Servers by jDrones

A Tale of Two Quadplanes

(dazzab) #1

When a friend offered me a HobbyKing Dark Wing, I jumped at the opportunity to build my first quadplane. Being a bit smaller than an X8 and constructed of a fibreglass fuselage with balsa wings, it seemed like a great opportunity to try building something other than a foamy. It was an interesting build but it ended up weighing 6Kg which which was a bit heavy for it’s size. So armed with this experience I ordered a Skywalker X8 for part two of my quadplane adventures. If you need slow flying with plenty of lift and long flight time, then the X8 is a good choice.

There are a few build examples of X8 quadplanes from other members that I have watched with interest. The main design decision appears to be whether to add the vertical lift motors on bars attached directly to the wing or to attach the vertical motor support bar to an extra section of wing and insert it between the fuselage and existing wing. I’m sure both systems work fine but I opted for adding wing extension for a few reasons. Having the motor bars attached to a wing extension keeps the wiring attached during transport and the extra wing section adds about 15% more wing area resulting in a bit more lift.

Adding wing sections was by far the most time consuming part of the conversion process. In our first build we put the motor bars underneath the spars and attached them to the spars with small metal straps. This minimised the number of holes drilled in the tube which is always good. In the X8 build we took a much different approach and drilled large holes in the square carbon fibre tubes to allow the spars to pass through. We inserted square aluminium bars in to the carbon fibre tubes to join two lengths of tubes and to add strength where holes were drilled. Although this certainly reduces the strength of the tube, the results showed that the tube was actually much stronger than required.

We have gone through three techniques to make the actual wing extensions. For the DarkWing build we simply hand cut foam sourced from the recycling bin with a hack saw. It was fast and dirty but worked fine as a prototype. We used vinyl wrap to add strength and mask our sins. Needless to say, I don’t recommend this technique.

To make the wing extensions for the SkyWalker X8 we made a hot wire foam cutter and sourced EPP foam which worked quite well given it was our first attempt. Using templates attached to the sides of a foam block, it was fairly easy to trace the hot wire along the templates to create the desired wing shape. Cutting little slits in the bottom of the templates to allow the wire in to the centre of the block for cutting out spar cavities was much more challenging. For the next attempt we had the wing extensions cut on a CNC hot wire cutter by the company which supplied the EPP foam.

After attaching wing joiners to both sides of the wing extensions, threading the motor wires and embedding the square carbon fibre motor mount bars, the hard part was done. I should mention that in this design I’ve done something many will consider questionable. I have put 10mm carbon fibre tubes inside the wing extensions and then cut the normal 8mm spars included with the plane so that they extend only half way in to the 10mm tube. This allows for 8mm spars in the wings to enter the other side of the 10mm wing extension tube creating a join. Some may think that this weakens the spar but my tests show that it’s actually quite strong.

With this design you can even remove the wing extensions for travel if required although that will mean having to disconnect the vertical motor wires from the fuselage. It’s not necessary for every day travelling but might be handy if packing in to a case such as when shipping by air.

Once all the hardware was assembled then all we needed was to enable all the Q_Plane functions in Ardupilot. It was finally time for the fun part – flight tests. Results ranged from bad to catastrophic with the plane rolling over after take off at the first turn. Voltage levels and current draw were quite suspicious as well. We concluded that there was a possibility of a ground loop in the wiring when we noticed some very hot wires in the power board cable. During these tests the plane would stall because the turn would increase while not responding to commands for opposite roll. Then the vertical motors would engage which allowed for a controlled landing, saving the plane more than once. But after we changed a power module and a bit of wiring around we ran out of luck. Once again the plane stalled on it’s first turn and gravity won when the vertical motors did not engage. After a long walk of shame, that was the end of testing. Back to the bench…

We replaced some of the 14 gauge wiring with higher capacity wire and changed some power connections from XT60 to XT90 to ensure that the entire wiring harness was capable of handling 90 amps of current. When the vertical lift motors are in use and the forward motor kicks in to initiate forward flight there is quite a surge of current which needs to be taken in to account. It is possible to run seperate circuits for the forward motor and the vertical motors which is how the DarkWing built earlier was wired up, but I prefer having only one circuit for simplicity.

With our new wing extensions and vertical lift bars replaced (one was broken in the crash) it was back to the field for some testing, tuning and hopefully an automated mission or two.

Flight testing went very well. There were a few minor issues such as a transition that didn’t work well. We are looking in to those. So for now things are looking good and we are ready for endurance and auto mission test flights.

A big thank you to my good friend John for spending Saturdays with me in his man cave working all this out. Further thanks to the Canberra UAV team who I’m very lucky to fly with for all their help over the years. And let’s not forget a huge ThankYou to the ArduPilot community for the fantastic work that fuels all this fun!

Build components:

Skywalker X8 black frame kit
4 x 3510-13 700Kv TMotor motors
4 x HobbyKing 40amp Plush ESC
1 x HobbyKing SK3 40xx forward motor
1 x HobbyKing 85amp ESC
4 x 14” RC Timer props ( two CW and two CCW )
1 x 12x8 prop
1 x Pixhawk flight controller
Flight control software - Ardupilot 3.8b8
2 x RFD900 radio modems
1 x 5v BEC for Pixhawk servo rail and backup power
1 x HobbyKing HV power module
2 x 4S 5800maH batteries to power motors
1 x status LED / USB extension for external mounting
XT-90 and XT-60 power connectors
Volo LEDs for motor arms nav lights - optional
Modified Aeronavic multirotor motor mounts (no longer available – substitute as required)
4 x 15x15x500mm square carbon fibre tube – if you can source 1mtr length then no joins required
2 x 12.8x12.8x250mm aluminium square tube (used to join carbon fibre tubes)
TOW 5.1Kg
Build photos:


Skywalker X8



Skywalker X8

More Quadplane Fun
(rmackay9) #2

Nice! Really great that you’ve provided so much info on the setup!

(gmorph) #3

Great write up Darrell. Always great to see successful builds and flights of experimental aircraft.

Thanks, Grant.

(Greg Oakes) #4

Nice right up and excellent project outcome Darrell.

(Graham D) #5

Do you know what caused the sudden bank with yaw with the X8? And the hard landing at the end of the video? Run out of battery?

(darrell) #6

The hard landing was battery exhaustion. The sudden bank is a bit of a mystery however we did have similar issues in forward flight before a crash. There were indications of a short in the wiring and most of the wiring was changed in the rebuild.

(Dale Rogers) #7

Awesome video @darrell, I’m just disappointed I missed all the action.

(Perth UAV - Sam) #8

Nice! X8’s are a fine looking flying machine, especially with the QP attached!. It looks cold at CMAC.

I would be interested to see just how much current is required for hover and in transition. In our little 2.5kg frame we only put the power module with current sensing on the forward motor because previously it had browned out and cost us an airframe. It was one of the most interesting aircraft crashes I’ve seen, with it doing forward somersaults because only the rear motors came on afterwards in the air on it’s way down. We haven’t had an issue with current draw since, despite running up to 130A peak through a single XT60 on transition.

(darrell) #9

The HobbyKing HV power module is rated at 80 Amps. The standard 3DR is rated at 60amps. However, I have two batteries in parallel and only connect the power module to one leg of the circuit therefore it’s only getting half the current. Then I calibrate the current using a meter and scaling the reading as required. I measured 25Amps during forward flight, up to 60A in hover and spikes over 90A during transition.

(Perth UAV - Sam) #10

Thanks for the info Darrell.

I assume the 25A is for forward cruise around 70kmh or is it climbing? Also I’m not quite sure if the forward motor is only connected to one battery or both, and if those current readings are per pack seeing the current shunt is only on one pack? I think you should be able to improve the forward current a fair bit still. From memory we ran a Turnigy G46 550kV swinging a 14x8 Aeronaut tri-prop forward propeller at around 14-16A @ 80kmh on the X8 (no QP!) with 4.4kg TOW. It also climbed at +60degrees still once it got going after a bungee launch.

Up to 60A seams within the ballpark for hover for that type of airframe and weight, so say probably around 100A no climb hover in QP mode? Transition spikes are the killer in general, I still expect flames and lightning bolts to shoot out of the QP! Those are welding currents. Do you know roughly what the QP setup weighs extra including the wing extensions, and what your QP lift to weight ratio is? I’m always very interested in QP performance data! :wink: Thx

(darrell) #11

@JeffBloggs - thanks for the info. I’m hoping to get the forward flight current draw down. I did discover that the motor I was using had a bent shaft and couple of other issues. I’m assuming that was from the crash before the rebuild. So I’ve put a new motor on it which might help.

I use two 5800 4S batteries in parallel that both feed both the vertical and the forward motors. But the power module is only in line with one side of the parallel connection. I’ve used that technique in copters quite a bit and Tridge suggested I do that to keep the current running by the shunt down. It’s probably not quite as accurate but it works well.

(darrell) #12

Did a couple of flights yesterday. The current draw during forward flight cruising was 12amps. Rough calculations indicate about 25 min flight time with two 5800mAh 4S batteries. I’ll probably add a third battery which will comfortably give me the range/endurance I’m aiming for.

(Andrés Ábrego) #13

3 5800s and all the extra weight (escs and motors) is still OK for that platform? I think that its becoming quite heavy even increasing the surface with those wing extensions…

(dazzab) #14

It’s certainly getting heavy, I agree. But others fly it with four batteries successfully and get well over an hour flight time. I’ll be doing more test flights soon to find out for sure.

(Perth UAV - Sam) #15

Our experience with a non-QP X8 is that the weight is not an issue for flying itself, but it does increase you stall speed and resulting takeoff/landing speeds. The higher wing loading also increases your cruise speed, which in turn improves your range at the cost of endurance, if you optimise the motor/prop for the new cruise speed. I don’t think the wing extensions are necessary for the extra QP weight at all. Just need to watch out for wingtip flutter at over 100kmh if it’s not glassed/reinforced. :wink:

We used 17Ah of 18650 for a good 75min endurance on 4.4kg TOW. The only issue with a non-QP X8 is trying to get enough static thrust for a bungee launch (critical at those weights) from a prop that has it’s pitch optimised for cruise. At WOT the 14x8 prop would partially stall on takeoff, so getting it moving with a bungee was important.

Using a QP X8 means that you can actually mostly neglect designing the prop for static thrust for takeoff because the quad gives you a unlimited runway in the air. Adding 3rd battery pack will be fine so long the COG is right. Using some LiHV Multistars would help even more, or using a batch of the new 30A rated 20700 cells would work too.

(Alexander Perez) #16

@dazzab I noticed you are using RFD 900 radios under Arduplane 3.8. Did you experience any issues with your rfd? I’ve been using it without a problem in 3.7.1. Now using 3.8 beta 6, I am unable to connect to my RFD. Any pointers?

Thanks in advance for your feedback.

(darrell) #17

Do you have the newest firmware on them? I think it’s required.

(darrell) #18

During tests today I swapped the 12x8 prop for a 13x8.5 which seemed to improve the flight a bit. The current draw during forward flight is now just under 10 amps.

(Alexander Perez) #19

Thanks @darrell for your response. What version are you using? I believe the highest I saw was v1.13 (as regular Sik). There is a multipoint mpsik 2.6 for the rfd 900+ also available. Which are you using? Is there a specific site you acquire the latest fw? Do you have a copy of your latest fw you can provide? Thanks in advance.

(Perth UAV - Sam) #20

Hi Darrell
That sounds good! Roughly at what airspeed was that at? From memory the 8 pitch speed should be about right for 75kmh cruise on similar kV. Just be careful with prop stall at low forward speeds using high pitch props, so keep your quad assist on until at least 16m/s (60kmh) or so to be safe whilst testing and avoid steep turns to avoid a flat spin. Low static thrust/prop stall makes it harder to get it to push out of a flat spin, but you also have the quadchute still to recover. :slight_smile: (Having COG 5-10mm forwards helps too and altitude)

It’s well worth finding out the optimal cruise prop however, seeing most smaller RC plane props are only running around the 20% mark at WOT, and substantially less at 30-40% throttle. Just for example if your setup is only 15% total drivetrain eff. improving it by just 1.5% is a 10% range improvement at no extra cost apart from the correct prop.