gke's blog View Details
Posted by gke | Oct 19, 2020 @ 02:40 AM | 4,994 Views
This is a much lighter plank intended for indoors but works outside in relatively still air. Based on my own planks and planks of fellow flyers Pete and Ray. The rounded "look" came from Pete's version inspiring me to build. Took probably an hour maybe two to build.

AUW 47gm. Span of mainplane 450mm with 55mm tips at around 45deg. Chord 180mm with elevon chord 40mm. Fuselage 350mm with wing LE at 110mm. 820 motor with random 55mm quad prop. Any old Rx brick will do with nano servos. Will probably work with brick and integrated linear servos.

Motor down thrust needs to be about 5 degrees. and even then maybe some throttle to elevator mix with negative exponential. Reflex now 5mm. More tuning required.

Don't bother with KF mods. Does not like Wind which is not surprising given its weight.

There is a short paper attached by Jef Raskin (inventor of the Macintosh) suggesting that 4% under camber at around 40% gives best performance for thin airfoils at low Re. A short Bio for Jef is here:


I found Jef's paper only in the last day or so but it is what I have instinctively used for my Arado and Horten "indoor" models. Test flights, not surprisingly, showed the pitching moment of the under cambered wing competes with the elevon reflex in a complex way with airspeed. Add in engine thrust and it gets very interesting. Nonetheless it flies quite well once the dihedral tips are removed. All...Continue Reading
Posted by gke | Sep 17, 2020 @ 09:37 PM | 2,495 Views
Here are some key measurements of my PopCorn constant chord flying wing or plank. The unusual tail boom allows “parking” into the wind at high angles of attack without the tail fin being masked. This is useful on the slopes or when doing steep descents into tight landing spots.

The wing is 900mm span with a 250mm Chord. The wing has a simple 3mm spar of 3mm hard balsa inserted by simply sectioning the wing at 700mm from the LE and then sandwiching the spar. The airfoil is an EMX07 by Martin Lichte (use Profili). The foam is as light as you can get -i.e. packaging foam. The trailing edge is cut off and a 3mm stiffener added. The new trailing edge and full width elevons are made from standard TE stock with a 1.5-2mm sheet extension to follow the lower airfoil profile - the exact shape is not that critical but the closer to the original reflex the better. I used bottom tape hinging. The ends of the wing are capped with 2mm ply for a bit of protection.

The Skin is 0.75oz glass cloth finished with 3 coats of water based polyurethane. Wetting the wing then wrapping the cloth around and pegging it on the clothes line at the TE works fine. There is no LE but a woven packaging tape added after does offer some protection. Wing build time is very fast - less than an hour including drying time if it is a hot day.

The fuse extends with the same profile as the bottom of the airfoil to the TE. There is just the single nylon bolt holding it on. Ply plating at the bolt-on point is...Continue Reading
Posted by gke | Sep 17, 2020 @ 04:53 AM | 2,359 Views
I already have many many flying wings my favourites being constant chord planks. Apart from the criticality of their CG they are a delight to fly and are very forgiving as it is nearly impossible to stall them. It seems that when folk buy one they then feel the need to rudder control and a horizontal stabiliser - very odd.

The Reptile Dragon, a near constant chord plank, is a modern take on Dragon Eye UAV used by the US Marines.


In a moment of weakness/curiousity I bought a discounted kit as there seemed to be quite a few adverse comments and it did not make any sense to me as it should fly well although the wing is a little on the thick/draggy side. The assembly manual gives the CG as 48mm behind the LE which is WRONG. Somebody may have seen 38mm in a draft of the manual and said "that cannot be correct" and changed it to 48mm removing pretty much all of the stability margin.

Maiden powered flight 28 Oct 2020. CG at 38mm AUW with lead ballast and 2200mAH 3s pack is 1035gm. Trim was about 3mm of up. Motors cause pitch down on launch but manageable at about 75% throttle. Nice and stable in flight which was no surprise. Neutral trim gives causes a dive - not divergent but not recovering either. Maybe just a fraction nose heavy.

Overall OK but not as nice to fly as my scratch built Planks.

Vertical stabiliser area looks too small so may add another 30mm or so to the TE to increase yaw stability. A tiny touch of differential thrust as rudder is an option for those cross wind landings.

Posted by gke | Sep 03, 2020 @ 05:44 PM | 4,866 Views
This is another item of lockdown distraction. Good brain exercise.

One way of controlling an indoor model is to use differential thrust and no control surfaces. This is most often done using brushed motors. Purpose built controllers for this form of control are relatively rare but there is an easier way.

Rifle around in your junk box for a "retired" quad with brushed motors. Mount the control board flat on the aircraft pointing forward. Then take the left hand motors and mount them pointing forward a little way out on the left wing but as close together as possible. Ditto for the left hand side motors.

Magically every thing except roll control will cancel out it seems.

If you use one motor from each side then you will get throttle up when you pitch down and vice versa when you pitch up. Sort of a rudimentary height hold.

As you see below the idea works with a retired Hubsan X4 controller.

You can see and hear the effect of corrections on the pitch axis where the planks are very sensitive. If you use a "Tiny Whoop" class brushed motor flight controller, running say Betaflight or iNav, you can tune the control gains easily while using only two motors.

Not too hard really .

20200826 Differential Thrust Steering Using "Retired" Micro Quad FCs Part 2 (0 min 41 sec)

20200726 Differential Thrust Steering Using "Retired" Micro Quad FCs Part 1 (0 min 25 sec)
...Continue Reading
Posted by gke | Aug 27, 2020 @ 01:26 AM | 7,251 Views
See my Horton 1200 Blog. For this build I scaled down to 620m span. AUW was 60gm running a 7x20mm brushed motor with a 55mm prop. Unfortunately the fuselage masked the prop and the motor simply did not have enough oompah! I also failed to notice the prop was slipping at full throttle which did not help.

I have rebuilt it with no fuselage and no KF plating just a simple under camber. The 8mm, or so, under camber in the middle of the wing has the effect of introducing washout at the tips and so almost no reflex trim is required.

The CG is 110 back from the nose. The prop is 115mm back from the nose. It does have belly strakes which "may" aid directional stability and also give something to hang onto to launch. Elevons are almost full half span and have a 20mm chord - top hinged.

AUW is 37gm and for now it still uses the 7mm motor.

I roughly dialled my dimensions into this CG calculator. I didn't bother with the bat tail.


I had too much up elevon on the flight of my V2. Yes it is also slightly tail heavy so I will move the CG 2mm (yes 2mm) further forward to where the calculator said it should be! Flies quite well but the pilot is definitely out of practice with the lockdown.

20200912 Horten IX 620mm (1 min 6 sec)

20200903 Horten 229 V2 (0 min 36 sec)
...Continue Reading
Posted by gke | Aug 27, 2020 @ 01:00 AM | 7,001 Views
There is another Blog describing my poor experiences with the Nano Talon.

I could have bought a ZOHD Drift but decided to drag the Nano out of storage and modify it rather dramatically. The dihedral was removed by the simple expedient of connecting the wing joiner points with a carbon tube. The VTail mount on the tail boom was 3D printed. It still packs up into the original box.

The lockdown prevents me from doing a full flight test but the short test flights show it does fly. The slight crabbing was due to a trim offset in the VTail - fixed.

Flights were conducted using a ZOHD Kopilot.

Nano Talon V2 Re-maiden Kopilot Manual Mode (0 min 11 sec)

Nano Talon V2 Re-maiden Kopilot Stabilise (0 min 10 sec)
...Continue Reading
Posted by gke | Aug 27, 2020 @ 12:18 AM | 4,780 Views
There is lots stuff out there regarding the Horten Brother's work. Check out http://www.twitt.org/Farnborough.html or search for ROYAL AIRCRAFT ESTABLISHMENT, Farnborough Hants, THE HORTEN TAILLESS AIRCRAFT, K.G. Wilkinson, B.Sc. D.I.C., Technical Note. Aero 1703, October 1945.

This build uses the design of Bobdaeronort. His plans for the 60" version were scaled down using the tiling feature in Adobe to 75% or 45".

I did not have access to Depron so construction uses 4mm modelling foam with kraft brown paper internal lamination . The elevons are covered with white printer paper. The canopy can be tricky to bend into shape so I built a simple jig. I cut a slot in the inner foam lamination for servo cabling before adding the top and bottom KF sheeting.

Maidened successfully 12 Sept 2020 in the rain. Motor temporarily a junk box old Eachine 250 Racer motor (BG2204 2300KV equiv) with a 5x4.5x3 cyclone prop. Flight at under 50% throttle so quite adequate. AUW with 2200maH 3S pack was 570gm. CG 191mm (15%) to 219mm (5%) back from tip of LE. Flown at around 205mm or 10% requiring 8mm of up trim at the slow speeds I was flying to stay on the Oval. Will go to 5% when I get some space.

Landing a bit suspect - adrenalin.

20200912 Horton IX 1200mm Maiden (0 min 43 sec)

Bobdaeronort's more adventurous maiden in 2013:

Horten Ho-229 Nurflugel Maiden (5 min 30 sec)
...Continue Reading
Posted by gke | Nov 28, 2017 @ 03:29 AM | 3,589 Views
My Nano Talon arrived with a number of problems. Banggood has, after some correspondence, agreed to a refund of $40 or send a replacement wing set. I have zero confidence that a replacement wing set, in the typical packaging for single items, would arrive undamaged in the 2017 Christmas rush and so opted for the refund.

The Talon, although having some interesting design features, is an expensive model by today's standards and I agree with these price points as being more rational:


The faults and fixes so far:
  • the black foam impact resistant (much like the S800 Sky Shadow which you could hammer nails in with) but the white foam comprising 75% of the aircraft breaks quite easily.
  • The right wing latch was installed back to front.
  • The aileron torsion drivers bind overloading the servo and thus prevent the ailerons returning reliably to neutral - fixed use silicon or teflon lubricant.
  • The wing spar boxes in fuselage were not glued and resulted in the end plate tearing out of the fuselage - fixed by gluing the components together.
  • More seriously one wing is bent and has a wash-in twist and the other wing has a twisted aileron - boiling water perhaps or a diagonal slice in the wing undersurface and re-glue.

Other fixes/suggestions:
  • Tape the underside of fuselage to reduce rash to the weak white foam.
  • Tape over the fuselage wing end plates.
  • The square tube wing spars could have a CF rod inserted in the round centre hole to prevent crushing breakages at the wing boxes.

The stabiliser has an Arm processor and oh so easily could have been iNav/Betaflight capable giving it a marketing edge. It also does not support CPPM or SBus, both trivial additions. As it is I will be using a UAVXArm board.
Posted by gke | Jun 04, 2017 @ 04:47 AM | 5,688 Views
Intended to be used as a fixed wing test platform for UAVXArm.

Bought on sale as a kit from Banggood but in hindsight could have bought a foam only kit for even less as I have plenty of carbon tubing .

Used PVA for assembly but left wings unglued to make it a bit more portable. PVA tends to crack on high-speed crashes rather than tearing up the foam. PVA also tends to foam but can be contained using masking tape until it dries. Wings are fastened using woven fibreglass tape which lasts 1-2 seasons in our sunshine. In this case the tape has been fastened on the underside. Winglets are fastened using hot glue. Control horns are epoxied to the carbon rod in the elevons using a slightly oversized hole in the foam.

Motor is an AXN-2208-2150 which comes originally from an AXN Cloud and runs a 6x4.5" prop used on some of my multicopters. Current is around 19A and I am using a 30A Turnigy Plush ESC. Servos are HXT900 9gm and while I would have preferred MG one of the last two I had was DOA.

I am using a 1300 3s pack in the GoPro camera position which will give around 45min in still air with an AUW of 475gm. Predicted climb rate is around 9M/S. A 2200 3s pack fits in the main compartment but pushes the AUW to 560gm and only gives another 10 minutes of flight time. The VTX and camera roughly another 50gm.

CG set initially at 125mm as recommended but moved back to 135mm. This is just forward of the main spar currently but I will most likely go back to 140mm which is...Continue Reading
Posted by gke | Jan 25, 2017 @ 04:32 PM | 5,193 Views
There are lots of camera gimbals out there so this is directed to people who have older gimbals compatible with the BruGi gimbal controller software. The gimbal I bought a year or so back was throwaway cheap but had closed source code installed. I elected to use BruGi, the development of which stopped around 2014, and to add temperature compensation for the gyros.

My version which carries with it the usual GNU "all care and no responsibility" is here:


The original code is here:

Posted by gke | Jan 02, 2017 @ 01:10 AM | 6,313 Views
Time to build a plank with a bit of "Wick" so this is a build of the 60" Zipper put with a PW51 airfoil . The Zipper was originally Birdworks way way back with an EH 2-10 airfoil but for more than a decade it has been available as a part kit from Off The Edge Sailplanes in Oz. Yes it is an old design but then not all new things are that good.


Mine will probably not be quite as pretty as that on the Edge page my first build being WBPU and a few layers of light glass cloth. I will add a thin strip of carbon plate as a spar to help out the somewhat thinner glass skin. Glenn aka "Zipper" offered to cut me a set of balsa elevons as well which I will sand to profile if required as I will see if the WBPU covered elevons prove stiff enough. I will most likely commit sacrilege and add a motor . Sort of functioning ballast.

There are a number of build threads out there and the outline instructions are on the Edge pages:

Posted by gke | Dec 27, 2016 @ 05:41 AM | 5,056 Views
Some time ago I did a quick wiki on how to apply firmware updates to earlier FrSky receivers specifically allowing the D8R to output serial or CPPM. This wiki became lost when Google dropped their support for projects but may be found here:


The original link was:


This will redirect you to my gitHub page but the link does not resolve to the actual wiki.
Posted by gke | Oct 31, 2016 @ 04:45 PM | 5,391 Views
I have a Dream Flight Weasel which has not had much flying mainly because the slope winds are not always kind.


This proved to be the case recently and Ray mused "I wonder what it would take to add a motor?". He has a Weasel as well but he knew I would immediately take the plunge .

So I did a quick mod and added a BE1806-2300 with a plastic 5x3 prop borrowed from my 250 Racer spares along with a Turnigy 12A ESC and 1000mAH pack. The weight has increased by a shocking 15gm . Others, of course, have also done similar conversions if you search them down.

The static thrust of the motor/prop combination is 390gm just slightly above the AUW of 380gm so vertical is possible. Pack is good for around 15 minutes at full throttle and an hour at cruise. Air to the ESC behind the battery could be improved by removing top hatch or putting a tunnel through to the launch finger hole?

The changes to the fuselage are straightforward with the removal of the rear wall of the original battery box and a ply firewall braced backwards with triangular ply doublers - the foam is too squishy.

I could have just cut the nose off and end glued the firewall but I wanted to preserve the lines reasonably well. I used WeldBond white glue which takes a while to dry but is slightly flexible and adheres to the foam quite well.

Test flight confirms performance and certainly easily vertical. Two longish flights and only 270mAH from the pack. A rather bleak damp day but caught some thermals as well - surprising .
Posted by gke | Jun 24, 2016 @ 09:04 AM | 5,770 Views
I thought it was time to try a "pure" flying with no vertical surfaces. This was partially inspired by Karl Heinz's YB49 scale build.


As far as I can tell the XB had the contra rotating props and the YB had the simple pusher prop configuration at considerable loss in performance. I don't fancy buying four of the HK contra rotating drives so YB-35 it is.

The airfoils for the X/YB-35/49 were NACA 65,3-019 at the root and 65,3-018 at the tip. These are quite thick and draggy and close to symmetrical requiring up trim in practice. Their low Re performance is not good at model scales. So in my case I will deviate a little and probably use MHxx cores at around 9% as I don't have the task of squeezing EDFs into the wing as would be the case with a YB-49 . I intend using elevons only, mixing rudder in as elevon differential and possibly thrust vectoring on the outer motors for yaw control.

Scaled roughly 1:20 to 2.4M span to cut from a 1.2M wide foam slab which is manageable. I will probably borrow the 4 ancient Turnigy 2213/20 motors with 3 blade 8x6 props from my Lancaster. With the 3S 4900mAh pack it should have an AUW of around 2.0Kg.

I could also raid the Lancaster's retracts in which case AUW would be around 2.5Kg. Initially it will be minimalist and hand launched .

WingCalc wing plans:

http://j.mp/29Sjep4 YB-35

http://j.mp/2aubLNd Using MH62 cores to hand

MotoCalc gives stall at 6M/S and 8M/S at 43% throttle for cruise. This is roughly scale AS if my comps are correct. Endurance at cruise is around 45min and ROC 4 m/S.

eCalc also suggests climb rate and endurance should by OK even if not vertical which was also the case for the original.
Posted by gke | Apr 20, 2016 @ 08:05 PM | 7,851 Views
Commenced build of an Arado 555 based on that of Josef Poisinger's 1200mm version:


I have used an 8% Phönix airfoil instead of the original 12% to hopefully improve glide performance.

Build will be 3/4oz glass directly over foam cores using water based polyurethane (WBPU) which I have used before as a covering technique.

Propulsion will be either 2-4 EDFs or single propellor interchangeable on plug-in pylons.

Cores weigh 160gm bare without wing extensions which will be solid balsa - foam too hard to cut accurately. I am using lightweight filler to repair minor and one major (photos) hot wire cutting problems. The foam I am using sands well and I may put a light wash of filler over the wing.
Posted by gke | Apr 06, 2016 @ 05:26 AM | 7,054 Views
I could not find any plans but this NASA Delta Wing seems to be based on an idea from a NASA schools project to teach basic aerodynamics.


From what I can see on the video it uses a KFm2 foil but so far I have not found any other details so my dimensions are all guess work.

RC Depron Delta Nasa Glider Indoor Flight Faszination Modellbau 2014 *60fpsHD* (2 min 18 sec)

The video shows twin fins to clear the prop slot. Mine also has twin fins (because it looks good) but with a tractor prop and no slot (because it will fly better - probably ).

Initial glide tests are promising. Foam is ColorBoard from Diaso as for the Arado build. Electronics are from a WLToys PoleCat (F939) for now.

CG position calculated using WingCGCalc. Latest wing plan used is HERE. Elevons are 30mm at the root tapering to 20mm. CG 108mm from cutoff LE for twin vertical stabiliser version flies OK.

Modified version has original vertical stabilisers simply moved out to form winglets which give some dihedral and the CG moves to 114mm. The winglets are set at 45 deg.

Summary: OK once I added a bit of washout to the winglets but not all that fun to fly as it requires continuous control input . Retired.
Posted by gke | Mar 22, 2016 @ 01:13 AM | 8,108 Views
I have a larger Arado 555 build based on Josef Poisinger's 1200mm version underway but on a whim decided to build a mini version for indoors. It uses Josef's wing outline directly and is simply scaled 50% to a 600mm span.


First version used all of the electronics from a WLToys PoleCat (F939). The geared motor decided to hack into the wing at full power and so I replaced it with 2 Chaoli 7x20 motors running reversed Hubsan X4 props.


Notional thrust ~50gm with AUW of the aircraft, including 200mAh 1s battery and rather too much glue, is 54gm.


Foam is ColorBoard from Diaso. Elevons are inboard unlike Josef's elevons which are outboard, as they should be, but this proved to be too difficult/messy.

Video maiden flight outdoors at VARMS:

20160322 20160322 VARMS Arado 555 Mini (1 min 30 sec)

Indoor after a number of mods written up below.

20160420 Arado 555 Mini Flying Wing @ Waverley (1 min 6 sec)
...Continue Reading
Posted by gke | Sep 27, 2015 @ 06:23 AM | 11,363 Views
Flight control firmware for multicopters and fixed wing aircraft based hosted by an Arm32F4 with MPU6000, HMC5983 & MS5611 all SPI.

Posted by gke | Sep 21, 2015 @ 07:16 PM | 9,460 Views
Because it was reportedly impossible I have added RTH, or more correctly Turn To Home (TTH), to the last OpenPilot release 15.02.02 for CC & CC3D. Testing has been using EasyStar, Radian and Phantom Wing. It will probably work on multicopters as well.

The source modifications and firmware files are here:

Posted by gke | Jan 01, 2014 @ 07:18 PM | 11,029 Views
TxAdapter_gke is an open source standalone adapter translating CPPM signals where available from a transmitter, to Hubsan and FlySky signals.

Almost all of the adapter code was originally derived from the work of PhracturedBlue and others for the open source universal transmitter "Deviation". Inspiration for this development, which was intended for my own use, is due to Midelic.


Battery Monitoring

When used for the Hubsan X4 products flight battery voltages may be monitored with the throttle being progressively scaled back when a preset voltage is reached forcing a landing before the battery is destroyed. A buzzer may be added to give an audible warning.

MultiWii Telemetry

The Adapter produces MultiWii GUI compatible telemetry for attitude and battery voltage display. The only restriction is that for 8MHz processors the MultiWii GUI baud rate is limited to 56K.

Processor and Pin Assignments

The Adapter will work on most, if not all, Arduino compatible boards. It will also work on DIY Atmel based boards. Programming is by use of the Arduino development environment. Uploading programmes uses the processors UART, or where available USB, interface. A programmer is not required.

The default pin assignments are the same as those used by Midelic are detailed in the config.h file. Other parameters including unique ID, battery cutout voltage etc may also be changed in config.h.

Important Note that the resistive divider as pictured https://www.rcgroups.com/forums/show...9#post27180258 is incorrect. The divider should be wired to reduce the 5V PPM signal to 3.3V required by the Arduino. Thanks to guytzoler for noticing this.

Transceiver Module

The Adapter uses the A7105 transceiver. The maximum voltage on any of the A7105's pins is 3.3V and for this reason a 3.3V processor should be used.