Ducted fan theory and practice

Tony,

Yes. It is Schubelers biggest and is intended to replace mid-sized turbines, but less than half the price including the built in motor and the ESC. I have seen it flying a large Ziroli Panther very well. Only arrived a week ago. Bought sufficient ThunderPower Lipos Monday to make up the recommended 14 cell 8000 mA pack. Ran it, and yesterday we measured it at the lab from which I recently retired..

I intend it for a T-33 that I shall scratch build from Balsa and ply. Can't afford these expensive composite jet ARTFs and don't like just buying a ready made plane. Keep me occupied through the winter.

Won't be any bigger or heavier than all these 100cc Extras! AeroNaut and others do similar fans. I have built a few 90mm models, and have one under construction at the moment, but around 4ft span they are a bit toy like and not really any better than a good foamie. I like small planes and will build more, though I have others up to OS 120 size, but somehow the small jets are too fast and twitchy to be convincing.

PS: It's all Richard Sharman's fault Seeing him flyinh his little ones  so well a couple of weeks ago kicked me off into getting on with the half-completed 90mm one (I have one flying already) and ordering this fan.

Edited By Mark Powell 2 on 15/08/2012 13:36:21

I have had a look at Schuebeler's website, those big units are very impressive.

Your figures seem to fit the theory. By my calculations, the output power is in the region of 7,5 kW so the efficiency is nearer 80%. This you could expect to be reduced with duct losses.

If a 14 cell lipo gives you about 52v, then 9,4 kW will draw about 180 amps. What size cables do you need to carry that current?

BTW, I used to race 100cc karts, I couldn't imagine bolting one of those engines to a model aeroplane!

Tony,

your figures are near enough spot on, The Voltage does reduce to that level on load. The wires are about 1/4 inch dia of copper, plus the insulation of course. The battery is the same. Schubeler fit the 6mm bullet connectrer between the motor and the controller.

No wonder dsome have called these thing 'flying arc welders'. Evem my Wemotec 90mm fans (I have and use two) drew 80 amps or so when I had not got the motor rpm/volt quite right.

!00cc cart engines? I wonder sometimes if we are going a bit over the top. A guy on another thread was tlhking of fitting a 33cc petrol engine to an ARTF 67 inch span Extra. My ancient Flair Astro Hog, 72 inch span, nylon and tissue covered, crashed and repaired a couple of times, therefore heavy, with a modern OS91 FX two stroke, (the one I mentioned in my earlier comparison comments) will go OOS vertically in a few seconds Why a 33cc engine? I'm over the top too. The original 1958 or 9 world championship winning Astro Hog did it all on a K&B 35.

The composite ARTF jets are crazily heavy too. The retract UC, wheels, and brakes alone on the twin turbine BVM Phantom weigh five pounds. Thats why I scratch build my EDFs. Cost, too, of course. The Ziroli Panther I saw, with this fan, weighs 28lbs. Nuts. Nearly bought the TopFlite Cessna 310 twin ARTF, for two 50-70 four strokes. until I saw the weight in their brochure. 81 inch/20 lbs. Crazy.

Edited By Mark Powell 2 on 16/08/2012 19:54:15

Mark, I wouldn't say that, what I am saying is that no matter how the wing itself actually works, the force that is experienced by the wing has to result from accelerating the air downwards.

For instance there was a device being played with many years back that used an array of downwards facing needle points and a lightweight grid of wires below them. You apply a really high voltage between the points and the grid, and as a result the air is accelerated downwards. If the whole thing has been built light enough, it will actually generate enough thrust to lift itself, although the examples I recall could not actually lift the required power supply, they had to have light wires attached. So accelerating air downwards is all you need to do to generate lift.

With actual wings, well, they obviously do accelerate air downwards, even at zero angle of attack with a flat bottom section like a Clark Y or similar.

John

John

Clark Y. I had a self built Spirit of St Loius with a Clarke Y section at a zero angle of attack, the real one also had a Clark Y with a zero angle of attack. But it is only relateded to a more or less arbitray datum line, often the engine thrust line. Few real aircraft have 'decalage' like some models, the term seems to be unknown in full size practice (I have never heard it mentioned). The aircraft actually flies with some positive AoA. (nose up if it isero zero). On my EDF models (I have only made a few) I use symettrical sections on both the wing and tail and set them both at 2 degrees. Keeps the fuselage more level in upright flight and thus looks better. Spils the inverted slightly, but I can live with at.

Any many modellers (not us of course!) put a Clarke Y on with the reference as the flat bottom. That is wrong of course, it has a considerable built in positive AoA if you do that. A Clark Y, at least in model sizes, is as near as makes no difference, a symmetrical section performance wise.

Not sure about the 'elecric' lift. It sounds very much like an ion rocket, which are sometimes used in satellite stabilisation. The working fluid, air in our case, is ionised and the ions are attracted to the negative pole, in this case the grid. Some of it goes straight through.The electrons are bled off from the needles, and travel down a wire with a heated end , below the grid,s o they evaporate off and rejoin the ions. Otherwise the ions would curve round and go straight back to the grid and generate no thrust. Never heard of anyone trying it it anything other than a vacuum before.

.

Does it really matter.

Most. if not all, are now taught circulation theory.

I myself did not get to grips with the idea. The plot starts with the wing stationary. It is then moved forward through a fluid. There is a starting vortex which is generated by/from the wing, which is linked to the wing by tip vortices.

Examination shows that there is circulation of forces. These are considered as vectors, when added together show an upward force. Viola flight is achieved.

Now I had trouble visualising this, as well showing that the vectors did do what the lecturer and book said. Never mind the encyclopaedia of maths that can be generated by the ideas.

As for Bernoulli, momentum and any other theory you like, if it satisfies you, well, great. As for me, as long as my models continue to fly, all is well with the modern world.

This is BEB territory, which he choose to deal with in a cursory manner in his recent RCM&E article, providing a sketch of the ideas and concepts. So I am guessing he thinks it is a little heavy and involved for a general discussion.

With nearly all these topics, at engineering level, the first thing to ask yourself i, how good is my calculus, as no scientist can resist bringing in very quickly the technique.

Edited By Erfolg on 17/08/2012 12:38:25

Erflog,

No it doesn't. A lot of it is 'fashion'. In current full size practice pupils are now often taught 'constant aspect' on a landing approach. You look out of the front at the point you want to land. You keep it only getting bigger and doing nothing else. If you are doing it properly the 'aspect' will remain low on the windscreen as you are slow and have the nose up. Keep it there and you can near enough forget airspeed, it will not reduce to a danger point.You will be fine. You flare, of course, at the end.

Simple and obvious. But it took a hundred years to catch on.

Erlog,

agree on the 'theory'. Model aircraft are amazingly uncritical. You can get away with almost anything and they still fly well. Current fashion on competition free flight,where perforance is everything, is the 'new' (HoHo) 'low drag' airfoils. As if no one had every thought of low drag before and as if our construction techiques and covering methods did not make the final airfoil wildly different from the nice computer calculated one.

'Dynamic Soaring'. New! Wonderful! Just Invented! What a clever fellow I am! Albatrosses have been doing it for millions of years.

Erm, what has approach patterns got to do with the ideas with respect to how lift is generated. Or is it a joke.

Mark

"It took hundreds of years to catch on"

Just how many centuries have we been doing powered aircraft approaches?

'Dynamic Soaring'

Surely the issue is not how long the Albatross has been doing it but how recently we (modellers) have discoverd just how to do it.

And finally on EDFs.

Does the air down stream of the fan have the same velocity across the full section of the duct? I suspect not. Lower towards the centre, higher towards to the wall?

Does the air approaching the duct have a constant velocity across the duct. I suspect it does, apart from the wall boundary layer.

Conclusion? The losses per unit length are higher for the exhaust duct than the inlet so for maximum efficiency for a given duct length put the fan right at the back with no exhaust tube at all.

Hmmmm.

Simon,

You misquote, then criticise.

"It took a hundred years to catch on"

Afraid I can't think of a smartass answer about albatrosses

Except "What's the difference bewtween albatrosses and albatrossi?"

"Albatrossi are more literate."

You could be correct about the duct. The old IC DFs usually had the fan above the wing. Most EDFs have it further to the rar, but it is largely incidental to offset the weifght of the batteries. My Panther's rear duct is only four inches from the rear of the fan to the outlet. With a small convergence.. And the  flat rear of the motor i s only a half inch from the end of the duct, so the flow is more annular rather than a circle. But it wasn't 'designed' that way. It just happened.

Edited By Mark Powell 2 on 17/08/2012 20:18:45

Edited By Mark Powell 2 on 17/08/2012 20:20:08

It's all Richard Sharman's fault Seeing him flyinh his little ones... So, it's all MY fault !!??? Well, I'm delighted that people are still discussing this topic! Long may it continue.

The plane I was flying that Mark saw when we met on the flying field recently was this one:

which is the Max-Thrust GR4 Tornado (with working swing wing). I got this going for fun, and to try out the swing wing concept before developing a larger example. It flies extremely well considering it's only a lump of foam - very fast, takes off easily, not too hard to land, etc. The swung back wing does allow the model to speed up a little, but makes the controls rather less responsive. The swung forward wing is used for take-off and landing with some success.

On the other hand, there is a lot wrong with the model - the intake lip is rectangular, not smooth; the ducting generally is rough; the swing mechanism obscures the intake duct; the outlet duct bifurcates poorly, etc. The basic geometry of intake size to fan size to exhaust size is good though.

Basic figures for the model are: 450 Watts, all up weight 390g (=2 lbs) so watts/lb is good.

Richard

PS. "circulation theory" for lift on a wing in my book. But, a recent Channel 4 program on the Colditz glider escape story repeated the "bernoulli theory" with very elaborate CGI animation diagrams - what a pity!

I just stumbled across this thread & thank you all for the fascinating reading. Yes, the Scharnhorst Paper (in particular Fig 3 on Page12 ) summarizes the inlet & outlet area debate very well for me. The link to the Speed 600 power systems of the time with a 1% calculation accuracy is close enough for me as well.

The Germans are deeply into EDF & publish a lot on their Websites. Wemotec recommends a 10cm long 75mm diameter nozzle as ideal for its 90mm fan, which goes very much in Scharnhorst’s direction.

When we only had low power GWS EDF units available the first trick was to smooth up the inlet ducting and use a yoghurt pot to reduce the outlet area. Also long exit ducts waste power. Even with the frugal power available my Stealth fighter flew that way.

For practical purposes I offer some rules of thumb I have picked up along the way:

1. Aerodynamically clean models fly at about 2/3 of efflux velocity
2. Lower wing loaded models (airliners) can fly at around 100W/lb
3. For a “fighter jet” model you need more like 150W/lb

For me point 1) above means choose a reasonable efflux velocity as drag is pushing the other way at the square of the airspeed so there is no point in burning energy to create excessive efflux velocities. That should help optimize overall system efficiency.

The Watts per pound numbers look OK as Scharnhorst nicely calculates that even a well designed EDF system is going to max out with 85% efficiency. So if you knock 15% off the Watts per pound numbers in points 2) & 3) above you end up in familiar territory for conventional models.

To answer Richard’s initial questions:

* choose the best EDF unit for a given model

* design a new model around a particular EDF unit

* check the performance of the model and EDF unit is as it should be

* improve the performance of a given model set up

My crude design rules are:

• Decide (or calculate) your model’s flying speed
• Add 50% to get the efflux velocity
• Choose the smallest fan that will give you the right efflux velocity for your model while burning the right Watts/lb as suggested above.
• Round the inlet duct nicely & keep the exit duct within Wemotec’s (or similar) recommended dimensions.

You have to hunt a bit to find efflux velocities so I would love to have a good source of well tabulated results if anyone knows one, an old example is here : **LINK**

Thanks in advance
Mark
PS I like airliner EDF models

Richard

Love the Tornado and I am glad it flies well.

I have to ask where did the info quoted below come from?

" Basic figures for the model are: 450 Watts, all up weight 390g (=2 lbs) so watts/lb is good."

Normal conversion used is 1 kg = 2.2046lbs

Andy,

Glad you like the Tornado. The figures I quoted are MY measurements on MY tornado just before takeoff. It should have said 930g (=approx 2 lbs) but I made a typo, sorry. Therefore Watts/lb in excess of 200 -- not bad ?

I am using an Overlander 2200mAh 60C 4s Lipo which is quite heavy for its size, but does withstand the 33 Amp current draw which is sustained. I have used the supplied servos, wheels, fan, motor, esc etc with a Spectrum 6200 Rx.

The question is: why does it fly so well ? Considering it's only foam, has standard components, poor aerofoil section, and lot's of other technical deficiencies, (presumably to keep the cost down) it shouldn't be as good as it is. But it takes off quickly, flies smoothly in up to 20mph winds, and lands predictably.

Richard

"Considering its only foam......"!

I digress so returning to 'fan at the rear'.

As I was building in foam, so light for its size, I tested a ducted coarse pitch (4.5x4.5) prop in a duct and found a very obvious difference between the inlet and exhaust air velocity.

Obviously in a parallel duct this could only be the result of a variation in airflow across the section of the duct so the prop, mounted as a pusher, was positioned right at the back, flush with the duct end.

Of coarse I cannot prove this is is a more efficient location unless I also reposition it in the body of the duct but the finished 36" Skyray certainly flies very nicely.

Very light (453g 1lb) it uses a relatively small battery (1500mAh 3s) yet has a respectable 10+ minute endurance.

But then of coarse it is made almost entirely of foam!

That is a good question Richard. Has there ever been an analysis of of exactly what makes one model fly better than another.

The basic parameters must be understood; wing loading, aspect ratio, thrust line, tail moment, control surface size, etc.

Are there other properties which should be considered, eg. polar moment of inertia, or is it just the inter-reaction of the basics?

I have been trying to find out what is the actual efflux velocity for a modern airliner engine.

The only figures I have found are for a JT9D. 868 km/hr (539 mph) for the fan and 1120 (696) for the turbine.

Given that the fan generates the majority of the thrust and the 747 with this engine has a maximum speed 594 mph it does rather confirm that planes can travel faster than their efflux velocity.

So given the very low pressures involved in an EDF is a thrust tube that reduces to less than the FSA "to increase the velocity" really a benefit?

Simon, are you citing efflux velocities when the engine is static ?

I think you have overlooked a few things Simon.

The first is that there will be a large increase in gas volume from the turbine.

The second is that the outlet from the shroud unit is lower than the inet to the engine assembly.

I can see that it is not a popular concept for some people, but this really is a mass flow problem, involving the momentum of he gas stream, at the engine unit exhaust.

I cannot envisage an engine where gas stream is not travelling at some speed greater than the aircraft, the speed being the internal gas velocity (in the engine) + aircraft speed = the speed of the gas at the outlet.