A Depron Seagull

[Simon Chaddock]

Plane Print are offering STL files to print an RC EDF Seagull with a 1500 mm span.

 

A truly impressive bit of CAD design and intended to be substantially printed in light weight 'foaming' PLA (LW-PLA). There are videos that show it flying well on just a 40 mm EDF.

The EDF version is expected to weigh 560g. It uses ailevators and a rudder for aerodynamic control. 

 

To me despite the use of LW-PLA that seemed a bit on the heavy side particularly for a 40mm EDF. I wondered what could done using a built up structure from thin Depron sheet.

Following my usual EDF practise the intention would be that only the duct would be printed so I started to create a 3 view drawing using MS Paint.

It currently looks like this.

No quite the elegant curves possible with printing, somewhat smaller at 1240mm span and a bit more of an 'aeroplane' using ailerons, elevator and a fixed fin.

Only a guess at this stage but it is possible the all up weight could come to within 250g. 😲  

However drawing is one thing but converting that into a load bearing super light structure is quite another.

This could take some time.

 

  

    

[Simon Chaddock]

As the inlet and exhaust ducts together with the EDF will be by far largest items in fuselage/body it seemed logical to get them built to make sure they would all fit.

The Plane Print Seagull has a rather unusual inlet arrangement with two wing root inlets coupled with another in the top hatch. These feed into what is in effect a plenum chamber. The EDF with its bell mouth draws its air from that chamber.

My own experience, particularly with small or low power EDFs, suggests the inlet duct geometry is just as important, if not more so, than the exhaust tube.

My own mantra is "whatever the air is doing disturb it as little as possible". This means a constant and sufficient duct area with as few changes in direction as possible.

To achieve a single inlet that would not detract too much from the 'visuals' I decided it had to be on the top just ahead of the wing. It helped both the inlet path and the exhaust outlet if the EDF was slightly inclined.

 

   

To keep the exhaust duct with its high speed airflow as short as possible I favour an "EDF at the back" principle with a large area inlet at least 1.2 times the FSA.

Having searched around I decided to really go for light weight and ordered a QX 30 mm EDF. It  only weighs 21g and delivers a claimed 220g on a 3s. This should be quite sufficient if the Depron Seagull comes in at under 250g. 

 Again it is one thing to draw and hypothesise but quite another to actually do it!😟

Edited October 19, 2022 by Simon Chaddock

[Simon Chaddock]

Given that the Plane Print Seagull is intended to be printed in LW-PLA I decided to order some to see what it was like to use and how it might be useful in the ducts for my Depron Seagull.

To my surprise it actually works rather nicely even with the maximum printing temperature (210 C?) imposed by the standard PTFE filament guide on the hot end of my Anet A8. 210C is at the bottom end of the recommended printing temperature so would limit the degree of foaming possible.

As the bead foams as it leaves the nozzle for a particular bead width and layer depth a much lower flow rate is used compared to normal PLA.

I happened to have a 'spare' PLA printed nose cone from my Skyray to hand so I printed an identical one on LW_PLA to give a direct comparison. 

Apart from the considerable weight reduction the other aspect that impressed me was the smooth surface finish. Of course LW-PLA is nothing like as strong but it has a significant flexibility so recovers from a deformation that would split PLA. A further less obvious advantage is the subsequent foam layer sticks very well such that it has virtually the same strength across the bead layers as along them, unlike PLA which displays a substantial "grain" effect.

Anyway this encouraged me to consider using LW-PLA for the Seagull's ducts. With such a light EDF the weight of ducting could easily become significant.

The first trial LW-PLA exhaust tube compared to an identical one in PLA. 

 LW-PLA weighs 1.73g. In PLA 3.6g.

Note that although both are printed with a single wall and with the same nozzle, the wall thickness in PLA is 0.4 mm but for LW-PLA it is 0.7 mm. Both thicker and lighter!😀

The test exhaust reprinted to exactly match the 30 mm EDF  

As the exhaust tube is under a positive pressure when in operation any strength limitations of LW-PLA will not be an issue.

The inlet duct on the other hand might have to rely on the direct support provided by any fuselage formers that it passed through.

The installation of the complete final duct could be quite tricky.

    

    

Edited October 20, 2022 by Simon Chaddock

[Stephen Jones]

When it comes to making flying birds this guy has it sorted just have a look at his videos all fly with no vertical fin just like a real bird.

 

 

Steve

[Simon Chaddock]

Very interesting but I note he has to use an outboard reflex trailing edge. This is the sort of thing the Horton brother used for their big full size glider flying wings.

It works but the margin of stability is much reduced.

 

My current thinking for the duct, although it may change, looks like this.

It does fit within the 'body' and is printed in LW-PLA. It weighs 5.4g so the complete duct, including the EDF, comes in at 27g.

It will glued together as a one piece unit so will fully support the EDF.

The duct does blow some air.

https://www.youtube.com/watch?v=09-ARDOgT0A

Just over 9A and 96 W on a 3s. 

It doesn't blow the door shut but then it is only a 30 mm EDF!    

[Simon Chaddock]

Planking of the body half shell complete.

 

All done in 2mm Depron.

Although a single straight through duct has advantages it does take up an awful lot of the available internal volume. Add to this the fact a flush top inlet would provide very little dynamic pressure from the plane's forward motion.

A bifurcated wing root inlet could be a better option although such a major design change is not ideal having actually started construction!

The necessary inlet shape would likely be more complex that I could realistically hope to print so it would have to be done via rather fiddly 'internal' 2m Depron planking. The position of the whole EDF & duct would change a bit so all the existing former cut outs would have to be altered.

Oh! well, nothing ventured nothing gained.  

[Simon Chaddock]

Opened up and enlarged the wing root inlet but before the duct itself could be planked an inlet disc was printed and fixed to the appropriate former to give something to actually plank to.

 

The inlet is as big as possible and still retain a reasonable inlet path.

I also printed a simple test stand to measure the EDF thrust to select the best design of thrust tube.

 As expected the greatest thrust is delivered from a bare EDF. The tube on the right is a constant 32 mm diameter of the inside of the EDF body. it the restricts to 28 mm dia equal to the FSA. The cut off section is 28 mm diameter.

The tube on the right has no restriction so is a constant 32 mm diameter.

The RH tube only delivers 72% of the bare thrust figure. The LH tube delivers a more respectable 92% so this is the exhaust tube that will be used.

The bifurcated inlet although quite generous at 1.2 times the FSA will likely cost some static thrust but without seriously spoiling the Seagull body outline there is not much I can do. I will just have to live whatever the resulting figure is.😉  

Edited October 28, 2022 by Simon Chaddock

[GrumpyGnome]

I find all your builds fascinating... keep 'em coming!

[Simon Chaddock]

Having established the most efficient exhaust tube the next problem was to fit it in the half body.

 

Suite a long winded trial fit & repeat process as the duct has to be 'submerged' exactly half diameter in with each former cut out being progressively sanded down so it fully supported the duct. Eventually satisfied it could be glued in place.

Normally the next step would be to add the other half of the formers and fully complete the body planking but leaving or actually cutting openings to install the electronics. However in this case the inlet duct planking requires access from both side so the other half of the body will have to be completed as a stand alone half and just hope the two halves match up!

Had I fully thought through the consequences of using planked bifurcated inlet I might have stuck with using the single printed top inlet!.

It is easy to be wise after the event but nevertheless building a 'stand alone' other half of the body is next.

   

[Simon Chaddock]

The two halves of the body shell.

It required a dummy former to be used to ensure the inner end of the inlet duct matched the inlet already built.

The two halves taped together after more gentle former cut out sanding so the two halves not only matched up but also fully supported the exhaust duct.

It will have to remain just taped together until the ESC, Rx and battery positions are established to achieve the CofG, wherever that needs to be!

 

[Simon Chaddock]

Obviously the wings are next.

All 2 mm Depron. My judgment is they will be plenty strong enough without any spar as such just a shear web between the ribs and rely on the skin to provide all the strength.

The ribs are printed. Single sided with 3mm diagonal bracing but in LW-PLA as an experiment.

.

Only the root rib is drawn in CAD all the others are % scaled in CURA.

The ribs are not flat bottomed so the lower skin has to be formed by hand manipulation to just over the required curvature so when glued with a bit of weight on the rib it makes contact over its full length.

It all take a bit of time but once all the ribs are glued in individual Depron shear webs are glued in between the ribs.

Once they are firm a bit of careful sanding to get them flush with the ribs and the manually formed top skin can be glued on. Again a multiplicity of small weighs ensure the skin is in full contact with all the ribs.

After 24 hours to dry it can be handled so a 3 mm Depron leading edge can be glued on.

Then it is just a case of "repeat" thee more times..

The four wing panels laid out against the body.

A bit stuck at this point as nothing can actually be assembled until the 3.7g aileron servo arrive and are installed in the wing outer panels to allow the servo wires to be run through the wing.  

       

 

[Simon Chaddock]

Just propped up but it does give an idea of how the gull wing will look.

  

Both wings and the body including the EDF inside together weighs 90g.😀

I propose to use a 450 mAh 3s at 43g.

Then there will be three 3.7g servos, the ESC and the Rx as well as the tail feathers and nose still to make but it should still be on course for sub 250g.

 

Of course how well it might fly is another matter! 😉 

[Simon Chaddock]

Whilst waiting for the servo I have started to install the RC gear. Its a tight fit!

The 450 ,Ah 3s battery position is guesstimated to achieve a CofG in line with the leading edge of the fan. There fortunately is some freedom as where to place the elevator servo which should give a reasonable CodG adjustment without having to add weight.

The final headache the wings will have to be glued to the body halves and all the servo wires installed before the body halves can actually be glued together! The battery will be the only item accessible once it is done. 😲 

Edited November 8, 2022 by Simon Chaddock

[Simon Chaddock]

With one servo available one wing can be completed and glued to the half body.

The crude adjustable jig is to make sure the other wing will have the same "gull" angle. 😉

The aileron servo is flush with the wing underside. To achieve a strong aileron differential the servo arm is angled forward 45 degrees with the servo at mid position. This will give about 3 times more 'up' travel than 'down'.

The servo wire was run through the two parts of the wing before they were glued together and the same for the wing to body. There is no wing reinforcement.  

With just the tailfeathers and nose/beak to complete the best estimate is 190g ready to go.😀

[Simon Chaddock]

The two halves are only held together with masking tape but it gives an idea of what it will look like.

No tail feathers yet. They are quite large so hopefully the tail moment will adequate. 

 

A crude balance with the 450 mAh 3s in suggests the elevator servo needs to be placed towards the rear end of the fuselage.

 

Edited November 18, 2022 by Simon Chaddock

[Simon Chaddock]

The elevator servo sandwiched between the tail pipe and the skin.

 

Everything, even the tail feathers, has to be installed so the two halves can be glued together as a final operation. Any subsequent alteration will require an area of the body skin to be cut away. As with many of my builds all the pictures are important to remind me exactly where everything actually is! 😲

After much soul searching I took the plunge and glued it together and left it clamped for 24 hours so now it looks like this.

 

Still the nose and battery hatch to do.

The span is 1210 mm. As above it weighs 130g.

The 450 mAh 3s LiPo adds another 43 g so maybe 185g ready to go?   

Edited November 20, 2022 by Simon Chaddock

[MattyB]

Looks good Simon - presumably you will test glide it over longish grass initially to finalise the CG before trying a powered flight?

[Simon Chaddock]

A test glide is always a toss up with super light planes. There might only be one arrival without damage!   

If the CofG is in a reasonable forward(ish) position it may be better to risk a powered flight as you can have a bit more time to obtain some sort of control even if it is only to 'influence' the landing.

 

I used my recently acquired lofting experience to print the nose/beak.

 

LW-PLA and hollow with a double wall. It has to be hollow as the Rx aerial is inside.

More a learning exercise than a necessity as it could have easily been carved/sanded from foam or balsa.

Getting anything like the required shape can involve quite a bit of trial and error as the lofting equations have their own ideas of what shape best links up the points. 😟

For  instance the Seagull nose is actually made up of seven different sized ellipses and a separate nose piece.

It even takes a couple of prints before the result 'looks right' on the Seagull.

The scrap prints so far.

Fortunately each only uses 1/3 m of filament and weigh less than 1g so really only time is needed.

Battery hatch next, some light sanding and then some thin paint.

A Seagull can have black areas on the wing tips which is likely a good idea for an RC plane.

White under and pale grey topside is a pretty good camouflage when viewed from below against the average UK sky. 

 

  

[MattyB]

35 minutes ago, Simon Chaddock said:

A test glide is always a toss up with super light planes. There might only be one arrival without damage!   

If the CofG is in a reasonable forward(ish) position it may be better to risk a powered flight as you can have a bit more time to obtain some sort of control even if it is only to 'influence' the landing.

 

Fair point. The other option is to "fly" it in your hand with someone else operating the elevator to ascertain the elevator trim and pitch response. If it's very sensitive to small movements it shows the CG is likely to be a bit far back, though it's not a truly scientific test.

 

I have done this with plank flying wings on the slope with some success, though the really important tweak for them is to have a mix that gives instant elevator trim on the throttle stick. This means you don't have to fly around with it out of trim for anything more than a second or two, and one flick of a switch then sets the current elevator trim as he new neutral (easily done in the open source TX firmwares). Maybe something you could think about for the first flight - you could have one flight mode with the throttle on a shoulder slider and elevator trim on the throttle stick for launch and initial trimming, and another with a conventional assignment for once it is trimmed out? Given it's a relatively standard/traditional aerodynamic layout though this might be overkill.

Edited November 22, 2022 by MattyB

[Simon Chaddock]

MattyB

Unfortunately your throttle/trim suggestion is a bit beyond my simple Tx but I do appreciate the trim adjustment problem with digital trim buttons.

On more than one occasion I have been glad of my old 35 mHz Futaba that has a slider trims so you can put it all on in one go!

I do fully expect that with its limited elevator tail 'volume' it will be pretty pitch sensitive. I do normally fit a stab receiver specifically for a maiden as it will act faster than I ever can but the Depron Seagull simply does not have the internal space.

A little voice in my head keeps telling me that the fin should be much bigger to ensure lateral stability with its short tail and those big wings. I may add a big fin extension specifically for the maiden

I have set the battery position so the CofG is a bit forward of 25% chord so it may initially require positive 'up' elevator immediately after launch but hopefully its very low wing loading and a reasonable thrust to weight mean things will happen slowly enough for me to respond.

 

Looking at the weather forecast, wet and windy for several weeks, I don't think I will be trying to fly it anytime soon.

 

[Simon Chaddock]

A bit of paint.

Just a 'wash' brush coat of much thinned match pot emulsion. Dove grey and white.

 

I think the dog's expression is asking "What are you doing? I was asleep!"

[Torsten Spitzner]

That looks the part. Am keen to hear how she flies..

[Max Z]

On 22/11/2022 at 15:13, MattyB said:

the really important tweak for them is to have a mix that gives instant elevator trim on the throttle stick. This means you don't have to fly around with it out of trim for anything more than a second or two, and one flick of a switch then sets the current elevator trim as he new neutral (easily done in the open source TX firmwares)

Matty, can you give a link to an explanation on how this is done? I am not unfamiliar with openTX, but would appreciate some guidance.

 

@Simon ChaddockNice! I do have a box full of blue foam bits that should build into a Gull-like glider sitting around for decennia, maybe I should give it a try....

[Shaun Walsh]

For true realism it needs a servo operated pump that dispenses white toothpaste from an orifice at the bottom.

[Simon Chaddock]

Given the rather horrible weather for even attempting to fly such a thing I have done a bit of decorating.

It now has a proper Seagull 'hooked' beak and its a bit creepy but that is a picture on an actual seagull's left eye.

And thanks to edit flip horizontal there is RH version on the other side 

Having a rather novel no spar lightweight structure I did my usual wing tip test just to be sure. If in a ready to fly condition it can be lifted by its extreme wingtips it gives a bending moment at the wing root equivalent to pulling about a 4g manoeuvre.

It flexes a bit but nowhere near to failure.

With a 1210 mm span it is a bit smaller than the 1500 mm span of the Plane Print Seagull but at 181 g it is only 1/3 the weight. 😉

A nose down thrust test on the scales showed 105 g. Should be enough, just.