• MR60

    @Frantz, thank you. I did not fly it in "real conditions", meaning far enough to judge if the antennas placement is ok or not. What I did do , as you can see on the pictures, is to separate the three antennas as much as possible, taking advantage of that long H platform : from left to right : 1.3Ghz video fpv, 2.4Ghz radio receiver, 433Mhz telemetry.

    Also the GPS module is put as far as possible from the APM.

    I did notice on another build a possible (not confirmed, would require more tests) interference on the GPS reception by the 1.3 Ghz antenna. As I unplugged the 1.3Ghz antenna I had no more GPS fix losses...

    (GPS operates in a close band at 1.2Ghz, so there could be some interference).


    @Criro, thank you. I did replace my landing gear this afternoon with the 3DR I reused and foam swimming noodles as shown here :

    I have to say it is quite practical for hard landings and much much more stable.




    As a conclusion of my first small flights with it:

    I am still very disatisfied with the flex of the middle piece, even with the two carbon tubes I added.

    I have to design a new frame, radically more rigid. I will be thinking of a H frame but with a X for motor arms(cross design) , as Frantz advised, separated in their middle (X intersection) with a spacer.

    I will post later about this new design I have in mind.

  • MR60

    End of my build, finally !!:

    Hey guys,

    I finally finished my first H frame build. I post some pictures below of the last progress :









    I did a couple of take-offs just to see the reaction of the frame. So far I'm happy with the incredible flight stability. Much more stable than the X frame (based on 3DR i had). I am very happy with the room that frame gives to place everything : camera, fpv, batteries, antennas, etc

    I'm not yet happy about : the take off has to be corrected with some stick movement because the geometry of the frame is not perfect and moved after I painted it. This is the biggest difficulty about building its own frame with home tools (without CNC machines and the like) : the adjustments are not perfect. I'm not happy either with the current landing gear as on the picture. It is too high and too instable when rough landing. It made me tip over ! So I am at the moment modifying this to re use the 3DR quad landing gear.


  • Hi all, what are the forces that act on the motor-arms / booms? The ones I can think of that are generated by the motors / props:

    1) lift generated by the props

    2) torque counterreacting the drag of prop-blades

    3) precession if the copter moves the props (acting as gyros) out of plane

    What else? Did I miss something (it's late night here...)?

    Now for designing a copter it would be nice to better understand the magnitude of these effects. I'd be happy to get links / infos to mathematical formulas, but to start easy some simple assumptions could make it more approachable. I mean, for a typical quad, what ball-park numbers are we talking about and what effects to we really need to consider in practice.

    This is a beautiful thread, community at its best and as it seems ending up in a better code for us all. Cool. Love it.

  • Developer

    Graham Dyer pointed me towards this discussion and it was a very simple change so I've gone ahead and added the "H" frame type to master.  The commit is here.  This change will go live with AC 3.0 which is a couple of weeks away I think.

    To give it a try you'll need the latest code and you'll need to set the FRAME parameter to "3".

    Thanks for the contribution!

  • MR60

    H frame built, next step:


    In order to have an almost perfectly rigid frame, I decided to use a light carbon tube.

    The picture shows a first piece of carbon tube that is being glued with epoxy to the frame. A second piece will be glued on the other side making an X with this first one. I first hollowed a U shape in the wood arms to receive the tip of the tube. Then I poured epoxy (slow dry) on it. I hope it will hold, I will keep you posted on the result.

  • MR60

    Follow-up of my build:

    Here how it looks with the Aluminium arms screwed on (top center plate is just screwed, not glued):


    I cut the wood arms shorter to save weight.

    And then I added some small wood "retainers" on top and bottom of the wood arms to fit the aluminium arms snugly:



    Next I will begin to drill holes for cables management, especially from the motors to the ESCs.


    Coming back on Ulf's explanation about inversing the motor's rotation and adapting the code : can you explain how the twisting creates an air flow that counter acts the rotation ? In fact on your two drawing 1 and 2, I do not understand how is the air flow generated from the twisting (intuitively I would have drawn the air flow arrows exactly in the opposite direction than your drawings, I must be missing something...)
  • The H-frame is different from X-frame, Arducopter needs modification to support it!!!

    The H-frame is different from the X-frame, if the frame is not perfectly rigid.

    Fortunately, it is easy to fix by changing all motor directions and a small modification of the Arducopter code.

    The reason is that the twisting of the frame results in airflows sideways that can either amplify or cancel the desired yaw rotation. Unfortunately, the motor directions for X-frame setup will work against the desired rotation. I strongly recommend anyone using an H-frame to change all motor rotations and modify the Arducopter code!

    * Using original X-frame setup (not recommended) is shown in picture 1.

    * The solution is shown in picture 2.

    * The code modifications is shown in picture 3. (Simply change AP_MOTORS_MATRIX_MOTOR_CCW to AP_MOTORS_MATRIX_MOTOR_CW and vice versa.)

    I have experienced this myself in practice. First my H-frame was impossible to control in yaw, and two opposite motors were spinning up very fast. Then I started thinking about the twisting of the frame. My H-frame works great now, after changing all motor directions and the Arducopter code. My H-frame is described here:

    The beauty of this is that the frame need not be rigid. Twisting of the frame is actually good for the yaw rotation! I would suggest that the guys developing the Arducopter code adds the H-frame option.




  • MR60


    I decided to start building my own H Frame. Based on all previous discussions about h frame and experience from others, my requirements are:

    -The frame must be totally square (position of motors must be spacially exactly in same spot as an X frame)

    -The frame must be as rigid as possible with no flex, but with the constraint of weight

    -The motor arms of the frame must be removable for easy transportation

    -The frame must absorb vibration as much as possible

    -The frame shall allow plenty of space for FPV and aerial photography

    -The frame must be easy to build with basic materials and cost no more than 50$


    WIth these requirement I looked around in the house. I had 4 used/spare 3DR aluminium arms. I like them for their rigidity and their vibration absorption. I went to the hobby shop to buy 2 wood booms 25x10 mm and a plank of poplar to buildas the central "box". These two would booms will be screwed to the aluminium arm tips.

    As i want something easy to dissassemble/re-assemble, i do not want to mess with screw nuts/bolts and screws. So the idea is to integrate the screw bolts within the wood arms permanently with epoxy resin. That way, I just need for assembling/dissassembling to screw two screws per arm.

    First step of the construction was to build the arms. Two holes per arm were drilled. On the inside face of the arm I drilled a larger hole to accomodate the metal bolt. Then I put epoxy all around it and let it dry.

    This looks like that:


    I repeated the operation 4 times for each wood arm.

    The 3DR aluminium arms will be screwed on them like this with two screws:


    Then I carefully marked the center of the arms, to start building the center plates : 80mm long x 10mm thick

    The positionning and marking of the arms and first bottom plate took lots of precision and time to have it perfectly aligned and square.

    They are being glued at the moment:


    The beans cans are useful at something, at least.

    The bottom center plate is not onmy glued but also screwed to the wood arms.

    The upper plate will only be screwed in order to easily access the inside of the center box. That will give plenty of space for cables and electronics.


    I am now stuck because have to wait everything dries, but will continue this post as building steps progess.

    (I have also two carbon tubes I am thinking of using along side the center plates to make it even more rigid.)


    more later...





  • I've read several of the responses here and I've got to kick in here.

    It is certainly true that none of the (arms) are truly rigid, in fact, there is so much variance in rigidity in both X and H frames that the point is totally moot.

    For our purposes using our flight controllers, the designation of X or H is totally meaningless, it is simply a description of where the motors are (on the corners or on the ends). X or +.

    Individual variances in airframes, weight, power and construction materials make much more difference than the X or H layout.

    That is why the PIDs are tune-able, stop sweating the whole X or H thing, for this class of vehicles, it is quite simply irrelevant.

    Build either one you want and call it an X, it will be perfectly happy.

    Generally H frames weigh slightly more, but their cargo carrying characteristics often more than compensate.

    The only significant difference is if you make your design non symmetrical (long and narrow or wide and short).

    But all that really means is that you might want to tune the roll and pitch PIDs separately instead of locked, that's it, nothing complicated - easy in fact.

  • Take a look at this also, folding arms. I don't have it, it is pretty new.

    H-Frame HK


This reply was deleted.