Tandem wing for high efficiency? Case Proteus

Homebuilt Aircraft & Kit Plane Forum

Help Support Homebuilt Aircraft & Kit Plane Forum:

Autodidact

Well-Known Member
Joined
Oct 21, 2009
Messages
4,513
Location
Oklahoma
All the books I have read can be construed as meaning that induced drag is a result of the tip vortices. The (detailed) math is very daunting and the illustrations are usually not very good and the explanations somewhat simplistic. It was only indirectly that I came to see the vortex field as being necessary and that was the suggestion that induced drag roughly followed the spanwise distribution, (I'm not sure how accurate that is - it may just be a "design" simplification of which there are so many). I see that I still have a very long way to go and may have to come to grips with the calculus that must be done if I am to truly understand it.

Math can be studied forever, but at some point, we must design an aircraft.

Peter Garrison comes to mind; the first Melmoth had short wings with tip tanks, the second modification had wing extensions and now Melmoth II has long, thin elegant wings that must be much more efficient than those first ones.
 
Last edited:

autoreply

Well-Known Member
Joined
Jul 7, 2009
Messages
10,765
Location
Rotterdam, Netherlands
I see that I still have a very long way to go and may have to come to grips with the calculus that must be done if I am to truly understand it.
Calculus (or math) is a communication tool, a language, just like English, Swahili or Dutch. You can understand and explain most things in multiple ways, and the mathematical one isn't worse or better as the English one, but it's certainly not the only one. The real problem is that many use a language that's not native to their audience or they don't speak fluently themselves.

I don't think there's a limit to how much one can understand when explained in plain text and some simple relations. I've seen several presentations where ridiculously complex issues were explained very understandable, no math involved. But truly bilingual people who master both English and Math aren't as common as many would hope for.
Math can be studied forever, but at some point, we must design an aircraft.
Fortunately enough, for a good design you really don't need more than squares and some products. Nothing to be afraid of, if not the shear volume of those calculations :gig:

Here's a book that I can wholeheartedly recommend:
Amazon.com: The Simple Science of Flight (9780262201056): Henk Tennekes: Books

I've never read a better book on aeronautics. I've never learned so much from a single book and I think that everybody that wants a better understanding of aircraft and flight should read it at least once. Great present for your (grand)children too ;-)
 
Last edited:

ragflyer

Well-Known Member
Joined
Apr 17, 2007
Messages
310
karoliina,

In general, the cg range is proportional to the chord. That's not a rule, but more a common result, if you raise the aspect ratio, your required horizontal tail (volume coefficient) goes down, so you make a smaller stab, resulting in a smaller allowable range for the c of g. Look at the gliders for some extreme results.
Autoreply, I am afraid this is a circular argument. For a given chord length the CG range is of course proportional to tail volume (TV): higher the TV higher the CG range.

However, as the TV (specifically the tail arm length in the TV formula) is denominated in chord length, changing the tail volume coefficient by changing the chord does not really change absolute CG range (it does change relative CG range i.e CG range in % chord terms simply because the chord as changed). IOW if you reduce the chord keeping every thing else constant, your TV will increase but this will not result in higher CG range*.

In the first order the CG range still stays the same. in other words if you double the aspect ratio of both the tail and wing of a given a/c while keeping all areas the same you still have the same CG range in absolute terms; if you had 10 inches before doubling the AR still gives you 10 inches*.

Slender wings have no disadvantage in terms of absolute CG range as such; sailplanes have limited CG range because they maximize aerodynamic efficiency at the cost of CG range by using smaller tails.

* Strictly speaking the doubling example is only true if both the tail and wing have equal AR to start with because the relationship between AR and lift slope is not linear. Neverthless restricted CG range is not intrinsic to slender wings.
 

Hot Wings

Grumpy Cynic
HBA Supporter
Log Member
Joined
Nov 14, 2009
Messages
7,928
Location
Rocky Mountains
I much prefer to stick to the real fundamentals. Induced drag is the energy it takes to accelerate the air downwards (hence the "lift induced drag"), nothing more, nothing less.

I think even you are still thinking one step too far up the ladder from the fundamentals. :nervous:

Put yourself in the place of each individual molecule of air, just happily bouncing off your neighbors in your own little part of space based on your your temperature above absolute zero.

Now add the wing crashing through this world and visualize how all the little upset molecules of air react to the wing, and their neighbors. These individual air molecules each behave according to Newtons laws. The may like their neighbors but they still need their own personal space (kind of like one of my cats, he's got to be near, but "don't touch me"). If you think at this level of aerodynamics you really no longer need to use higher level concepts such as pressure, density, or Reynolds numbers.

This doesn't help much with the design phase, unless you have access to a multi petaflop computer, but it does help with the understanding.
 

autoreply

Well-Known Member
Joined
Jul 7, 2009
Messages
10,765
Location
Rotterdam, Netherlands
Autoreply, I am afraid this is a circular argument.
That's completely true.

If we have a conventional aircraft with low AR wings, we need a big tail. Increasing that tail area a bit for more cg range is a relatively low penalty. But if we now have very slender, high AR wings, we need a lot less tail. Thus the relative penalty for the same cg range is bigger, even though the absolute penalty in tail area is the same. Then it starts to make sense to make the tail lifting (canard) and increase the distance between the tail/canard. If you continue that, you end up with a tandem wing where both wings have approximately the same area and you thus have a much bigger allowable c of g.

@ Hot Wings, you've just nailed both subsonic and hypersonic aerodynamics in what, 10 sentences :gig:
 

Hot Wings

Grumpy Cynic
HBA Supporter
Log Member
Joined
Nov 14, 2009
Messages
7,928
Location
Rocky Mountains
both subsonic and hypersonic aerodynamics in what, 10 sentences :gig:

Sorry, I'll try not to be so wordy in the future. :emb:
 

Hot Wings

Grumpy Cynic
HBA Supporter
Log Member
Joined
Nov 14, 2009
Messages
7,928
Location
Rocky Mountains
both subsonic and hypersonic aerodynamics in what, 10 sentences :gig:

Sorry, I'll try not to be so wordy in the future. :emb:
 

Rick McWilliams

Well-Known Member
Joined
Feb 27, 2011
Messages
274
Location
Santa Paula CA
This is a fast growing post. Roncz, Raymer and others agree that span is the determining feature for induced drag. The root of the confusion comes from aerodynamicists dividing everything by wing area and dynamic pressure, and whatever to get a dimensionless coefficient.

I would like to point out that transfer of momentum to a mass of air generates lift. The energy in that downwash field represents the power required. Momentum is porportional to M V. Downwash energy is portiotional to M V^2. Doubling the span 4x the mass flow, and 1/4x the downwash velocity for the same lift, 1/4x downwash energy.

Long span airplanes have less downwash angle, and thus the tail is more effective. A long tailed or big tailed airplane will have the aerodynamic neutral point further aft. The center of gravity can thus move aft, staying ahead of the neutral point. A long tailed conventional airplane will the tail neutral or lifting.

Winglets are always a point of contention. They increase the mass of air acted on by a wing less per unit lenght than a wing extension. They also do not have as much increase in wing root bending moment as a span extension. The really lively discussion comes from extreme taper and 3D wing tips where that portion of the wing might be said to produce negative induced drag.
 
Last edited:

orion

R.I.P.
Joined
Mar 2, 2003
Messages
5,800
Location
Western Washington
This is a fast growing post. Roncz, Raymer and others agree that span is the determining feature for induced drag. The root of the confusion comes from aerodynamicists......
You probably could have ended you sentence right there and been right on.

Long span airplanes have less downwash angle, and thus the tail is more effective.
Well, yes and no. Depends on what the tail is doing and what types of forces it needs to function within the full flight envelope. In most conventional cases where the CG is sufficiently ahead of the neutral point the tail will have some level of down force on it. For those cases the downwash is useful since the angle of the downwash flow is utilized for generating this balancing tail force.

A long tailed or big tailed airplane will have the aerodynamic neutral point further aft.
Possibly but not always. This is where simple statements and generalizations can get inexperienced designers into trouble.

A long tailed conventional airplane will the tail neutral or lifting.
While there is at rare times a very narrow part of a particular trim envelope where the conventional tail is lifting, it is a very limited occurrence - the condition is more likely to meet instability criteria rather than that of trimmed level flight.

Winglets are always a point of contention. They increase the mass of air acted on by a wing less per unit length than a wing extension. They also do not have as much increase in wing root bending moment as a span extension. The really lively discussion comes from extreme taper and 3D wing tips where that portion of the wing might be said to produce negative induced drag.
Usually an academic discussion at best. Winglets do work but may introduce compromises that may negate the benefit. When you're a 747 flying with a wing load of around 170 psf, the winglets work great - results from several sources including David Lednicer and Aviation Partners, indicate a commercial airplane's total drag reduction (cruise) in excess of 14%. But as the airplanes get smaller the benefits decrease rapidly with wing loading - similar tests have shown reductions on airplanes like KingAirs on the order of only a few percent.

Interestingly enough, high end gliders have also shown measurable benefits of winglet incorporation.
 

autoreply

Well-Known Member
Joined
Jul 7, 2009
Messages
10,765
Location
Rotterdam, Netherlands
Usually an academic discussion at best. Winglets do work but may introduce compromises that may negate the benefit. When you're a 747 flying with a wing load of around 170 psf, the winglets work great - results from several sources including David Lednicer and Aviation Partners, indicate a commercial airplane's total drag reduction (cruise) in excess of 14%. But as the airplanes get smaller the benefits decrease rapidly with wing loading - similar tests have shown reductions on airplanes like KingAirs on the order of only a few percent.

Interestingly enough, high end gliders have also shown measurable benefits of winglet incorporation.
To the best of my knowledge, this has not that much to do with wingloading, but everything with flight regime. Cutting some corners here, but airliners cruise roughly at best L/D, so 50% of their drag is induced, so a reduction in induced drag has a major influence.

Even if you're cruising only 50% faster than best L/d-speed the induced drag is only 16% of your total drag and thus adding winglets yields only 1/6th of the gain as seen in the airliners, further offset by the extra frontal drag of the winglet. So by the time we're in typical GA territory (cruise twice or more the best L/D speed) even a huge reduction in induced drag is barely noticeable, since total induced drag is only a few percent of your total drag.

I've heard from several GA designers that later added winglets and almost all of them add them for better climb, especially in twins (engine failure), often allowing for shorter T/O's or higher gross weights. The ones I know of are sized for neutral cruise performance. Their large reduction in induced cruise drag is only a small portion of the total drag, but this is just enough to offset the extra frontal drag of the winglet. Gliders too usually have their cross-over-speed around the highest cruise speeds, 150 kts or so fully ballasted and only then the straight wings win.

Even for the largest sailplanes, winglets are still beneficial and have a significant drag reduction, even with aspect ratio's north of 50. That, despite the much less efficient winglets, given their very small Re, many have a chord far below 5"

IMHO, their only real drawbacks are design (and production), they're very hard to design correctly, as documented by Maughmer, google his name for several reports.
 
Last edited:

fly2kads

Well-Known Member
HBA Supporter
Joined
Jan 2, 2010
Messages
1,818
Location
Justin, TX
IMHO, their only real drawbacks are design (and production), they're very hard to design correctly, as documented by Maughmer, google his name for several reports.
I have read several of Maughmer's winglet papers, and a thesis by one of his grad students. I find it telling that Maughmer repeats over and over that it is much easier to make things worse than to make them better!
 

Rick McWilliams

Well-Known Member
Joined
Feb 27, 2011
Messages
274
Location
Santa Paula CA
Orion,
The aerodynamic neutral point always moves aft when the tail is larger or further aft. The Nemisis NXT was initially unstable, adding 30 lbs of tungsten to the spinner was not enough. Adding 4 ft2 to the tail brought it back to stable, the center of gravity was even further aft.

There are very few conventional airplanes that are long enough or have a sufficiently large tail to have zero or positive lift. Usually wing pitching moment requires negative tail lift to trim. Since the tail will have a negative lift should we add a lift lump to the center section of the main wing?
 

Autodidact

Well-Known Member
Joined
Oct 21, 2009
Messages
4,513
Location
Oklahoma
Raymer and others agree that span is the determining feature for induced drag.
Now I'm starting to see this. If you hold span constant and increase "A", the result is more induced drag so aspect ratio alone does not decrease Di. The one factor common to any decrease in Di is span. But an increase in "A" alone brings a decrease in area and an increase in Cl and therefore an increase Di. The source of the confusion seems to be what is implied; you have to do some mental gymnastics to get at the essence of the concept because what is said is overly simplistic...."Span is the determining factor"....When none of these things - span, aspect ratio, area, etc. - can be changed without causing a change in the others. If all I did was increase the area (and now I'm contradicting my above statement), induced drag would go down then as well since Cl would go down. It's more what doesn't decrease Di than what does, i.e., it is aspect ratio alone that doesn't decrease induced drag.
 

topspeed100

Banned
Joined
May 4, 2009
Messages
4,063
Location
Oulu/Finland
For a plane to fly well, it must be beautiful.
— Marcel Dassault
I like this comment the most.

I saw a huge eagle the other day with a telescope while birdwatching ( for the first time im my life with a telescope ). It has a huge AR...but also very short fuse ( moment arm )...and it is big...therefore the AR is not exceptional but considerable...there must be a reason for it to have developed as such ?

Also Proteus has a mission profile to meet and thus it developed into what it was..flying very high..low speed etc. ...all peaking to maximise the performance.

http://commons.wikimedia.org/wiki/File:NASA_Proteus_aircraft_in_flight.jpg
 
Last edited:

ragflyer

Well-Known Member
Joined
Apr 17, 2007
Messages
310
Perhaps the following will help clarify a lot of what has been argued here but then again ....

1. Best L/d is not a function of weight: in the first order, in speeds well below supersonic, (all things being equal) changing weight only effects the speed at which best L/D occurs not best L/D per se. There was a rather amusing mythbuster episode where they tested if a concrete glider was feasible and they used L/D as the metric. Of course even a lead glider is just as competitive as a carbon fiber glider of exactly the same shape if L/D is the sole metric.

2. Best L/D is directly proportional to the effective span and inversely proportional to square root of the drag area. Hence span is the biggest determinant of best L/D. I find the formulation:
best l/d =0.89 *effictiveSpan/ sqrt(flatPlateDragArea) rather elegant. It boils down efficiency to the two simple constituents: span and drag area; nothing else matters (AR, weight span loading etc.).

3. All things being equal higher AR will result in higher L/D. This is quite simply due to the fact that if all things are equal and AR increases then its just another way of saying that the span (and hence L/D) has increased.

4. Both 2 and 3 are correct and are mathematically equivalent, however fixating solely on AR in comparisons of L/D can be misleading, as for example in the classic case of B47 versus vulcan*

5. If you are solely interested in minimum sink or minimum power then the biggest factor is span loading: Minimum power for level flight varies by the 1.5 power of span loading but only 0.25 power of aerodynamic drag. This is what Paul MacCready exploited to win the Kremer prize for human powered airplanes.

6. The speed for best L/D and minimum speed are both directly proportional to the square root of span loading. This basically says that while you can compensate for poor aerodynamics with increasing the span you pay for it by a reduction in speed for best L/D.

*autoreply I recall in the past you have questioned the Vulcan - B47 comparison but I am afraid to say it is very much true. You can find references by a number of leading authors: your countryman Troenbeck in subsonic aircraft design, Ramer, and Prof. Mason at Virginia tech; the data very much checks out basic sanity checks as well.
 

Aircar

Banned
Joined
Feb 20, 2010
Messages
3,567
Location
Melbourne Australia
It's not exactly just span that is the determinant --it is the area of the streamtube being deflected that deterimes the mass flow --for a biplane the stream tube is a 'racetrack' shape being a circle equal to span PLUS the gap between the wings (roughly) --for a highly staggered biplane it can be TWO streamtubes --thereby giving ONE HALF of the induced drag for a given span (Slingsby sailplanes built a 'flying billboard" consisting of two sets of wings separated and joined by a LONNNG triangular framework that carried a lighting system --the length would have been fifty chord lengths or more so that each end of the aircraft operated in isolation but for less separated wing systems (eg Proteus) there is still an increase of mass flow and so reduction in induced drag .

the weight of a cantilever wings rapidly chews up any theoretical advantage as aspect ratio increases -- firstly because of the spar car material needed to resist the bending from a concentrated payload and secondly because as the aspect ration goes up the wing thickness and hence spar depth goes down so needing more material and finally the distance that the spar caps (and webs) have to carry the lift loads back to the payload increases -- thus the total spar volume rises very rapidly --a 1.5th plus power actually . I once graphed the empty weight of Glasflugel sailplanes from the 13.6 metre salto to 15 metre libelle 17 then 19 Metre kestrels to the 22 Metre H 604 -- all carrying the same paylod, same material, same manufacturer so 'isolating' the effect of span to a degree -- you can do it yourself and come up with a formula that also accounts for the extra fuselage mass needed and landing gear beef etc --it is more than the 1.5 power ( I might be able to locate that study if interested --the figures are on the web )

Gliders don't count POWER just sink rate which is deceptive since you soon reach a point despite increasing straightline performance where there is NO payload possible --ie ZERO useful work and ZERO efficiency as a load carrier. It takes less power by far to fly the 13.6 metre Salto than the 22 metre super Kestrel. Getting around the structural cost of low span loadings is what is the real metric for progress in balancing aerodynamic against structural efficiency (and why so many homebuilts are built with short 'Herschey bar' wings .

The fallacy of high aspect ratio was shown by the Kansas university re winged C 210 ("redwing" project I think) that was slower than the standard wing because of HIGHER induced drag (operating at higher Cl ) --it IS span loading rather than AR and also Effective span loading allowing for the real stream tube(s) involved . I still find it hard to credit that carrying a high tail down load can benefit induced drag (downlifting costs the same induced drag penalty as uplifting --or side forces as with winglets ) Non planar wing tips could probably be modelled by a second streamtube attached to the main circle of span diameter but not really sure ( A sort of "mickey mouse head" shape but smoother somewhat )
 

topspeed100

Banned
Joined
May 4, 2009
Messages
4,063
Location
Oulu/Finland
How about the tandem wing ?

Mignots Flying Flea and Quickie by Rutan also sports this type of wing.

I assume dragonflys somehow have the "same philosophy".....

What if you mix the wing even more radically and introduce third set of wings...or tailfeathers actually...not in the Piaggio 180P Avanti II design sense but sorta like a a lifting body + control surfaces ( flaps; ailerons ) and tail feathers...is that doomed to fail..or is a B1B actually already as far as you can go on that road ?
 

Himat

Well-Known Member
Joined
May 5, 2011
Messages
2,868
Location
Norway
Long and interesting discussion, one important point:

Calculus (or math) is a communication tool, a language, just like English, Swahili or Dutch. You can understand and explain most things in multiple ways, and the mathematical one isn't worse or better as the English one, but it's certainly not the only one. The real problem is that many use a language that's not native to their audience or they don't speak fluently themselves.
And this is important.
Also, when translating from one language to another in some cases there is no good direct translation.
From scientific math/ mathematical physics to plain spoken english (or any other spoken language i presume) it's not just some cases, it is quite common.
The different translators then end with not just translating the information, they make their own explanation, and put different weight's to the different terms.

Like Cdi=Cl^2/(phi*AR*e) e=1 for eleptical wings, less for anything else.
Then to reduse Cdi, is it best to reduse Cl needed or rise AR?
Remember that changing one is impossible without changing the others as they are related.

And if this was not clear, it's maybe because I try to explain something in a language that's not my own:)
 

karoliina.t.salminen

Well-Known Member
Joined
Oct 7, 2008
Messages
407
Location
Finland
Long and interesting discussion, one important point:
Like Cdi=Cl^2/(phi*AR*e) e=1 for eleptical wings, less for anything else.
Then to reduse Cdi, is it best to reduse Cl needed or rise AR?
Remember that changing one is impossible without changing the others as they are related.

And if this was not clear, it's maybe because I try to explain something in a language that's not my own:)
I speak English everyday at work exclusively and also think in English, maybe I am reading you still wrong or something, but your
statement makes me to wonder (this whole discussion is getting rather strange imho, started from a technical question about "why's and why nots" for
a specific configuration):

Are you saying that there is something strange in reducing Cl needed (alone?)? That in other words is called reducing weight.
Reducing weight reduces drag due to lift because less lift needs to be generated. Weight can be reduced without changing AR, despite
changing AR will change the weight too, but if we consider from aerodynamic standpoint, what strange there is
to make the plane lighter? Burt Rutan has said: if you build a part, throw it into air and it comes back, it is too heavy. Well maybe only
ufos that repel gravity are light enough to meet that criteria (just kidding), but the bottom line is that; to make your plane perform better, one thing you can do
is to make your plane lighter. Also John Roncz mentioned in one of his presentations "In order to your plane to have less induced drag, you shall make it lighter".
I don't see anything strange on that and it does not have direct relation to AR (you can make the plane lighter or heavier without changing AR).
If I want to climb with our Diamond higher, (thus allow it to function with less thrust available), I will not pack it full of stuff, but try to keep it light. Talked also to Jim Price at
Oshkosh and he told that he had taken measures to reduce weight when he did his altitude record 35000 ft with his normally aspirated Long-Ez. So the Cl needed alone without change
in other parameters works based on empirical experience. I don't see anything strange on that, but I think it is strange to try to see something strange on that.

P.S. Aircar's reply is interesting.
 
Top