After looking at kits, but never finding quite what I wanted, I decided it is time to design and build a bespoke plane for FPV thermal soaring.

Went for span of 1800 mm with a two piece wing. Just about right to pack in the car and have some performance. Wing construction will be traditional foam using hot wired cores. Each wing half is in 2 equal length pieces, since it is much easier to hot wire short spans accurately and also allows something approaching an elliptical planform for the wing. Airfoils will probably be based on Mark Drela series for DLG etc

Ailerons in 2 parts per wing so can do camber control and crow. Traditional tail as I reckon I can embed an antenna in the fin.

The motor is a classic pusher mounted behind the wing which works well for FPV and is easy to launch. The wing is on a pylon allowing the motor to sit in line with the wing to avoid having to angle the motor to get a good thrust line. The pylon allows a slimmer and more streamlined fuselage, though I will probably use a slab sided balsa fuselage, maybe with some triangular fillet so I can round off the corners.

The camera goes in the nose of the plane. I don't currently plan on pan tilt, but can always make a new fuselage pod if I decide to do pan tilt later.

The flight battery is planned to be a 2200 3S type since I have quite a few around.

As regards electronics, I plan to use a flight controller with RTL, with GPS, compass, airspeed, altimeter. I like an OSD and having plenty of instrumentation when I'm soaring

I used OpenSCAD for the initial sketch. If you are into programming I really recommend trying it out. Attached is the Openscad source in tranquility.txt. (rename to tranquility.scad to open in OpenSCAD)

Looks cool! Following with interest

Thanks Elmtree.

Mods as shown in the attached pic.

• Made the centre panel ( which splits in the centre for a 2 piece wing for transport) flat, which will make constructing the joiner easy.
• Decided on a V-tail after all. With the high wing a normal tail is very vulnerable on landing. Also a V-tail is very light compared to other types.

Now trying to figure out how to place of equipment to provide good separation while remaining tidy. The initial placement

• Vtx in the nose, just behind the camera
• I hope I can slide the ESC ( purple) in beside the battery, also need to allow some room for the leads and battery connectors, meanwhile making sure that the battery is easy to change without disturbing any wiring.
• Behind the battery is the flight controller/OSD. It looks quite big, but that is to allow spare room for connectors and wiring. The wing pylon will have to be able to come off to allow access. Not sure how that will work yet!
• RcRx in the back of the fuselage pod behind the FC. A FrSky D4R-II or similar.
• Poking out of the wing leading edge is the airspeed sensor. It may be affected by prop wash but the plan is that its main use is for TEC vario calculation in conjunction with the Baro so should be OK for soaring.
• Green object in the wing is GPS, possibly with a mag. There is a mag on the flight controller but it will be affected by currents around the flight controller with battery so close. Anyway GPS essential for RTL and for an antenna tracker!
• Small dark objects in wing and tail are the servos. 2 servos per wing for split control surfaces to allow for good camber control and crow braking. servos in the tail are heavy but makes pod less cluttered
• Finally in the tail is a 433 MHz telemetry modem. I hope to use a bare HopeRF HM-TRP module with a whip antenna for light weight.

.

EDIT. I put the scad file on github, as well as my initial thoughts on aerofoils in XFLR5
https://github.com/kwikius/tranquility

Still working away at this project. I decided I would like to try a built up wing, since it can be made much lighter than a foam wing.
I recently discovered a Free application that helps to design built up wings. It is by Curtis Olson who is also one of the lead developers on FlightGear.
The app is called madesigner and is available on Github.
https://github.com/clolsonus/madesigner/
There are some docs here
http://madesigner.flightgear.org/
Big warning. There are lots of bugs with it, however, after some tinkering I managed to get something useable.
Here is a shot of some madesigner output for Tranquility project. Not obvious but the wing sections are cut down for sheeting. The current plan is to use 5 mm x 0.5 mm carbon spars, for which cutouts should be just visible in the first pic.
The app also produces output for making wing ribs as shown in the second pic
EDIT: I added some more of the ribs. They get printed out on several sheets (Note that the numbering is wrong. I may try to do a bit of work on the app, as it does do quite a good job!)

What airfoil is that?

Originally Posted by Gratter

What airfoil is that?

root is MH32, chord of 180 mm, 0 degrees washout
middle is AG25, chord of 170 mm, 0.7 degrees washout
tip is AG27, chord of 120 mm , with 1.2 degrees washout.

(The madesigner app automatically blends wing sections together so the in-between sections are a hybrid)
Dont forget that 1mm has been shaved off top and bottom of the wing section in the pics for sheeting and cap strips.

Not all ribs are shown. I added another of the sheets. I think the app has got the numbering wrong. It has quite a few bugs, but looks adequate to get the job done!

I spent a bit of time trying to optimise the wing in XFLR5. The XFLR5 .wpa file is on github if anyone who wants to have a go at optimising it better:
https://github.com/kwikius/tranquility

It evolved. Now around 1M span and pretty light. Has ben a real pleasure going back to balsa. Having access to a laser cutter makes all the difference! I gave up on MAdesigner, it just doesnt have enough options and being Python is dog slow. I just worked in OpenSCAD to design the wing ribs and pod. Still not finished but flyable and flies great.

On the second flight on a nice hot day, nearly lost it in a thermal, so looking forward to getting the FC and FPV gear on it now!

NICE JOB! ! ! !

The design of the wing and tail has a LOT Of free flight style to it.

Originally Posted by BMatthews

NICE JOB! ! ! !

The design of the wing and tail has a LOT Of free flight style to it.

Thanks. I am really glad to have got back to good old balsa wood and balsa cement from mainly using Epoxy , foam and FRP. (I still have great nostalgia for my old Keilkraft Caprice, one of the first models that I managed to build straight and true without the wings twisting due to dope warping the structure ).

After initially using a very thin CA for gluing the first wing, and getting concerned about how much the balsa soaked up I decided to use balsa cement wherever possible half way through the second wing ( I also bought some higher viscosity CA glue). The second wing came out 3 grams lighter than the first due to this, still a bit heavier than I had hoped at 25 g bare, maybe due to the carbon spars.

I was initially planning to cover the structure with tissue and dope, but ran out of time/patience and used Solite.

With the reports that Solite is no longer being produced, maybe I will have no alternative than to use dope and tissue on the next one, which should give it an even more free flight look.

FPV gear installed. It fits, just about!

On my previous Lidl FPV plane, I didnt have a flight controller or OSD and I really missed the instrumentation like direction and distance form home and the ability to flick a switch and for the plane to fly home on its own.

For Tranquility soarer therefore, I decided I needed the same level of instrumentation as on my larger FPV models.

The flight controller can be seen at the rear of the hatch, or at least it can be detected from the connectors going into it.
The FC features a graphical OSD, a full deck of sensors.: gyro, accelerometer, baro, I2C. , SD card, analog to digital converter, 7 servo out. 5V out to power servos and several uarts. It was designed long and thin to fit this type of fuselage and in fact the fuselage pod was designed around it.
The Flight controller runs a custom flavour of ArduPilot.
It is in more described in more detail here:
https://discuss.ardupilot.org/t/ardu...chip-osd/11583

The RcRx is an FrSky D4R-II, quite old tech now. It can be seen sitting on top of the flight controller, covered in Velcro from a previous incarnation. At some point I hope to swap it for a newer FrSky RX, but for now, it just fits!

The VTx in the nose, ahead of the battery, is the Immersion RC Tramp 25 mW (EU) version.
The camera is a RunCam Micro Swift, which provides a really good picture quality and some advanced features in a small form factor.
Not visible, but nestled under the flight controller, is a 433 Mhz 3DR telemetry transmitter. This is used to transmit GPS info to an antenna tracker on the ground, so allowing a good range with the 25 mW VTx. The telemetry antenna comes out of the tail boom.
The battery is a 1000 mAh 3S Turnigy. I changed the connectors to the smaller JST style ones.
The motor is a Turnigy 2730 1500 kV with a 5.5 x 3.5 folding prop.
The ESC is a Turnigy 6 A esc, mounted inside the fusulage pylon. The Motor Prop combo draws about 5A at full bore, so just in spec. The Pylon acts as a chimney to cool the ESC.
There is also an airspeed sensor which can be seen protruding from the wing root.
Finally there is a motor current sensor hidden down in the fuselage pod.

As can be seen, despite the small size , the plane has a lot of functionality, and I hope the instrumentation of altitude, airspeed, artificial horizon etc, will come in useful for thermal hunting.

So far, the model has only been ground tested with the FPV gear , but all being well, I am now really looking forward to trying some micro sized FPV thermal soaring with the plane.

Thanks much for providing the openSCAD file - really fun to play around with.

Originally Posted by Achi

Thanks much for providing the openSCAD file - really fun to play around with.

Achi: No problem! To see how things look and how equipment can fit in, OpenSCAD works pretty good.

I got the plane in the air over the weekend with all the FPV gear in. See the attached screenshot.

Even before flying, the first thing I noticed was how easy it is to get out and fly, compared with my larger soarer with its 4 piece wings and 2 piece fuselage. This plane is small enough to put in the car in one piece.

The maiden FPV flight lasted around 35 minutes with plenty of battery left. The motor was just used to get to altitude then switch off and look for thermals. Climb rate is OK but not vertical. I found myself ascending a few times with power off but nothing dramatic, but it was early in the day. However everything worked pretty much as on my other planes including the antenna tracker and the return to launch capability. I am not quite sure that the airspeed calc is right yet though. I think it is reading much too low.

Here is some more info about the design of Tranquility V1.

The total wingspan tip to tip is 1088 mm
With all FPV gear on and a bit of tail trim ballast weight of 410 g
Wing loading of 10.5 oz/sq ft or 2.9 kg/m^2
Wing area of 0.142 m^2 ( or 1.52 sq ft)
Root chord of 162 mm. Root section is Mark Drela ag03
Tip chord of 121.5 mm. tip section (before the round wing tip) is Mark Drela ag11.
The semi span of the straight part of the wing is 432 mm.

The wing was designed so it can be built on a flat building board, without needing packing . Ease of building on a flat board was I believe part of Mark Drela's original concept for these sections. Despite being built on a flat board, the wing has some washout, since the root section has a bit more curve up at the underside at the leading edge, while the tip has very little, thus giving washout. The washout is intended to counteract tip stalling of course.

The wing was built in one piece and then the flaperons were cut out afterwards so that they follow the wing section nicely.

The wing control surfaces ( aileron/flaperon or what you will) are designed with a chord of a larger ratio of the wing chord at the wing root than the wing tip, so that as the flaps are deployed downwards, the angle of attack of the root increases more than the tip to counteract tendency for tip stall, when flying slowly with the flaps on. Hopefully This combined with washout will give benign handling at low speed!

Wing construction uses a 3 mm * 0.5 mm carbon strip with a web of 3/32 balsa. The joiner is a 6 mm carbon tube sliding in a aluminium tube set at an angle into the wing spar webs so that only a straight joiner is required. The ribs are laid up diagonally to make the wing torsionally strong.

The wing is mounted on a pylon, so the motor can be at the back and a camera stuck on the nose The pylon is mounted quite a long way back on the wing chord. This allows for a larger hatch and also should be better aerodynamically, since the leading edge flow is not disturbed so much and the pylon counteracts the reduction in size of the fuselage pod at that point. Not been tested in a wind tunnel though!

As to performance, it is early days yet to decided on the capabilities of this plane. I have only had one FPV flight so far, but as far as the overall concept of a fully featured micro FPV soarer is concerned, I am hooked and already thinking about V2.

Thats awesome, congrats. What will you change?

Originally Posted by Achi

Thats awesome, congrats. What will you change?

Very early days but inevitably buillding leads to thinking how to do things better next time.

Firstly weight or maybe more important wing loading.

With the FPV gear on the wing loading came out quite high for a thermal soarer, especially at this small size, so need to look at how to get things lighter.
If the dead weight cant be reduced then maybe stretch the wingpsan to 1.2 m. The wings are pretty light at around 25 g each and the carbon spar with web is very good in bending. It is easy to pick the plane up by the wingtips. Therefore stretching the wing wouldn't add much weight, but the small wingspan and overall compactness of this one is appealing!

Second the wing construction.

The design and look of the rib arrangement between the leading edge and the front spar was fun to make, but dont think it saves much weight over a D box. A D-box is better aerodynamically too, and so , at higher speed, there would get better glide angle and overall torsional stiffness with a D-box. It would also be easier to build. I'm glad I did this one as it is and I would keep the same construction to the rear of the main spar though.

Third the arrangement of the controls for the V-tail.

The tail control servos are forward in the plane wire linkages used to save weight at the tail, but I have ended up having to use 10 g of lead at the tail anyway to get the CG right. The control linkages along the tail boom and the servos sticking out the side of the fuselage are not that neat either. If the v-tail servos were in the tail, then the fuselage boom tube could be detachable behind the fuselage pod and this would make the plane able to fit in a smaller box for transport. (It is tricky making the boom detachable with the servos where they are, because of the problems of connecting the linkages when you put the plane together).

Caught a proper thermal yesterday evening. (It was a hot day so not that hard !)

Tranquility first thermal (2 min 1 sec)

Did you use openSCAD to generate laser-cutting files?

I really like the D-box arrangement of Vladimir's model's Elf- carbon molded d-box in a free-flight like construction method. Very time-consuming but rewarding if you make it. I hotwire a foam wing, and used that to make a D-box mold for a straight wing a few years back and it worked really nicely. Then I added an i-beamspar similar to yours (vertical balsa and spanwise uni-carbon, with balsa ribs going to the trailing edge spar.