Ardupilot Mega Twinstar Build


Here’s a build log for my Ardupilot Mega Twinstar II. Hopefully the included photos will be useful to anyone trying to decide which airframe to buy and wants to get an idea of the space in this model.

Overall, I am very happy with the Twinstar, my build didn’t require many modifications to the standard kit and the instructions in the box are pretty good. I’ve upgraded it to a brushless setup with a 3 cell lipo for power as it is cheap and sooo much better. I’ve included a full parts list at the bottom.

The great thing about the twinstar is it comes with an APM sized hole on the CG! (which by the way is a good thing for the IMU).

Okay, so a little bit of shaving is required on the left hand fuselage half. And as I later discovered, its best if you put a straight pin header on the airspeed sensor connector, not the right angled one shown here. Otherwise you can’t pull it out between the wing bolts so easily.


This next picture shows my chosen component layout in the fuselage. I’ve gone for servos in the standard location for simplicity. The receiver I am going to put in the rear as the servo connectors on the APM must point backwards, and I have a switching mode power regulator just below the rear wing bolt.


If I want to change the receiver in the future, I’ll have to cut it out, but I imagine I can patch it up again no problem. If I need to rebind the receiver that’s also covered, as I put a y-lead on the battery connector, one end going out the side of the fuselage, the other end to the voltage regulator.

I cut out a hole for the servo wires from the receiver and then hot glued them in place with the APM in the fuselage, being careful not to get any glue on the APM itself. This means that the APM can be simply pushed into the servo connectors and all 8 inputs are wired up in one go. I made my servo wires to fit, and in retrospect, I would have bought some 8 long female headers instead of servo connectors for the APM end. I also cut a slot out for the USB connector.


On the other fuselage half, a hole for the output servos must be cut. Here the rudder servo wire can be seen passing by. The wire is embedded in the foam, so as not to interfere with the underside of the board, and then appears through the newly cut gap. When installing the APM, that wire is going to get pushed back into the compartment behind.


The pitot tubes go in the nose, so the airspeed indicator should probably go somewhere near by. Here’s my choice. I have put velcro hook and loop tape on the sensor so if it needs re-plumbing when I end up bending those tubes, its easier to remove and put back. For the connector, I used a 90 degree header, but soldered it at an angle, so that the output is pretty much in line with the fus.


I then cut a v shaped channel in the fuselage half for the tubes, tacked them in place with CA, and then offered the other fuselage half up, squeezed them slightly, leaving an imprint of the pitot tube location. Then I cut a corresponding v.


With everything in place the halves are now joined. A piece of card can be glued down below where the APM will sit, to stop it making contact with the Elapor foam.

Here is a picture showing the string that I’ve attached to the board to help pull it out of the hot glued servo connectors. As it is a snug fit after all :) When it is in place, take a look at the two cutouts on the fuselage in-sides where the tailplane snakes are installed. These provide excellent ‘rails’ for sliding back a piece of card to above the APM, providing protection from the wires coming from the wing.


The GPS unit goes nicely in the canopy, which essentially provides it with a massive amount of protective foam. I didn’t feel the need to have it protruding out the top either, I’m sure the foam will not interfere with the signal. Oh, and it’s best to remove the 4 pin header (only required for FTDI connection if yours wasn’t pre-programmed) or you’ll end up spiking a wire or battery!


The motor supplied is about the same diameter as my selected brushless motor, only longer and 20g heavier. So the mount requires shortening. Being an outrunner, the outside of the brushless motor rotates, and so clearance between it and the flanges on the mount is more critical- so these will require filing down a little bit too.


I’ve installed the motor mounts facing the opposite way to the way shown in the manual such that the load from the motor isn’t pulling on those four feeble looking self tappers that hold on the mounting plate.

I wanted the ESC to be in the airflow to give it a good amount of cooling, but didn’t want it to ruin the aerodynamics of the wing. So I opted to cut into the motor nacelle slightly and CA the ESC in place such that the smooth side is flush with the surface of the wing. The only real protrusion into the airflow is the capacitor.


The wires (all of which I had to extend) can then be run into the foam as per the manual, before gluing on the spar cover. As there are quite a few wires running down here, I tacked them in place with some CA to help alignment and minimise trimming of the cover.


I also broke the connection on the middle red wire coming from the ESC to disable the included BEC. This means that I don’t have to use the less efficient linear BEC and as an added bonus, can power on without the wings attached if testing.

I used a smaller servo than the recommended, but with enough glue I believe it won’t move around within that over-sized cut out. Also, the decals can be used later to smooth the wing where there is now a big gap.


Finally a power wire needs to be made with 3.5mm gold plugs for the battery end, positive male, negative female. The other end is connected in parallel to two (female) deans connectors for the ESCs as well as a single JST (male) connector for the radio power supply (the switching mode voltage regulator or “UBEC”). The lipo is held in place with velcro, for which I bought the branded stuff, the kit supplied didn’t appear to be all that strong. I originally went for counter rotating props in this, but my APC props were so out of balance that I had to put some cheapo GWS ones on until my blade balancer arrives.

I took a lot of pictures during the build, so more are available upon request. Here’s the part list, which after postage, without the transmitter and receiver comes to around £500 (in the UK). Also the take off weight is 1.3 kg

AirframeMultiplex Twinstar II1
Motors1300kv 28mm by 30mm 50gr outrunner2
ESC30A unbranded2
UBEC3A Switching mode unbranded1
ServosEmax ES08D 8gr digital servos4
Lipo3000mAh 3 cell Loong Max Tipple1
Servo wire6 metres is plenty!1
Servo connectors10 pack male1
Power connectors2x Dean pairs, 3.5mm gold, JST male1
GPS receiveruBlox1
GPS adaptordiydrones adaptor1
GPS cable30cm is more than long enough1
Pitot tubesBought spares!2
Airspeed sensorMPXV7002DP1
XplaneVersion 91
Transmitter + ReceiverSpektrum DX7 + AR7000 - this may change when I add telemetry1

I have successfully flown this a handful of times now. It flies superbly- very stable with plenty of climb out performance. I haven't worked out the flight time yet as I've been tinkering a lot, but I reckon it could comfortably cruise for 15-20 minutes.

The stock settings for the APM work very well indeed. As do the default FBW-B speeds- the minimum of 6m/s looks very slow and I was worried about it stalling at first, but it had no problem in keeping its height and attitude rock solid :)
E-mail me when people leave their comments –

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

Join diydrones


  • how i have learned the static tube must be in one of the sides of the fuselage and the air speed tube should be on the front in the airflow ..
  • @Peter looks like top tube is the static tube. you can see the tiny holes in the side, and the end should be capped. the lower tube looks like the pitot tube (hole in the end)

    My question is on the tubes as well. I thought rule of thumb was twice the width of the fuselage away from the aircraft, what made you decide to put them so close?
  • I now see that the tubes look different :-) that might explain it.. no hole in the top of one ?
  • Very nice ! But are the airspeed sensor correct mounted ? Am i wrong if i think only one tube should be on the front in the airflow and that the other should be not in the airflow, and the speed is the difference in pressure between the tubes ? Keep up the good work !
  • Admin

    I used the stock Multiplex nacelle portion of the motor mount, but substituted a Micro Dan aluminum firewall for the stock plastic firewall. The aluminum firewall has cooling vents and drops the motor center 1/16" lower than the stock mount. This provides more clearence between the motor and the nacelle ribs. Micro Dan has indicated that he is interested in making an all aluminum motor mount for the Twinstar II once he gets his present backlog of machining completed.

  • great job ...
  • Developer
    Good build log, I like your methods, For a good non 2.4 ghz radio I use a JR XP7202, It uses same programming as DX-7 on 72 mhz I tape my servo connectors together as a unit, I saves time! have fun flying!
  • Very neat setup! Well done!
This reply was deleted.