DH90 Dragonfly after-build - RC Groups
Thread Tools
Feb 21, 2012, 10:25 AM
eye4wings's Avatar
Build Log

DH90 Dragonfly after-build

The DH90 Dragonfly is one of the most elegant aircraft of the late 1930s yet is very little recognised, most people assuming that it is a DH89 Dragon Rapide from which it was developed. At first glance this is understandable but there are a number of differences the first of which is that it is a smaller machine carrying only 5 including pilot.

One of the more significant changes was that the fuel tanks were moved from just behind the engines in the Rapide, where they had presented a fire risk, to the wing centre section which was thickened to accommodate them. Of the 67 made, only one Dragonfly was ever lost to fire.

Today only two examples remain airworthy, one in the UK (G-AEDU) and the other in New Zealand (ZK-AYR) a photo of each of which is posted below.

The Wikipedia page is here...
And photos of both aircraft are available on several sites across the web including this link on a full-size forum...

My purpose in starting this thread is to make available photos and detail that didn't make it into the article publishing my model design in RC Model World in December 2010 so that any of you interested in doing your own build to my drawings have the maximum of information to hand.

And of course if any of you are already building, so that you have a ready-made thread to post into.
Sign up now
to remove ads between posts
Feb 21, 2012, 10:53 AM
eye4wings's Avatar
My model's scale was decided by the fact that a friend was building a 96" Dragon Rapide and, having seen several photos of Rapide and Dragonfly in formation I decided that it would be a great idea if I designed a Dragonfly to the same scale so as to be able to recreate the pictures at model scale. His Rapide was a couple of months ahead of me so I designed mine and started building in an attempt to make up for lost time.
In the event I was test flying two years ago while his was still building... and it's not finished yet... Let's be realistic - it's not going to happen is it?

So the model was designed at 1:6 scale at a span of 86" and with wings permanently rigged ready to simply plug into the fuselage and be fixed with wire pins top and bottom. The fuselage access door was designed to be opening as a hatch to allow the batteries to sit on the cockpit floor, and in the event I devised a hook system to fix the top wings in place, one each side.
The wire pins to hold the lower wings in place were to be situated in the engine nacelles, which therefore needed hatches in the side.

In fact I have to confess to a deviation from scale that I had to adopt in order to have the wings fixed as a pair. The inboard struts, an inverted V shape, on the full-size connect at the top of the rise to the thickened wing centre section on the centre line of the nacelle, but I had to move them to the bottom of the rise so that they could be removed as part of the wing.

Confession is good for the soul! (Though you probably wouldn't have noticed anyway!)

The plans are available here...
Feb 21, 2012, 12:32 PM
eye4wings's Avatar
I've noticed here on the forum that faced with a new build, quite a few will start by building the tail feathers - tailplane and elevator, fin and rudder, before progressing to the 'meatier' bits.
So I'm probably going to surprise some of you by saying that as I don't have a proven design to work to and therefore have no certainty that the 'easy bits' will be quite as required when I get to the point of requiring them to add to the model, I prefer to jump straight in at the deep end and, so far as is foreseeable having produced only outline drawings at this stage, tackle the tricky bits first.
This I do on the basis that in the event that I cannot lick the hardest part, then I haven't wasted a truck-load of time and materials getting to the point of finding out.

So first job was the undercarriage.
"What?! Why? - you've got nothing to stick it to yet!"
Well, there's madness in my method!

The histories of both Rapide and Dragonfly show that one of the commonest accidents suffered by these types was ground looping, and faced with sudden side pressure the undercarriage was prone to collapsing sideways, letting the outside wing hit the ground and suffer damage as the aeroplane slewed round on its crumpled wheel. It could be a costly incident.
As I considered the design I realised that all that surrounding trouser which everybody else seemed to add on after the undercart was complete and working independantly was, apart from providing a little streamlining for the airflow, pretty much going along for the ride.
If I was going to carry it about then I decided that it would have to work for its living!
Another factor I had to consider was that my engineering skills are worryingly low. If I can't bash a bit of metal to shape with a hammer or cut it with tin-snips or drill it with a hand drill I'm pretty much stumped. So the simpler it is the better.

The Dragonfly's undercarriage could hardly be simpler from that point of view. Brass tube, cut by rolling on a flat surface under a knife and piano (music) wire cut by filing round and bashing back and forth held in the vice until it fractures is about all it takes. The right-angle joint between legs and axle is two thicknesses of tube half on the end of the legs then filed on the side with a rat-tail file until I can get a drill bit to stay in place then the axle can be inserted below the wire leg and ordinary electrician's solder run in to hold it still. Or epoxy could be used - but either of those methods don't risk melting the plastic of the wheel (if done with a moderately powerful iron.

A spring sits on the leg before insertion into the vertical tube and again there is a choice of two methods to hold the leg in place. Either a block can be glued onto the side of the trouser to stop the wheel falling out or one leg can be long enough to go right through its guide tube so that half an electrician's chop-block connector can be screwed onto it via a small hole in the side.
Feb 22, 2012, 09:00 AM
eye4wings's Avatar
The springs were chosen in my customary 'flying by the seat of the pants' method of squeezing between thumb and forefinger until I found some from the spares box that felt about right.
The theory is that each one should carry a quarter of the model's estimated AUW without too much compression so there is plenty left to take the landing shocks.

Once completed the spats were taken in hand and bashed onto the workbench to confirm in my mind that they would be up to a heavy shock.
Dropping the wheels out and replacing the springs is not a big deal anyway... a five-minute job.

Anyway having got that far I now needed to build the wing centre section to fix them to - so a few words of explanation for design choices made...

The power system was going to be a bit different from the last few models because this was to a smaller scale and, since I wanted to fly with plenty of capacity in the batteries, I did not think that I would have room to accommodate the packs in the nacelles with the ESCs and motors. Therefore I planned on using the cockpit floor for the batteries and, to keep power wiring short, housing the ESCs just behind the wing leading edge. For this purpose I planned to include a flattened out section of card tube as found in the middle of kitchen roll. This I hoped would be large enough to allow air around the RSCs to take away what little heat was built up. I chose a couple of HK 40A ESCs which would be more than adequate for my expected 25-30A draw.

In the event (and here is the advantage of posting the build AFTER completion and test flying - you can be warned right at the outset of any development work that altered the design) The ducting had to be cut away together with part of the wing skin under the centre section and the ESC cooling air taken from the airflow below the wing.

This is the centre section as it was originally built.
Feb 22, 2012, 09:38 AM
eye4wings's Avatar
Now I did say in the above post that that was the way the wing was originally built. In fact the spars were later strengthened with a 6mm x 12mm Cyparis top and bottom member for the main spar and another vertically inserted to assist the rear balsa spar on which the flaps hinge.

The maiden flight was not the happiest of occasions. In fact just about everything that could go wrong, did go wrong - but waited until lift-off to become clear.
On the climb out I realised that one of the aileron servos had stopped operating some time during the take-off run and that I could only turn the model left. On the downwind leg at about 50 feet altitude the model lost power and it was clear that I had no time to explore the aileron issue. Swinging onto the runway line I then could not stop the left bank and the model continued turning left arriving at ground level cross wind. I took a very untidy model home for repairs to one wing and the wing centre section, which had a nacelle hanging off it and was crumpled at the fuselage port side.

Now I mentioned above that I used Cyparis to strengthen the centre wing section - but this wood is not widely known by that name so I include a picture below so you can see the difference between it and its' nearest competitor, Bass.
Our Canadian buddies seem to recognise it as Port Orford Cedar but I was told that it is the timber from the Leylandii a fast-growing fir which has caused a fair bit of friction between neighbours over here.
I get my supplies from here...

Whatever its name where you are I cannot recommend it strongly enough.
As I have said in other places I will never use Spruce again. Even Bass has to give it best as it is not only light but straight-grained and flexible - all of which are most desirable for our models.
Feb 23, 2012, 12:31 PM
eye4wings's Avatar
Next stage is to finish the stub wings out to the nacelles. Nothing unusual about this, installing the tube ducts that will carry the ESCs and wiring between batteries and motors and then completing the sheet. I left the area clear within the fuselage so that servos and receiver could be carried as low as possible, their position would allow for seating to be installed at a later date with the gear under the seat (I never did it!).
A little trial and error took place in fitting the 1/8" balsa soft sheet that sloped down to transition the thickened centre section to the outer wing level. I fitted it to the lower rib first and glued it in leaving plenty at the top so that once the glue was dry I could just trim it to the line and sand it in nicely - well, fairly nicely.

Now to get on with the fuselage sides...
The sides are cut and sheets glued together from meduim 1/16" sheet balsa and a stringer of 1/8" x 1/4" medium balsa glued on top and bottom. Formers are economically cut from 3/32" medium sheet balsa and from the photos below you should be able to see that they are not quite flat. There is a very slight curvature which the part formers give the sheet - and this imparts quite a bit of rigidity to the sides, which will be drawn together later at the tail end. But for now let's just look at the joining on of the sides to the wing centre section.

I set up two of the side formers (one on each side) on the front spar and cut through the top sheet. These are raked forward and you will need to check the angle of these carefully - I didn't and had to do an adjustment later in the build to correct the mistake. Much easier to check it now.
To glue the sides on a strip is glued onto the top sheeting and the sides fitted and glued to this. I have only fitted one side at this stage so it can be seen how it is fitted.
The other side is fitted just the same way of course.
Feb 24, 2012, 12:58 PM
eye4wings's Avatar
That mistake I had to correct - it happened because I glued the sides on too far forward on the wing section. Consequently the tail was about 1.5" too high compared to the wing incidence. Can't have that - even if I did adjust the tail incidence to be right compared to the wing it would still look wrong. So I cut straight down from the top, through the top longerons and sheet sides leaving only the bottom longeron, then inserted a triangle of balsa and some extra sheet doubling the side. (see below).

Having checked that the fuselage sides were fixed vertical to the wing I then added the partition at the rear of the cabin to make sure it stayed that way - but leaving a gap at the bottom for all the control cables to pass though.

I also cut out the two 1/16" ply top wing faceplates and measured rather more carefully in view of the earlier mistake (more haste less speed) to make certain I glued them on in the correct positions.
I chose to cut out the door at that stage and set it aside for later. That sounds organised doesn't it? What it means is that I set it down elsewhere on the workbench where it could get buried under offcuts and balsa dust like everything else in the workshop and be found by recourse to sixth sense and sending sniffer dogs (fingers) into the dusty detritus.

Now to tidy up the nose...
Feb 24, 2012, 01:43 PM
eye4wings's Avatar
Now the eagle eyed will have noticed that some extra work has gone on in that last photo in that the cockpit floor has been installed. I took that slightly out of order because it is vital to make sure the centre of the fuselage is fixed so that building up the structure either forward or aft will not be able to pull it out of true.

So the nose...
The cockpit floor sits on a single spine member with slots cut to accept the bottom parts of the nose formers. While this is being assembled it can wag around a bit side to side, but the moment the 1/16" medium balsa sheet floor is added it becomes very much fixed so again check it is straight before the glue sets.
Some care is still required however as there is still the ability of the nose to twist.

Perhaps I should say something about my thinking on coping with the various parts of putting a model together. I know that many of you use proprietory jig systems which require a fuselage (for instance) to be built in halves, either with horizontal or vertical spines which then require the two parts to be joined. Or some designers build latticework sides and then, once these are assembled together add on curved formers to get the fuselage shape required. Often this is then skinned with sheet to which the finishing is then applied.
Over the years I realised that it suited my methods (and my meanness) better to start from the outside sheet and work in. This had the benefit of saving a lot of balsa that would have been used to support the sheet.
I served my 'apprenticeship' building kits like many did - and soon found myself thinking that the methods we have traditionally used, and therefore come to think of as 'right', actually resulted in us building in far more structure than is actually required to fly. Okay maybe when IC power ruled, but for virtually vibration-free electric power quite over the top. I therefore trimmed down all my structures by about 50% and found that the weight saved resulted in the model flying in a more scale manner. And when I had finished with a model (the Trislander was the first to suffer this abuse) the guy who bought the model from me promptly installed 4-stroke engines in it! It was, I thought, designed light for electric and I warned him that the tail motor would probably tear itself off the fin - but it didn't! I was STILL building too heavy!

So back to the nose... Once the bottom section of the formers are fixed it is simple enough to glue side sections onto them and the top sections onto those, then the sides can be glued on and held in place with rubber bands. At that stage no more torsional movement.
Anyway, time to show you the cockpit floor in construction.
Feb 25, 2012, 04:19 PM
eye4wings's Avatar
So nothing particularly unexpected about pulling the nose into shape. Once the vertical sections of the formers are set up (glued face to face onto the bottom sections through slots cut in the floor), and the top parts added the sides can be pulled together and glued onto the formers held either by pins or rubber bands (I recycle those the postman drops all over the streets - they're poor quality but, provided they haven't been out in the rain they last a few months (did I mention I suffer with chronic meanness?)) I just hate to see wastage!.

Unless there are particular reasons for not sheeting some part of the structure as it's building I do so - I just like the encouragement the steady emergence of the shape of the model gives. So the nose is sheeted - soft 1/16" balsa - of course and then on to the next tricky bit - the cockpit roof.

It took me quite a while peering at the available photos, estimating the positions of the glazing and trying to figure out how the structure must have gone on the big one, and I am aware that the drawings are a little difficult to decipher, so a few photos should make the intentions clear.

For the top skinning I have a rummage around for all the offcuts of soft sheet left over from earlier processes then choose those that are long enough to plank over the double curvature. That's why I work in a mess of offcuts and balsa dust - even quite small bits of balsa can still be found a useful job to do - if they can be found at all of course!.
Feb 25, 2012, 06:27 PM
Gravity sucks.
mrittinger's Avatar
Very nice work, excellent craftsmanship!
Feb 26, 2012, 03:49 AM
eye4wings's Avatar
Thank you Mark.

Maybe not quite a 'LOL' moment, but gave me a wry smile. My temperament is such as to have me always dashing ahead with things... always after the quick effect! Trouble is I find myself getting myself into problems and always seeming to have to find solutions to get myself out of them.
Maybe that's part of 'craftsmanship'?

I was once called a 'Balsa Wizard' - by a non-modeller ... I didn't know that was what wizards were made of, never having knowingly met one!

Thanks again. All comments welcome!
Feb 26, 2012, 07:57 AM
Registered User
Bill Smudge's Avatar
Wonderful work Robin!!

Watching with keen interest. Certainly not a "standard" fuse building technique..

Feb 26, 2012, 09:19 AM
Big gov never Works
St. Martin's Avatar
Robin, that is beautiful work. The Rapide has always been on my to-do list for yrs.

Feb 26, 2012, 11:53 AM
eye4wings's Avatar
Ah Fuzz, thankyou, but you might just have fallen into the category of my first statement... the one about the Dragonfly often getting mistaken for a Dragon Rapide?
Perhaps I could encourage you to stand out from the crown a little and buy my Dragonfly plan instead?!

And thank you too Bill, a little lateral thinking goes a long way!
Like I said before, if you're going to be cutting out lightening holes (a waste of material that took years growing) why put the material there in the first place?
Just part of my general modelling philosophy! Maybe I'm just a bit rebellious and see no particularly good reason for 'following the crowd'?
Feb 26, 2012, 12:46 PM
eye4wings's Avatar
Time for another stage I think... let's look at the power source and do a bit more to the nacelles before the airframe puts on much more weight (and size).

I don't recall doing any calculations for the Dragonfly. It was just one of a line, moving from one model build to another you get to know what will work and what won't. We all have our ways of doing things, our favourite motors and so on. Mine was set on the present route by some TowerPro motors on sale at 20 apiece - for which money I could have either the TP3520 or the TP2915, 600W and 400W respectively, either of which would happily swing a scale prop for the kind of model that I enjoyed flying - 30s to 60s civilian types.
Since this model was smaller than previous ones that had used the larger of the two motors I simply decided that I would use the smaller, higher revving version on 2S and see how it went.
These days there must be dozens of motors that are in the same category - notable the Turnigy offerings that are even cheaper. And to my mind there is no particularly compelling reason for paying more... you could spend 300% more - but I can guarantee that you'll not get 300% more performance! Diminishing returns!

So the nacelles were simply built forward from the wheel spat units, which offered a very solid foundation on which to build a floor of 1/4" medium balsa. This is the chin of the nacelles and has a large hole at the rear to allow air to pass down the trouser and out past the wheel.
To this is added the 1/4" 5-ply motor mount and the two softwood bearers that brace it back to the main airframe. Again I used Cyparis as I had it ready to hand, but ordinary softwood from your local DIY store would do the job - and be a lot cheaper! (there I go again!).

Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Yippee! 2 X 6 first flight after 11 year build grog85361 Sailplane Talk 9 Jan 04, 2012 10:16 PM
Discussion A little after the fact but why don't folks build arms out of CF? spacecase Multirotor Drone Talk 23 Sep 21, 2011 11:01 PM
Discussion After Jan. 2005, there will be no more GWS Dragonfly with EM300 in GWS new production GWS4CEO GWS (Grand Wing Servo) 37 Jul 05, 2005 08:28 PM
Dragonfly main frame twist after assembly? meatbomber Coaxial Helicopters 1 Apr 01, 2004 07:54 PM