TYPE) = 1 --> Use of external FBL tail control[/li][/ul]

Would this setup be usable? Any limitation? Did anyone test something similar? I guess that with Flybar Mode Selector set to 1 and in any other mode that ACRO, all the GPS functions (loiter, RTH,…) are still available. If someone could confirm this…

By the way, my last question is about the SBUS output on the PixHawk. Is it functional and what are the output signals (raw stick inputs or values from the control loops)?. Of course, the idea here is to use SBUS from PixHawk to external FBL controller…

Thank you very much in advance for any time used to answer these questions :smiley:

Best regards,


After looking on the internet for some time without finding anything else than other people requesting the same thing, I do a little UP for this post…


The wiki was quite out of date on this issue, and I have updated it to reflect the current situation (but have more work to do!). As of AC3.0 (IIRC) we have been capable of full acrobatic flight, including full flips and rolls past vertical. I actually enjoy flying in acro, and do high speed flips, rolls, or stationary flips, also Split-S, Hammerheads, etc. Also, all commands are now done in “body frame” not “earth frame” as previously stated in the wiki.

I have not done any aggressive 3D stuff, and there is some risk that the EKF will get upset in this case. But you can certainly have fun with sport type flying.

Hi Rob,

Thank you very much for your answer :slight_smile: I read your update in the wiki. Of course now that I know more about the suject, I have some more questions :
[li]About the agressive 3D possible issue in ACRO mode, which element in the loop could cause a failure? I guess the control loops in ACRO are “just” on the rate so no problem at this level. How about the rate estimation by the EFK? Will these value be affected by strong shaking, even if they are close to the sensor in the EFK process?[/li]
[li]For a well tuned controller, what are the performances and felling of the control loops compared to a standard FBL module (eg. during strong collective pitch variation, in heavy wind, during high speed flight,…[/li]
[li]How do you manage the problem that may happens when switching ON the Altitutde Control module (eg. when hovering in ACRO or STAB, the collective pitch stick will be above neutral position, especially in negative pitch is needed for inverted flight)[/li]
[li]How do you manage the risk of dismaring the system in inverted flight? What are the collective stick position needed to arm/disarm and the time needed at this position to disarm?[/li][/ul]

Thanks for the support :slight_smile:

The problem with aggressive 3D acrobatics is going to be the EKF. The problem is that our current Acro mode uses the attitude from the EKF to achieve a very tenacious attitude hold that does not drift, etc. It also gets it’s rates from the EKF, instead of the gyros directly. I have had a case where the EKF went bad in flight due to a bug in some beta code, and I was able to land it in Acro mode, but it was difficult as it kept moving around on me. In fact it was somewhat of a crash landing, I just got it over the runway about 2 feet off the ground and killed the motor.

I haven’t had a problem since that bug was fixed, but I want to caution people as, I can’t actually test the 3D performance myself, as I’m not capable of piloting those maneuvers.

What is needed, and we have talked about this, is a fall-back mode which uses pure data from the accelerometers. Something you could switch to should the EKF go bad. I expect this will happen for AC3.4, it’s pretty high up on my list.

As for the performance relative to other controllers, hard to say, I haven’t flown any others. Nor have I ever seen any kind of data from other controllers. Best I’ve seen are commentary from users about how “locked in” something feels, which is pretty subjective.

For collective control when switching modes, there really is no strategy. It’s up to the pilot to do what is necessary. If you are Stabilize mode, hovering with the collective above the center, and switch to Alt Hold, it will start climbing unless you move the stick to center. There’s really nothing we can do about this. We have thought of ways to solve it, but every method leads to other, much worse, problems. At the end of the day, it is still up to the pilot, to pilot the vehicle.

For disarming in flight, nothing can be done. If you hold full negative collective, and left yaw for 3 seconds, it disarms. Again, this is a limitation of doing extreme acrobatics with a full feature autopilot. Doing full negative with full yaw for 3 seconds should be extremely unlikely however, even with the most advanced maneuvers, I think?