Sailplane performance

[northwest]

Even as a sport pilot I find that flight performance is what really matters, and can make a big difference in price between models.

I'm planning to do an experiment but first I'd like to hear your views.

Are these the main performance objectives:

-withstand great launch forces and launch very high
-climb in very light lift, but still handle well in the wind
-wide speed range (very slow to very fast)
-long hang time
-very flat glide to range out really far and actually come back
-Be able to get out of dangerous lift without damage
-land precisely

For my experiment:
IS IT BETTER TO REDUCE DRAG THAN TO REDUCE WEIGHT?
I did my best to build as light as possible. The models came out exceptionally light. I discovered that a sailplane can actually be too light! It can't penetrate the wind, has great hang but poor range, and performs very well only in calm weather. Now I have just tried ballasting in my very light AVA and it helps a lot!
But what if our models had half the drag? I have shifted much more effort towards drag reduction and had much success. Now I feel that DRAG is the greatest enemy...and weight is secondary.
Less drag makes a model launch higher and fly faster (giving more range and covering more "air distance" in search of lift) even if hang time is the same.

I decided to take an inexpensive "beginner-type" model and make it into a VERY high performance plane by drastically reducing drag (not weight). I chose the Dynaflite Daydream (all wood, under $40) just to see what could be done on a very low budget. Many of us can't afford to spend hundreds of $$$. The results will be in a new thread called Modified Daydream.

[bobby legue]

Im looking forward to this thread.

[Jurgen]

I think there is a trade off in 'minimising drag' vesus 'construction strenght'. Used materials might force to settle for a search in optimum balance between them. Good luck with the experiments, nice initiative ! Jurgen.

[aeajr]

-withstand great launch forces and launch very high
-climb in very light lift, but still handle well in the wind
-wide speed range (very slow to very fast)
-long hang time
-very flat glide to range out really far and actually come back
-Be able to get out of dangerous lift without damage
-land precisely

For my experiment:
IS IT BETTER TO REDUCE DRAG THAN TO REDUCE WEIGHT?
I did my best to build as light as possible. The models came out exceptionally light. I discovered that a sailplane can actually be too light! It can't penetrate the wind, has great hang but poor range, and performs very well only in calm weather. Now I have just tried ballasting in my very light AVA and it helps a lot!
But what if our models had half the drag? I have shifted much more effort towards drag reduction and had much success. Now I feel that DRAG is the greatest enemy...and weight is secondary.
Less drag makes a model launch higher and fly faster (giving more range and covering more "air distance" in search of lift) even if hang time is the same.

I decided to take an inexpensive "beginner-type" model and make it into a VERY high performance plane by drastically reducing drag (not weight). I chose the Dynaflite Daydream (all wood, under $40) just to see what could be done on a very low budget. Many of us can't afford to spend hundreds of $$$. The results will be in a new thread called Modified Daydream.

Your criteria are good and your objectives are good, but it all seems very subjective. What are you using for your measurements?

Altimiter? Stop watch?

Or is all a matter of "how it feels to me?"

How are you reducing and judging drag? How are you measuring the reduction?

I am not critical of your project, just trying to understand what you are doing.


As far as too light, one of the "truisms" that I now accept is that you can make a light plane heavier, but you can't make a heavy plane lighter. In the case of the AVA, many pilots have ballast in "most" of the time. But that does not mean that they would want the plane built heavier, just to add weight. Because, when the conditions are right, they can shed that weight.

One of my critiera for any new sailplane is the ability to accept ballast. Too light is not a measureable number. It would be too light for some given condition. And the wing airfoil would be a major factor in that characteristic. What is too light for an MH32 might not be too light for an AG40 or an RG15, for example. What is your criteria?

I have an AVA. In wind under 5 mph, at 44 oz, about 5.7 oz wing loading, it is not too light at all. In fact I might even wish I could shed another 4 ounces, but it is wonderful just as it is.

However, as the wind picks up, I find I am feeding in more and more down in order to penetrate. So, one might say it is too light for the higher wind conditions. And where is the point where I am adding too much down? When would I be better off with more weight and a higher wing loading?

By adding weight, have I acheived a better L/D or better sink rate, or does it just feel better to me? Am I a poor pilot who does not know how to handle a plane in the wind? In most cases a heavier plane handles better in windy conditions but does it really fly better? Does it thermal better? Does it hang better?

How do you measure these?

If the wind comes up and I start to feel uncomfortable, I add ballast. Now it is not too light anymore, for me. So the flexibility to take ballast is an important feature to me. Either via a ballast tube or via a ballast box. I would want to be able to add at least 25% ballast to any sailplane.

As for drag, you see many of the super planes with the control rods utside the boom. It would seem that this would induce more drag. Yet the super pilots that fly these super planes don't seem concerned. They feel it is better to have a straighter control rod than to reduce drag. A trade off has been made.


Launch height is very much subject to pilot skill. Other than the strength of the wing, to withstand the forces applied, pilot skill is probably more important than airfoil in achieving height on launch. Two different pilots will achieve different heights with the same plane.

Landing precision is another point that is very pilot specific. This was really brought to my attention when I purchased the AVA. I have been flying full house planes for several years and I have to say that I have never been real good at precision landing. Yet, I put the AVA on the tape almost every time.

Why is that? Clearly it is the pilot, not the plane, for I think most would say that a full house glider affords more control and therefore can be landed more accurately. So, how are you measuring landing accuracy?

The challenge is not whether your criteria are valid, they are very valid. The challenge is how to measure these things and how to judge the the plane vs the pilot.

[Double Pappa]

This is an awesome thread and a great idea. As a builder these are the things I have to juggle constantly. I will be following this closely.

As for criteria.... If we compare our sailplanes to other performance vehicles and toys I guess the only criteria is feel. No two drivers will have their race cars set up alike. I guess a good way to measure or guage some of these things wold be to look at whether or not the plane can be adapted to feel right to the pilot.

It is my opinion that many of the factors we look for in a sailplane are intangable. They are measured in how the plane makes us feel or how it feels to us. I can give an example. The bow I was shooting in tourmaments was not the top of Martin's line and was not better than other bows.....it just felt right to me. It felt like it was an extension of me when I shot it.

My little DLG I have been playing with just "feels right" to me. I have no idea how to measure what it can do even though I know what I can do with it.

This is going to be a great thread in so many ways.

[danstrider]

By adding weight, have I acheived a better L/D or better sink rate, or does it just feel better to me? Am I a poor pilot who does not know how to handle a plane in the wind? In most cases a heavier plane handles better in windy conditions but does it really fly better? Does it thermal better? Does it hang better?

How do you measure these?


Fortunately, there is an easy answer here...
The speed polar is a plot that shows horizontal velocity vs vertical velocity.


http://static.rcgroups.com/forums/attachments/1/0/9/9/0/a1868692-83-minimum%20sink.jpg
The first important item is the minimum sink speed, shown above. It is a horizontal line just tangent to the polar. Finding the matching airspeed gives the speed for the lowest sink rate. This means that if you fly the plane a bit faster OR a bit slower, it will sink at a faster rate. I believe this is why the local guru will say to fly just a little bit faster than stall to float around all day.


http://static.rcgroups.com/forums/attachments/1/0/9/9/0/a1868693-21-max%20L%20over%20D.jpg
The next important item is a line drawn from the origin tangent to the speed polar. This is the maximum horizontal to vertical ratio that is possible. The slope of this line is your L/D ratio (i.e. if the slope is 24horizontal to 1vertical, the L/D is 24:1). Note that the speed for maximum L/D is faster than for minimum sink. I think this is why the local guru tells newbies to "speed up a bit" when you're coming in short for a landing: you can get more glide distance. I think this is also why the same guru will tell you to feed in a few clicks of down "to get the plane on step," referring to getting the plane up to the max L/D speed.

The next item is weight:
http://static.rcgroups.com/forums/attachments/1/0/9/9/0/a1868714-17-ballasting.jpg
As you increase weight, the speed polar shifts radially away from the origin, proportionally to the original distance from the origin. Since the lower speeds are close to the origin, the minimum sink speed does go up, but not by a whole lot. Since the fast end is quite far away from the origin, the faster speeds are affected a lot more. This accounts for why you can add substantial ballast and still fly the airplane slowly but have much better high-speed penetration.

I don't have a handy plot showing the effects of flaps (or reflex), but you can change the shape of the polar, mostly constricting horizontally and gaining some additional sink. Reflex stretches horizontally. I think this is the point that is lost on histarter ... his "old tech" Oly may indeed have a lower minimum sink rate, but the ability of a Supra to change camber really opens up the performance range to make tailored high-and-low speed efficient regimes.

Also, just a general comment to the original question, full-scale sailplanes went through a similiar learning curve. In the early days of soaring, the gliders were being tailored for minimum sink using lightweight construction and highly undercambered airfoils at the expense of a narrow speed polar. Fast forward to today, the highest performance sailplanes are considerably heavier and indeed have a higher minimum sink speed, but they pick up huge increases in L/D by being so drag-consious. This means their speed polar is much more open horizontally for a wider speed range.

Indeed the weight is important, but being super clean is better than being super light.

I suggest reading Helmut Reichmann's Cross-Country Soaring for the speed polar stuff. There are several other sources for designing...

Just my 2c.
Dan

[aeajr]

Awesome post Dan!

[tonyestep]

The quickest way to see how a high-performance plane oughta be configured is to fly one or watch one fly. Check out a Supra or Pike Perfect (or similar plane). They float better than the old-fashioned gasbags, can work much smaller and lower thermals, yet they launch higher and come back from downwind in a way that can only be described as infinitely better.

The ingredients are strong construction (composite materials), modern airfoils discovered by the use of various descendants of Dr. Eppler's code (especially Xfoil), and careful attention to drag reduction in details such as wing-fuselage fairing, fairing of external horns, internal horns (via RDS), proper fuselage shape, and so forth.

Rather than reinvent the wheel, see how others went through the same quest. First, read the history of the Mantis, one of the great modern designs (Tom Kiesling) and its descendants in RCSD:
http://www.rcsoaringdigest.com/pdfs/RCSD-2003/RCSD-2003-12.pdf
(see page 14).

The Supra was the eventual outcome of that evolutionary process. You can check out the Supra pages at:
http://www.charlesriverrc.org/articles/supra/supra.htm
There are numerous pages; spend some time with them all.

Then check out Philip Kolb's comments re the Pike Perfect design at:
http://www.f3j.com/perfect.htm
Scroll down the page to see the polars and read about the design thought process.

When you're ready to rock n roll, download Xfoil at:
http://web.mit.edu/drela/Public/web/xfoil/

[jrerickson]

A few things to think about:

1) The wing is worth about 90% of the drag on the plane. Wing planform and airfoil are very important. A highly cambered airfoil will have great hang time but lousy penetration because of the drag. Conversely, a nearly symmetric foil will penetrate well but won't hang like the cambered foil.

2) Drag squares with velocity. The implications here are that as you go faster there is not only just a little bit more, but A LOT more drag. This shows up on the ping off your launch and when ranging ("running") between thermals. There are three parts of the drag equation. The first is vortex induced drag (lower speeds) then wing profile drag and then finally then parasitic drag (higher speeds). Induced is the wing part, parasitic is the fuse part.

3) It's hard to have things both ways. A good plane for the wind (low drag, heavier) will not be a great landing plane. For landing you want something that has a very low stall speed, so a thicker airfoil or effective flaps make a big difference. Light is good in the landing zone because you don't have to worry as much about managing momentum.

3) The suggestion about an altimeter is excellent. It's much easier to dissect accurate data than to go by how you felt.

4) Be subjective and test at different times of the day. When it's the "noon balloon" a Dynaflite Skeeter is going to feel like a Pike Perfect. A good sales tip is to fly the plane you want to sell into a great thermal than hand the buyer the transmitter :-)

The same test on a cold windy morning is going to show you why you can buy 15 Skeeters for 1 Perfect.

John

[tonyestep]

"...A good sales tip is to fly the plane you want to sell into a great thermal than hand the buyer the transmitter..."
========================
Ah yes, I've seen it work many times.

[Gary Binnie]

Great thread, I'm glad that Dan posted the polar diagrams which should explain things clearly.

If you extend the airspeed line left of the origin for headwind and extend the sink rate line upwards for rate of climb you get more speeds to fly, the basis of MacCready theory (after Paul MacCready).

More here: http://en.wikipedia.org/wiki/Speed_to_fly

Interesting how this all gets put to use for model gliders but the principle should be the same.

On drag reduction the major effort for full size competition crews is spent on cleaning and polishing the wings, especially the leading edges. Bugs on the LE have a marked effect on best L/D, flight computers have an adjustable input for the pilot to enter a percentage performance loss based on the amount of flies he can see (known as the 'bug polar'). This only really applies to laminar flow wing sections. Old gliders will fly fairly happily with wet wings from rain but modern comp gliders have warnings in the flight manuals not to even try to take off if it is raining.

Helmut Reichmann's book is very good but even he said that some of the wing polishing time could be better spent in the air practising, I think he was right there.

Winglets are another recent development for full size gliders, the idea is that they reduce tip vortices (reducing drag) at thermalling speeds when induced drag is highest (especially when ballasted) with possibly a slight penalty at higher speeds.

I guess good telemetry from the model and post flight analysis is very important here, I'm just a flat field sport model flier but have quite a few years instructing in the big ones.

I had the pleasure in my early gliding days to go cross country with a world champion who said the secret to it all was to 'fly in the lift, stay out of the sink'. Wise words but I never quite got the hang of it like he did!

Cheers

Gary

[schrederman]

Drag increases with the square of speed... true to a point. Parasitic drag, that is, However, induced drag decreases with speed. Some of the full-scale stuff actually tests out at a higher L/D ballasted than when running empty, due to this phenomina.

What you're experimneting with has been tried by all the old heads on this board... Changing sheeting schedules of the wing, narrowing fuselages, adding fairings, etc. Not meaning to discourage you by any means. What you're doing is interesting to us all, especially if you have some solid method of measuring.

Jack

Oh... yeah... don't forget the winglets! :p :D :p :D

[northwest]

Ok. Glad to see so many replies and really appreciate the links to all the useful info.
The measurements will be scientific but simple. The results should lead to reasonably valid conclusions.

-I will be launching from two hi-starts (one large nesail pinnacle and one dynaflite H.D.) which will be pulled to specific tensions in lbs.
-I have a zlog altimeter for launch heights.
-model weights/ballast will be measured from a digital scale (to .05 oz).
-an assisiant will run the stop-watch.

Statisically, readings will be the average of several under specific conditions.

This will take time as mods are completed and flights accomplished in a range of conditions. Also expect comments on overall handling, feel and comfort level.

I already have a used Daydream without repairs. I flew it several times before this idea came about. It flew ok, like the average beginner's plane but I already started modifying to make it better. Results will be taken from three different wings: (1) the stock wing, (2) a modified wing thinned considerably(especially the tips). (3) a wing made from spare AVA wing-tip panels. I have flown with options 1 and 3.
(The AVA tips make a totally different aircraft but I'll describe that in my next reply, and there are no statisitcs for this yet).

Test Bed Mods:

-Dowels for elastics removed, front bulk head re-done/rear beam installed to accept bolt on wing.
-tail surfaces removed, new ones built. Razor sharp leading /trailing edges. Stab has re-enforced hinge-line spar to withstand hi-speed launch stresses.
-the 2 deg wing incidence at the wing saddle reduced to zero.
-ballast area prepared inside the belly...consists of industrial strength velcro (female) glued to the walls and floor. The male velcro will be glued to the lead ballast strips.

[AMBeck]

A couple things you mention might make your life interesting. Sharp TEs are usually a good idea (although occasionally a Gurney flap will pull a separation bubble off the upper TE and reduce low speed drag). The sharp stabilizer LE can cause big problems. While it may lower drag around zero tail angle of attack, it will greatly reduce the tail's stalling angle of attack. That could be fatal for an all flying tail. If you're using a fixed stabilizer/hinged elevator, deflecting the elevator will cause the tail to go out of it's low drag range. Tail stalls are at least as exciting as tip stalls. LEs are round for a good reason. Changing your wing incidence will change your trim speed, but won't have much effect on overall drag. Getting rid of the rubber bands and dowels is a great idea and will pay off. Another trick that will pay off is to put a fuselage vent in the tail. With a slight vacuum in the fuselage, leaks around the hatch and wing root will blow into the fuselage and not cause separations in a very critical area.

[lincoln]

While it's worth doing, you can only get so much out of aerodynamic refinements. Some of the best designers around now have put very large chunks of their lives into studying aerodynamics. If you are not prepared to do the same, I suggest your best bet is studying what they do and borrowing it. The more you modify your Vista to be like an Allegro Light, the better it will fly.

Handling:

what you want is low radius of gyration, high damping, plus a combination of twist distribution and planform that gives you a stall you can live with, neither too soft or too sharp. To get low radius of gyration you make the tips and tail as light as you can, and maybe put just a bit of extra weight at the cg. To get high damping, use a fairly long tailboom. You can make the tail surfaces smaller if you do that, and still get a net gain in damping.

drag reduction:
probably hard to do better than the Drela foils, which are significantly better than previous ones, although, IMHO, not revolutionary. Use a real airfoil, if thin, on tail surfaces. A good one is maybe 6% thick, front an ellipse with high point maybe 23% or so, then straight lines tangent to ellipse, back to t.e. Seal control surfaces, seal wing/fuselage joint, maybe make little fairings at junctions between various aircraft parts. Sand fuselage to have rounded edges.

However, you can do all this and still, if Joe Wurtz shows up at your contest with a Gentle Lady, he'll still win. One of the fun things about our hobby is that at a contest, pilot skill dominates. Now and then, some expert or other will show up at a contest with junk. Chances are, if that expert was going to win, he'll probably still win flying the junk.

[N9DP]

I’m glad you elaborated on your lab technique in your second post. After reading the first post I had the impression you were seeking to morph your $40 kit into a $1500 Pike Perfect!

You noted that: “Statistically, readings will be the average of several under specific conditions.” Good lab technique also requires you to compute the standard deviation (sigma) of at least five trials. Then when you compare the means of A vs. B, they should differ by more than one sigma, else the results are indeterminate due to experimental error. In other words, if you plot error bars on your means, if they overlap, you don’t know if your change was significant.

My concern in the design of your experiment is that the wind may be a dominating random factor. Quite frankly, you are starting with a model that is fundamentally draggy by design, and cleaning up all the intersection drag and tweaking the incidence is not going to make a huge a improvement in performance.

I suggest you compute the mean and standard deviation of five or more trials of the same configuration on several different days. This will tell you what your ability is to discern small performance changes because you will have a good handle on your probable error in measurement.

Note on my old boss’s (VP Research) door: “In God we trust – all others bring data.”

Good luck and keep us posted.

Dennis (retired experimental physicist who makes a science project out of every hobby!)

[luvF3b]

Here's an answer after too many wines!! and from a non aerodynamics guy, perhaps Mark Drela can provide corrections/comments.

Can an glider be too light? Short answer is yes. Because of increased profile drag due to low reynolds numbers.

I'll have a go at the looong answer:

The performance of a thermal glider is related to L/D ratio for penetrating, and (root L squared)/D (ie power factor) for minimum sink . In other words, the performance all relates to the relationship of lift and drag at various speeds. The lift required for a given flight speed and wing area relates only to the weight. (See point 1 below). Drag is the real variable.

The drag on a plane is the result of parasite drag, vortex induced drag and parasite drag. At slow speeds vortex induced drag predominates, at medium speeds all three are important, and at high speed parasite drag and profile drag predominate. Parasite drag is simply a matter of reducing fuselage cross sections, and not poking things out into the air flow (perhaps a bit simplified). Profile drag and vortex drag are the important things given a clean design.

So assuming slow speed flight for a given weight, there are a number of factors at play, Addressing firstly profile drag:

1) the weight drives the required forward flight speed for a given wing lift coefficient.

Cl= M X 32 / 0.5 x K x V² x S

where Cl=wing lift coefficient, M=model mass (lb), K = constant related to the fluid properties (air), V = velocity (ft/s) and S= wing area (ft²)

So, for a lower weight, lower flight speed results at constant lift coefficient, or conversely, for a given flight speed heavier weights need a higher coefficient of lift from the wing. Given the "C" shape of most foil polars, reducing weight (ie reduced Cl) can reduce profile drag by moving closer to the apex of the "C" at a given (slow) flight speed, ie moving into the "low drag bucket."

2) Because the weight drives the forward speed at a given Cl, heavier gliders will operate at a higher Reynolds number.

Re=6400 x c x v

where c=wing chord (inches) and V= forward flight speed (ft/sec)

Looking again at profile drag, the drag coefficient of the foil reduces for a given Cl when the Reynolds number increases; so, increased weight = increased speed = increased Reynolds number = lower drag coefficient for the profile. Temper this with the point made in (1) because the increased weight may result in moving outside the "low drag bucket"

And now to vortex induced drag:

3) A higher lift coefficient means more vortex induced drag. Vortex induced drag is related to the square of Cl, and aspect ratio. So for a given flight speed, a heavier model will need to operate at a higher Lift coefficient (see 1), and will therefore produce more vortex induced drag.

At some lower weight, there will be an optimum between lower drag due to reduced Cl moving closer to the apex of the "C" (ie into the low drag bucket), increased drag due to lower reynolds number, and the lower drag due to lower vortex induced drag (ie lower weight = lower Cl for a given speed).

As you can see, this really is a multi dimensional problem that involves variables such as the weight, wing area, aspect ratio and profile.

A VLM program such as XFLR5 (freeware) is a great program that can provide the answers you need for the Ava. :confused:

Cheers,

John