Tall Paul

C-130, L-1011, have inverted sections for the horizontals.

UStik

Yes to the last two paragraphs, but objection to the Telemaster stuff!

Thanks for putting my points about big static margin and inverted stab airfoils on full-size airplanes into better words. But what about my explanation of the upright stab airfoil? (I can sleep well, I offered that explanation.) May well be that the old designers thought of a lifting stab, we just don't know. The designer of the original STM definitely had a free-flight background (multi German champ, second at WC, but he had been a Heinkel factory pilot flying He-162 as well), but I think he might have not used a rear enough balance point to get the stab lifting because the decalage is quite big (4 degrees aerodynamic). (And a slab stab would be just parallel to the flat wing bottom giving the same decalage.)

Nevertheless, he designed the STM in 1967/68 as a load carrier, as what it was explicitely advertised. Of course, the other intended use was a trainer, and they even dared to say that it "facilitates proceeding directly from free-flight sport models to multi R/C flying". This remarkable qualification was due to the well-known, docile flight behavior and light net weight. But it was built sturdy enough to carry that same weight as load (11 lbs AUW, still only 19 oz/sqft wing loading, but that was just the legal weight limit). Docile flight behavior was a prerequisite also for the load carrier, but it was not the typical self-righting trainer and a full-house ship, so trainer was secondary use.

The Bridi version you may know gave up the sturdiness and recommended a more forward balance where the stab does not lift and the static margin (pitch stability) is bigger. That's a pure trainer which has to be beefed up to be a heavy-load carrier. Static margin is still only 15%.

BMatthews

I may be remembering wrongly in that case. I thought the regular 40 size Telemaster came out before the STM. If I am wrong on my dating I'll just claim Senior's Memory Syndrome and we'll forget that these last couple of posts even happened....


We may never know the real reason behind the big tail volume coefficients with their strongly set back balance points and lifting stabilizers but I strongly suspect that it was originally used to cheat the minimum wing loading rule used in competition free flight in the US. Later, as we can see in the present rules for FAI models, the specs were changed to total horizontal surface area to counter that trick towards creating almost tandem wing models. I'm not exagerating either. We can find lots of examples of models from the mid 50's to early 60's where the tail moment is so long and the stabilizer area so large that we see balance points specified at anywhere from 70 to 120% of the wing chord. That's right 120%. The balance point being indicated behind the leading edge.

Anyhow I still doubt highly that the designer of the STM intended it as a HEAVY load lifter. Certainly for camers, piggy backing sailplanes and other such moderate load use it excells. But if we loaded it down with a 15 lb payload I suspect we'd be back to a seeing some of the problem issues from the SAE models show up.

You're right about the tail should not stall before the wing. That is why the upside down airfoils on the big lifters. The idea being to allow the tail to reach a higher (negative) lift coefficient before seeing a lift drop off related to the stall of the tailplane and the resultant loss of torque for lifting the nose to flare the model or even just to fly it level. Such occurances during a landing flare where the stabilizer gives up tends to bend the nose of the model into rather artistic free form shapes

In the case of the STM we have a wing that is lifting upwards and wanting to pitch itself into outside loops due to the pitching moment. Two things are countering that. The location of the balance point well to the rear wanting to torque the nose up and the tail which carries a portion of that rear balance torque. So why doesn't the STM suffer from lack of stabilizer authourity when landing? A bunch of numbers would need to be run but I suspect it is because of the low pitching moment of the STM airfoil compared to the strongly aft loaded airfoils used on the SAE and other serious load lifters. That and the fact that they run with too small a tail volume to allow the stabilizer to reach the required lift coefficients to deal with the pitching moments of a high wing loading and high lift coefficient during take off and landing. Maybe more during landing.

It has always amazed me that our specs for calculating the tail volume coeficient do not include a spot to insert the pitching moment. I know from experience that I can make a model with a symetrical airfoil fly very nicely in a stable manner and have no surprises with a stupidly small stabilizer and put the balance point at or just barely in front of the neutral point and fly it without issues. Now if I take the same model model planform and balance point but switch to a Selig 1223 what do you think is going to happen? I'm going to suggest that it'll mostly be a question of how many pieces there would be. If we were to run some numbers and study a chart of how the pitching moment of a heavy lift airfoil alters with airspeed and lift coefficient I suspect we'd see a pretty big change in tail load over a typical flight. It may well be that the strongly nose down pitching force from a heavily working airfoil of this sort needs to be looked at and the tail size and moment selected to move the balance point at a still safe stability value to a location where it is back far enough that it aids the tail instead of fighting it. At the same time perhaps there is something to be said for using a compound hinge for the stabilizer that allows us to more smoothly alter the camber of that surface by using two or even three segments along with the fixed front portion to delay the stall of that surface. After all, we just gave the airfoil a "flap like" amount of camber to allow it to operate at extremely high Cl's at relatively low airspeeds but then we leave the tail that has to control that wing with a flat airfoil that can only alter camber at one hinge point and suffers from separation bubbles and stalling at very small angles of deflection? Sorry but that part makes very little sense.

My apoligies if some of this seems cloudy. I'm throwing it out as it comes to mind so some of it may seem a bit jumpy. But these are ideas that on their own I've thought about from time to time. In particular the one about our stability margin CG locators not allowing for airfoils with different pitching moments.

da Rock

I second that motion... uh, wondered about that too..... It most certainly works against the tail. Always thought there ought to be a "traction circle" for the tail.

Doc.316

Yes that is the one Tarus Flyer...the channel wing would take off stol...ok in climb....slow...everything you would want in a SAE plane...except for making the structures guy go crazy...lol.

I think your analysis for free flight is spot on. The analysis seems good also. When loaded the planes I have flown did exhibit a strong downward pitching moment (due to lift or weight being in front of the cg or a combination) even with the high neg lift aerofoils (upside down)....the worst of these was the flying wing (think horton) that had to be flown with almost all up elev travel. This was probably due to the fact that the wing had a small tail volume coeficient (no tail really) to counteract the weight.

So thinking about this the way to go about getting something that would fly well....is to have an adjustable elevator incidence...Start flying it with no payload...then add payload and retrim with the elevator incidence "screw" until trim was obtained with that weight (or a close to trim condition at least) then keep flying and plotting untill a chart was derived for the airplane of "trim screw position" versus weight was obtained.

The other option would be to move the cg around with weight to obtain the zero weight trim...of course this would change your stability...which if you had to put the CG behind the neutral point to trim might not be very good.

So this sound correct?

Steve

UStik

Oh, what have I done to deserve this? I'm suffering from the Senior Telemaster Memory Syndrome, as well as several others, especially cloudy explanations. My apologies!

You may have a look at the OEM's Telemaster 1973 catalog page in the attached pdf. 11 lbs was the legal weight limit for all models, and with a barely 1 hp .61 engine of the time (1968) and a low-pitch prop you just had a decent thrust/weight ratio. The design task was similar to the SAE task, getting maximum payload from limited AUW. The designer solved the problem by choosing an optimum wing loading, then designing the structure for minimum weight. "Aerodynamically", it would carry much more load if only strengthened and powered better.

I think the flat-bottom airfoil (not Clark Y, please) and the cambered tail were given. You gave me two good new ideas. Doc, you mentioned "design rules", and I think both were such rules, or a habit back then. The other idea is that not any cambered airfoil is used but really flat bottom (the rest may vary) just for ease of building and trimming. I remember that sometimes C/G position was given and decalage had to be trimmed by shimming the stab.

The original STM even had a tail hold by rubber bands, but for the contest I thought of the modern Giant TM first. It's basically a scaled-up version of the generic layout. The general weight limit is given also in the contest: 55 lbs. Net weight of the Giant should be 23 lbs so it could carry 32 lbs. It would need strengthening in some points but it flies gracefully. You'll argue that's not really heavy load, but again that's my point. You have to find a total optimum, and I guess that leads to moderately high-lift airfoils and rather big wing areas. Probably the old designers used only those conventional design rules and knew by experience that a design would work. They just tried out how well it worked, and the Telemaster turned out to be a masterstroke. (The designer had tried out many design parameters and built really hundreds of free-flight models just for that.)

Telemaster is even a rather conventional design. I don't know what extreme designs were used at contests, only the free-flights with club-and-stick shaped fuse and a small stab on a long tail, the stab flat-bottom airfoiled. The STM's stab has 24% of the wings area and the moment arm is 2.75 times the wing chord length. Tail volume is 0.665 according to the Plane Geometry tool mentioned above. Though simple it's rather sophisticated and accounts for the fuse in the calculation of the neutral point. The NP is independent from wing pitching moment, as usual. I don't know either why. Pitching moment is accounted for in the calculation of balance or of incidence angles, respectively, and shows what you describe: Exchanging the flat bottom airfoil for a symmetric one makes the STM tail lift instead of pushing down. I guess the free-flighters were trimmed to float at minimum sink rate, depending on the contest rules, and that would require a rear C/G making the stab lift. The small pitch stability wasn't bad since the model wasn't sensitive to gusts. They even built forward fins to have nearly no yaw (directional) stability.

Now if you take the Telemaster design and morph it to really heavy load you're asking for the troubles described. Look at the heavy jet transports or even fighters with their small wings. They are fast even with their high-lift devices, they need much thrust even at high speed, and they need several tricks to get the most out of the design. The transports all have moveable stabs in addition to the elevators, even though that's partly due to their big AOA range. But many WWII planes had (in-flight) moveable stabs as well, I think to adjust decalage to flight condition. Smaller planes with a forcedly forward C/G (C177) have just an all-moving stab, sometimes with inverted camber and even anti-servo tabs for even more camber. (Yes Doc, I agree.) But these moderately heavy planes have/need quite much power and quite effective flaps. (There seem to be more tricks, like forward-camber airfoils where the center of lift moves forward at higher AOAs.)

Meseems these are two extremes: The heavy small plane with high-lift airfoil, aerodynamic tricks, complex build, much power; the lightweight big plane with moderately cambered airfoil, plain simple, easy build, medium power. Why not search for the optimum in between? I mean, planning/engineering only replaces chaos (random design following "rules" ) by mistake (missing the optimum), but isn't that the fun of it? And the modern tools make it easy to run many numbers and calculations. We don't need to build hundreds of models today.

Finally, I looked into a 1940 mechanical engineering encyclopedia, airplane chapter. There is a "how to" or workflow for designing an airplane. After specifying the basic givens it says "find optimum wing loading". Later it tells how to find stab size and decalage. They use a simple force/moment plan (sort of side view) and derive a total-wing-moment curve over AOA. Then the stab is chosen to give a matching curve and stability over the whole AOA range. I don't know who came up with the neutral point concept and when, though I find it convenient and use it. But obviously there are other simple ways. And to the rule that the wing has to stall first, I even forgot that it is that general and applies to the inverted stab as well. I thought of the more popular version that even in extreme cases, for instance when airspeed is nearly lost in a gust or wind shear and the airplane starts falling due to wing stall, the stab must not stall. With a flat-bottomed wing (camber may be 5%) that requires a cambered stab as well to bring the lift-over-AOA curves to accordance. Maybe the Telemaster designer had that in mind, if he didn't simply follow a "rule".

Tall Paul

I've always considered the lifting horizontal on a free-flight the speed controller.
These are single speed airplanes, and the horizontal is the thing that keeps them that way.

UStik

To illustrate the airfoil moment conundrum:

The S1223 is impressive, lift coefficient up to 2.2, even though at big drag (0.04 to 0.05), as supposed. Measured values, but unfortunately no moments, as usual.

The only airfoil I found that is remotely comparable is HQ 2.5/9 with flap. It's a glider airfoil, the 22.5% chord flap for camber adjustment. There are measurements with 7.5 degrees deflection illustrating the problem, I mean knowing true values.

JavaFoil (I think Xfoil the same) over-estimates lift (for S1223 even more) and even mis-predicts the moment coefficient characteristics. The down-pitching moment (negative values, reference point is 1/4 chord) increases with AOA (left picture, diagram on the right, the kinks in the curve are stall) and the values are too big.

The measured values (right picture) are a bit smaller and decrease with increasing AOA (the curves drawn the other way around, more down-pitch being higher). Different curves are for different Re numbers, by the way.

So the measured values are just the opposite of what the calculation predicts. That's not the same for all airfoils, but even worse, you never know in advance what difference will occur. These calculated curves are just unreliable.

But what could the measured behavior mean? With a constant moment coefficient, absolute moment would be proportional to dynamic pressure (speed squared). The faster you fly the harder the tail has to push down. At first I suspected the calculated behavior would exacerbate that, but now it seems to be even better knowing the measured behavior. When the wing is pitched up to higher AOA the moment coefficient decreases, so it helps instead of resisting the pitch change.

Now who knows the true values of a S1223 (or similar) and who has thought out the airplane's behavior or has a tool to evaluate it?

By the way, a S1223 wing is really hard to build.

BMatthews

But any plane at any speed that has some measure of postive stability is a single speed plane. Once the trims are set it will always try to return to that trimmed airspeed. We just cheat by adding things like downthrust so the power can go towards making the plane faster. But kill the engine and allow the plane to glide and it will assume the trimmed airspeed where the nose up from wing and tail balance the nose down forces of the CG being in front of the neutral point and the forward pitching moment of the wing.

And in fact I can show you personal examples of where a lifting tail free flight model was NOT a single speed model. I've got a personal example here where the wing had a wider than normal chord. The model flew well with postive pitch stability unless the speed got a bit too high. At that point the combination of wing pitching moment and stabilizer lift became too much for the leverage of the lift of the wing and tail and it would bunt over to a vertical dive. I solved the issue by making a new stabiliizer with a less cambered lifting section.

Most folks not involved with free flight would expect such models to be trimmed for a higher stability margin. But in reality the contest models are balanced on the ragged edge with only a very little stability margin. This being so the model can be trimmed for the relatively strong power portion of the flight where we have to live with the fixed trim and use just downthrust and a spiral climb so the "loops" performed under power are canted to the side of a spiral path up a column like shape. Because of this the balance of the other things that affect the model are critical.

I saw another example of this. A fellow test glided his model and set the trim for a nice boyant test glide. Early power flights had the model set up to transition nicely and glide in a most wonderous way for one circle and then DT down. I was there for the first go at a contest flight. THe model SCREAMED up and when the engine quite and the model transitioned it was so high it looked as though a max was a given even without thermal help. We watched it for the first circle... so far so good. THe next circle was a little faster. THe third faster still. Now ALL eyes are watching because the morbid side of us all knew what was coming It never made a fourth circle. It had sped up and transitioned to a most amazing vertical lawn dart dive. There wasn't enough left to repair... which for us free flighters is pretty bad indeed. In this case he basically had coarse ground balsa dust in bags... The point here was that the early test glides and test flights fooled him into thinking all was well. But some bit of turbulence that created a slight speed up was enough to upset the delicate bubble of stability it had and produce the death dive just like my case. Yep, we run 'em REAL close to the edge.

Doc.316

Bmatthews...I just love the smileys in your last post.......I can tell you either must have been in competiion with the guy or everyone thought he would win hands down and didn't think they would have a chance....I can just see everyone getting their "circular flow" books out when it was starting to go down.

Well, what everyone probably doesn't realize (and this is a fact with the teams that I have worked with in both the AIAA and SAE competitions...(and it is kind of interesting). The teams with one or two modelers on the team that fly RC do not do as good as the teams with NO modelers on them. Ok...tell me why that is? I do know that I give "hints" to the teams....I am thinking that the teams that have non modelers pick these up...and the modelers tend to do their own things. I did tell one AIAA team "lightness is the key to winning" ....(and they built a really really light plane)...in fact it would have been up there if the gearbox had not come apart...(sigh)...I think they were still in the top 15...with only 2 flight scores.

I think really what we should be telling these teams is: Multiple designs could win if the team gets together and works well with each other (I have had some teams literally explode....kind of like the free flight plane...lol)...and each part of the team (aero, structures, etc) optimises (bad spelling) each segment of the airplane. Like this team...I would give them this "hint"......1.Find a "serious pylon flyer" where you are from....2. Let him run in your engine.... 3. Let him "tune" your engines to the fuel used...(Let me tell you something about engines...when you tune an engine for max power (and that is what you need to lift a load)...you need to look at the timing (think glowplugs) since the heat of the glowplug determines timing....along with compression ratio...er thickness of the head gasket...er.....4. Buy 4 "stock" engines...one will be better than the others. (hey, it costs some money to win...I used to buy 6 at a time for pylon racing). I suspect the advantages of doing this will outweigh all the drag reduction techinques you can come up with!......wow 4 Jetts are going to cost the university some money....(sell the ones that arent "the one"...)... This is my final "hint"...can't wait to see the Utube video of the Jett with the reduction drive....

Steve

UStik

Amazing stories, I just can't understand the explanations.

In fact you lowered the stab's aerodynamic incidence angle, and just setting more decalage would have had the same effect, wouldn't it? And I suspect the tail was not lifting, at least not at higher speed, it was just not pushing hard enough. The AOA range where the stab really lifts is very small, only around speed of minimum sink rate.

To me, that sounds suspiciously like spiral dive, after all the model sped up only in the turn, and it rolled (or it wouldn't have been a spiral). I understand that the model had small pitch stability, but that also means it had to speed up quite a bit in turns to pitch up enough to have the higher lift needed in a turn. That might have led to a bigger lift coefficient than that of spiral stability. Small pitch stability allows only wide turns, and a gust resulting in some roll makes the model curl. Or did I miss something?

BMatthews

Ustik, let me try to explain it more.

The model in question is my own design P30 rubber model that used a wider than usual chord to get more wing area. But it also had a roughly 30% stabilizer area with a decently long tail moment. I made it this large to allow a strongly rearward balance point location so I'd get some extra lift from the tail. Basically making it a semi tandem wing. I set the CG location by trial and error using test glides and observing the speed and radius of recovery from minor stalls. This produces a trim which is only minimally pitch stable but just right for free flight contest flying.

Now, there are two forces acting to pitch up the nose on such a model when gliding. One is the nose up torque from the lift of the wing. If the wing lifts at the 25% chord location and the CG is back around 50% as it is with this model then the lift of the wing is trying to rotate the nose up obviously. The second force is provided by the center of drag being higher than the true 3 dimensional location of the CG. This is a speed dependent factor but one that is often overlooked in types of flying other than free flight models.

On the other side of the coin we've got two forces trying to pitch the nose down. First is our pitching moment of the wing. In this case due to the wider chord than typical for free flight it is a sizeable amount. Second is the stabilizer's lift contribution. Yes, lift. With the balance point set to around 50% obviously the stab has to provide lift at any point in a very wide range of flight situations. The stab gets it's lift from the angle of incidence relative to the wing and the lift coefficient curve due to it's airfoil. In the case of the first stabilizer it had this was fairly high. I'd chosen to use an airfoil with a high camber value but still one that was only about 2/3 the camber value of the wing's airfoil.

Now that the stage is set let's look at one of the lawn dart flights. All would be well under power. The model climbed nicely off the launch and maintained a nice bouyant flight until the prop goes into freewheel. On a calm day with light lift the performance was superb with good lift riding ability and easy max time flight scores. But when things get more strong and turbulent the model would be tossed around more by the shear effects found at the edges of a thermal. If one of those hit it just right the model would either be pitched into a dive or the air would hit it in a way that it thought it was in a dive. (keep in mind that with an 80 gram "bit of nothing" flying at a fast walking pace with lots of surface area local turbulence effects are highly significant to how the model flies where on bigger and heavier craft it's just some minor bumping around) When this would happen I'd see the pitching moment and stabilizer lift overcome the rearward CG and drag effects and over she'd go into a bunt with no recovery. The only bright side is that because it was an 80 gram bit of nothing with a lot of drag from the wing with it's undercambered airfoil and freewheeling prop up front the "CRASH" was more of a mild "doink" where it would tap the ground and then just fall over on it's back or belly.

Now there were two things I could have done to fix this. The simple one would have been to shift the CG forward a bit. I did that at first and retrimmed the downthrust to counter it. But all that happened is that the bunt occured at a slightly higher speed. A further shift forward would have fixed the issue but then the powered portion of the flight would have suffered. Instead I altered the lift dynamics that occur between the wing and tail by making a new stabilizer with an airfoil having less camber. Because the tail lift is speed sensitive this gave me a less steep build up of the lift in the tail. But the wing's lift build up from speed was still the same. So there would be less of a pitch down effect from the combination with a speed rise. That change allowed me to move the CG back to the original spot (the tail sections were weighed and were the same to within a small fraction of a gram and the kidney bean size lump of modeling clay used to move the CG forward was easily removed). But now the torque effects of the wing's pitching moment and the stabilizer lift added together produced a curve that did not over power the nose up torque effects of the wing's lift and the drag center. With the reduced tail section camber it never again had any issues with recovering from a stall. Even deep stalls were recoverable without issue.

What I've described here is a very delicate balance. It also shows that there is more at work when it comes to stability than just the factors that are included in the usual online CG location calculators for models. For example indoor duration models often fly with 0% or even slight negative stability margins. Yet they are stable. They managed this because the wings are mounted to the fuselage with posts so the true 3 dimensional CG is well below both the lift of the wing and the drag of the wing. The vertical planform layout provides the stability forces that are missing in the balance point location to achieve a similarly minimal but still adequite pitch stability.

As for my buddy's model that went in it was definetly not a spiral stability issue. Seen too many of those not to be able to recognise the tightening turn of death. In this case the circle didn't tighten. In fact it either stayed the same or opened up as the speed built. Free flight contest models have way too much dihedral to end up with spiral stability issues in any case. In fact to aid the model in being able to hunt around in lift we tend to make the vertical tails undersized if anything. So if we run into any spiral stability issues it's on the side of dutch rolling due to too small a vertical tail. Nope, this was a speed induced bunt that looked like it was very much of the same sort as my later P30 design produced. In fact it was my buddy's earlier experience and hearing him talking about it in the aftermath that gave me the idea of making a new stabilizer with the flatter camber value airfoil.

Anyway, to try to bring this back to the load lifters and given that we often do not look at all the factors let's look at what this might mean to the lack of elevator authourity in a load lifter during landing. We've got a honking big wing with an airfoil that looks like it was designed in a "flaps deployed" manner due to the camber and strong aft loading of the curve in the case of the more recent airfoils from Eppler and Selig. THis produces a hellishly strong pitching moment. But it also allows the models to fly at a slow airspeed. But now we're asking the tail to do it's job of controlling a wing that is able to fly at a slower speed so the tail has to achieve it's lift requirements with a low airspeed. Doesn't seem quite fair to the tail now does it? So how to make it so the tail doesn't need to work as hard? One way is to make the tail larger than it needs to be for simple stability so it reduces the area loading to achieve the lift required. The other would be to make the airfoil more effective like we did with the wing. Hence the upside down airfoils. But perhaps there are other options. For example multi hinges that provide a smoother surface for the variable camber. Perhaps a full flying stabilizer that also has multi hinge points so that it is split into 4 chordwise segments so at the same time it is altering it's overall angle of attack it is also altering it's camber in a much smoother manner than you can get with a single hinge line? What about a one piece flexible skin on the lower side with overlapping sliding plates on the upper side so you get a smoothly variable camber surface on the more critical lower side of the stabilizer?

I guess the point is that if the load lifters are subject to losing pitch authourity at landing speeds that some additional thinking needs to be placed in that area. Optimization of payload location in both the fore and aft and vertical directions is one area and further thinking about how to optimize the stabilizer lift at the wing's landing airspeeds is another.

UStik

Thank you so much for these comprehensive explanations, especially clearing up my spiral dive suspicion. We're not on the same wavelength, though, since your explanation of the "lifting stab" issue is even too elaborate for me. In fact, the first 6 paragraphs are quite clear to me but seem to miss my point (if I don't miss something again).

I just can't believe that you think the stab provides lift in a very wide range of flight situations. I think the range is rather small and even exists only if the plane is balanced on the verge of stability, as you described. Obviously, you had overdone it (or underdone) on your model, I mean too little stability for the cambered and wide chord wing. (But if you enlarge chord for more area the plane might fly slower and you're just trading chord for speed keeping pitching moment the same.) My point was that it has basically nothing to do with stab camber, that's all. After all cambering an airfoil simply shifts the lift-AOA curve upwards, making the geometric zero-lift AOA negative. What counts for stability is only the aerodynamic AOA (relative to zero-lift) and the "aerodynamic" decalage (difference between zero-lift AOAs of wing and stab). So less cambered stab and more dihedral are equivalent, aren't they?

As an example I have only the Senior Telemaster at hand because I have its parameters in the spreadsheet tool mentioned above. It doesn't explicitely account for the vertical positions of wing, stab, fuse, and C/G, but one spreadsheet calculates only one static flight situation, anyway, so you have the correct C/G position in the plan view and the side view as well. Notwithstanding any misunderstandings on my side, this tool says the STM's stab is neutral (zero lift, 0 degrees aero AOA) at only 2.8% static margin. There are still 3 degrees (aero) decalage (2 deg geometric) and thus some stability due to the quite big -0.11 moment coefficient of the flat-bottom wing airfoil. These values are for 33 ft/s cruise speed at 0.5 Cl.

Only half a degree more decalage brings the plane near to stall (20 ft/s, 1.3 Cl) and makes the stab lift (4 degrees aero AOA). Or a not unusual 17.8% static margin now brings the stab back to neutral. The recommended 14.6% static margin requires 6 (5) degrees decalage to slow down the plane (23 ft/s at 1 .0 Cl) and make the stab neutral. That's what I mean saying the range in which the stab is lifting is small. For me there's no lifting stab per se (per camber). I'm aware of the forces and moments of longitudinal balance (at least I think). No need to go off-topic any further.

I'm as well aware of the design of heavy load lifters. Aren't they called STOL airplanes? I flew one with double slotted flaps but also fixed leading edge slats. For roll stability and control, the ailerons were drooped with the flaps but not as much (and inner and outer ailerons differently). Shoulder-winger with low C/G and low thrust line. A quite small stab is enough (due to the LE slats?), just moving for trim and elevator for camber, and thick blunt airfoil. Trimmed full up for flare, it needs 30 lbs stick force for down elevator during approach, then pull stick and power lever at the same time for flare. But talk about go-around, full down elevator and flaps reduction needed - vertical moment arm between wing drag and thrust. That's why I meant that little tail power is not that bad as long as you have flaps and propwash.

Again, that was not my point as to the contest plane. As Doc recommends, I would build light - and I would build big. A simple design, quite normal and not extreme (lift and moment coefficient). It would be easy to fly. A thick airfoil, a tube spar to bear load and torque, ribs and film covering, glue only where needed would make it lightweight. I'm not biased by any design rules or habits because I just don't know/have any. So I wonder why you are so sure about that one and only design concept.

Did I miss something again? Is it possible that we tune in on the same wavelength?

BMatthews

Stabs provide lift in negative or postive directions based on where the balance point is located in relation to the 25% chord point and taking into account the pitching moments, any drag moments and lift axis to 3D CG locaton. In the case of a design with a symetrical airfoil and the 3D CG located right at the chord line on the 25% mark the tail lift will be zero for any stable trim speed. That's because a symetrical airfoil has no pitching moment and the lift force resolves as upward from the 25% chord point. As things change and we move the CG back this alters and for small amounts of camber and design changes we often see the tail lift swinging from positive to negative depending on the speed and attitude. But as the balance shifts more grossly back from the 25% point the tail operates more and more in the positive lift side until at some point it never actually goes to zero lift or negative at all (unless upside down of course). In the case of free flight models or STM's with 50% of the chord balance points it's very likely that the tail will be lifting at all times during upright flight at reasonable climb and dive angles. But free flight models don't stop there. Numerous contest FF's have their balance points even more back from there. For a while it was the style to operate huge stabilizers on long tails and run with balance locations back around 80 to 120%. BUT! They did not defy the rule that the balance has to be in front of the neutral point. It's just that due to the extreme planforms the neutral points were actually that far back.

As Mark Drela posted a few years back you can use the same balance point equations we use for "conventional" models to find the neutral points and %stability balance points for tandem wings and canards. You just input the rear wings as grossly oversize stabilizers. THe equations will happily tell you where it has to balance and the amount of stability you'll have with that location. And with such extremes obviously the "tail" is very much lifting at all times. In the case of the canard it's lifting far more than the "wing" up front.

In the case of my model the CG at the original position even after the new tail was added. The new tail was then shimmed to restore the glide to a floaty minimum sink speed. So with the CG at the original position and the tail shimmed to restore the glide it stands to reason that the new tail was again producing the same amount of lift as the old one at the same minimum sink trim speed. So far so good? The difference occured when the model sped up either due to an induced dive or from a hard speed gust that made it appear to speed up. The new tail allowed the nose to rise where the old tail was overcome and pushed forward into a vertical dive. Yet BOTH tails were stable at the original minimum sink airspeed and for a small variation in speeds in the case of the original one. So the only thing left out of this is that the original higher camber tail section produced too much lift compared to the wing at higher speeds where the new tail's lift was reduced. Yet at the min sink speed both were fine and even the original tail would recover from gentle stalls. OUt of all this the only factor that changed was the tail section.

I know it sounds like hocus pocus. I suspect that what I need to show would be a study on the model showing the Cl to alpha curves for both the wing and the two tails and overlap them in a way that represents what is occuring on the model. I'd further need to know the angles of attack of the wing in minimum sink flight and during the bunt event. If I could do that accurately and faithfully to what occured in the real world I suspect that in the first case we'd see a situation where the tail lift at some alpha sloped up faster than the wing lift. And combined with the other forces it was too much and the model went into and stayed in the dive. Meanwhile the lift curve of the flatter section apprently has a flatter or non crossing over curve so it avoids this buildup that overpowers the wing.

Keep in mind that through all this that all that changed was the tail section airfoil and whatever tail incidence angle that was required to retrim the model.

UStik

Aha, indeed I missed the re-trim. Such a diagram would be really helpful but also hard to do, and I don't want to keep you busy, either (though I love to read your explanations). But different lift curve slopes as a reason? The tail more efficient than the wing, and less efficient with the less cambered section? Hm.

Stability is always limited, it just should be limited outside the practical AOA range. It was you who called that "bubble of stability", which is small and delicate in case of a floating free-flight. Stability range depends on decalage, which has to be small on a floater, though. I guess your model was on the verge of stability, anyway, and somehow the cambered tail gave a smaller bubble than the uncambered. How about the pitching moment of the tail? After all it has to be added to the total balance of moments.

By the way, I looked up the definition of neutral point again. It's the point at which there is no change in pitching moment with angle of attack. At least it should be, but seems most calculators take only the main influences into account. And maybe it's an idealization, anyway, I mean that there may be no such point.

My problem may be my different thinking. I imagined less alpha of the model with more speed. Wing downwash (angle) destabilizes because it diminishes at lower AOA / higher speed. If the model is trimmed for minimum sink and the tail lifting, I guess it won't even lift anymore at merely 1.5 times the floating speed. But that might suffice to increase the wing's down pitch enough to overcome the effect of the rear C/G and a small stab down force. It's a nonlinear relation, what makes it quite complicated.

Dang, the spreadsheet tool even does take most influences into account. It has at least a wing Cm0, but no stab Cm at all, and it's an approximation, anyway. Besides, it's still too tedious to try several cases to get a whole curve. XFLR5 might do the trick, but the calculated moment coefficients are unreliable. Let's stop here, thanks again for your explanations.

BMatthews

You're right in that there is a very small bubble of stability on most well trimmed free flighters. And you raised a point about the tail's pitching moment as a factor that I had not even considered as a second nose down pitching force. And with such a delicate balancing act as we have in this situation it could well be a contributor. As well there's the downwash effects that would have also changed as the speed built and the angle of attack reduced. And there's probably some other effects going on that neither of us are seeing.

In any event the idea of my story about this particular model was more to point out how the stability margin calculators we use don't take quite enough of the factors into account. I think that the reason they do as well as they do is because they are set up for the "average" planform and "average" airfoil by using some fudge factor constants based on real world effects. But add in some oddball design choices such as deployed flaps or a strongly cambered wing or a radically offset heavy payload suspended below the wing and bad things raise their ugly heads and we end up with convoluted discussion threads such as this one

Lnewqban

http://www.geistware.com/rcmodeling/...calc/index.htm

UStik

Nice article, but a serious calculator does work the way described there. For instance, the spreadsheet tool I'm using is based on a thorough balance of moments and it works really well. The "popular" method of static stability margin or other ones are included in the tool just for convenience, but in no way as a basis. You may choose an initial design by specifying it and its calculated for an existing design as information. As long as it works for your design, why not.

Limitations of such a tool come from neglecting some factors that are neglectable for the intended type of design. The spreadsheets for instance are made for modern glider designs with conventional and V-tail. So there's no thrust, no biplane, no flying wing, no lifting fuse, no vortex generators, and even no cambered tail. But the drag and lift share of the fuse is factored in as well as downwash, at least as far as possible at all. After all the available practical methods are only approximations. There are no better moment coefficients factored in, probably because available values are guesses or unreliable (or plain unknown), anyway. Apart from that, the tool could be even used for other types of design if only the factors relevant there would be included in the balance of moments and if only these factors could be reliably specified (so not for "extreme" designs). Maybe the static margin concept would be useless, but it isn't needed, either.

Glauert's stability coefficient may be better than static margin, but it isn't even mentioned in the 1940 mechanical engineering encyclopedia. Instead, the explicit balance of moments is described, just mentioning only the factors relevant for "usual" designs (to keep it short). And these factors are pretty much the same as those included in the spreadsheet tool. In any case, you can't expect more than approximations, despite all efforts.

BMatthews

While looking for some articles about V tail sizing I ran across this stuff from Don Stackhouse. I don't know his background but he's quite well respected by a lot of the folks with letters behind their names when it comes to this stuff.

THe first few paragraphs in this article about V tail sizing pertains to the stuff here in a big way;

http://www.djaerotech.com/dj_askjd/d...l-formula.html

That led me to look at the articles about tail sizing. Here's a list of various responses from searching on "tail volume" on his site;

http://search.atomz.com/search/?q=ta...y&c=10&m=1&s=0

In more than one spot he repeats the need to design the tail volume with a eye on the rest of the model. In particular his repetition about how planes with flaps require a tail volume on the higher side of adequite to control the increased pitching moment of an airfoil with flaps deployed. And of course we come back to how similar the heavy lift airfoils with their heavy aft loading cusps resemble more regular airfoils with flaps hanging out. So apprently it wasn't just me that felt that some accomadation for the pitching moment is needed in our stability analysis. Suggesting that a Tvc on the larger side be used is just a way of fudging for the lack of including the pitching moment in the original equations.

Also in a few spots he mentions that the heavier the model the more need there is for larger horzontal and vertical tail volumes to control and damp out the momentums that the greater weights will generate. Again this sounds suspiciously like some of the issues that are going to occur with a model carrying a heavy payload. Granted by concentrating the payload closer to the CG location some of this will be avoided. But when the payload is 100% or more of the airframe weight even a concentrated mass is going to produce some interesting effects.

And of course designing with a larger tail volume means using a bigger stabilizer or longer tail moment arm. Both of which we know will provide more leverage to deal with the wing and payload pitching issues.

UStik

Oooh, now I feel I'm beginning to see what I can learn here!

Those tail volume issues should be self-evident. You reminded me that I had mentioned the damping issues earlier. (My memory is like a sieve.) The tool I'm using is based on a balance of moments and uses static margin as well as tail volume (maybe more) just as handy parameters. And obviously they even work, at least there are no hidden fudge factors at all. I suspect the static margin concept and the tail volume concept work really well for "normal" designs. The more the design departs from conventional wing/stab apportioning, moment arms, and symmetric airfoils, the more these parameters get questionable.

Now all balance calculators are based on the concept of neutral point. Even if the calculator does a correct balance of moments, it's based on the neutral point concept, which is the point at which there is no change in pitching moment with angle of attack. Again, that is based on the assumption that there is such a point for the wing. Hence it is really correct to use only a constant pitching moment coefficient (Cm0). Such a point may even exist, but probably only for "normal" AOAs (it's a linearization, isn't it?), and the whole concept goes out of the window near stall. Since the calculators work with derivatives it might help to use panel method tools (XFLR5) instead, even if the calculation of flow is still unreliable just near stall and especially at model Re numbers.

BUT, knowing the limitations of these tools, why not try them, anyway? Maybe they're not that terribly wrong and get us in the ballpark at any rate. After all, the alternative would be knowing nothing (and only racking our brains), so these tools are just better than nothing. I did only a quick check but it even shows strange things going on. (Or not that strange?)

Of course, the Senior Telemaster. I have its geometry and parameters in the Measure spreadsheet, which shows good accordance with reality. (Real incidence angles as well as flight speed are reproduced.) The derived characteristic values are transferred to the Design spreadsheet where some parameters are modified. Wing area and static margin are kept the same throughout.

First experiment: Weight 30 lbs instead of 6 lbs (net weight), meaning four times its own weight as load. Wing loading is 52 oz/sqft, not really a nice value. Instead of the flat-bottom airfoil (which was working at 0.52 lift coeff at 33 ft/s cruise speed) a S1223 airfoil was assumed, which might have up to -0.36 pitching moment coeff (instead of -0.11 for flat bottom). Flight speed was set to 35.5 ft/s so the wing is working at 2.2 lift coeff, which is the measured maximum for S1223. Even though the calculator enlarged the stab (and tail volume) a bit, the pitch damping ratio is now well below 1 and I should have modified some parameters to get more stab area or a longer tail (I could have done that). What looks reasonable are the incidence/attack angles and the -0.48 lb stab lift. There's a big aero decalage needed (small geo decalage due to -13 deg zero-lift AOA of S1223).

Second experiment going to the extremes: Even 40 lbs weight (70 oz/sqft wing loading), 36 ft/s speed needed despite the calculated ("theoretical", not really possible) 2.8 max lift coeff. Even smaller pitch damping ratio (of course) but smaller decalage and 1.07 lbs positive stab lift. Here we are in the dreaded small stability bubble, aren't we? But, as said before, after take-off I would prefer to fly at least 1.5 times stall speed, anyway, needing only 1.3 lift coeff and much less decalage, but still some stab lift in this case.

I don't know if the downwash calculation is still correct (but why not) and if the vertical C/G position is factored in, but the results the tool is yielding look reasonable. There are so big variations in decalage between the plane being empty and loaded (and between cruise attitude and lift/flare attitude) that an elevator alone is not enough and a moving stab is needed. Because the stab's lift is always quite small not even a stab cambering (elevator) might be necessary, an all-moving stab(ilator) should suffice. Now I know why the heavy jet fighters have one. And like there, much thrust and braking action are needed to accelerate and decelerate the big mass even to/from the quite low flight speed.

Even though I still don't know why things change that way over the load range, this experiment seems to show that a heavy load lifter design has to deviate from the conventional design "rules" in some respects, but that it can still be researched with a conventional design tool based on the neutral point concept. A small plane with really big wing loading would be hard to design for the whole weight range from empty to fully loaded. A bigger airplane seems to be easier to fly and to design, if only built light. Did I forget something?

BMatthews

I can see the light bulb over your head from here....

Truth be told I've never gotten as deep into the analysis from a numbers standpoint as you're going but the concepts you're working with show that you're on the right track.

And think about it this way. We take a 6 lb model and load it down with 400% times its weight in payload. Meanwhile the C17 Globemaster III with an empty weight 282,500 lb only goes up to a max takeoff weight of 585,000 lb. That's only a little over a "measely" 100% payload. Given this is it any wonder that the issues of dealing with a STM at 30'ish lbs or an SAE load lift model with much the same sort of empty and loaded weights needs that something extra in terms of study and providing enough aerodynamic shaping to deal with such an issue?

da Rock

And how often do those model's 400% payload match the unloaded CG EXACTLY. It would have to do that to keep from moving the CG.

Placement of cargo really has to be an exact process. You don't load 400% in a little differently than last time and see no difference.

You'd assume every team would design their loads to exactly fit the cargo compartment and fit it in one orientation only. And expect there to be a CG check after loading. If not...........

Tall Paul

I did a suspension test on one of mine, finding the vertical c.g. empty and loaded.
It really moves -down- when loaded! Several inches on a deep fuselage.
I think most of the collitch kids understand the need to have the ballast not move around in flight. There is (or was) a standard size for the ballast compartment.

da Rock

The idea isn't that it might move in flight.

It often does not occur to the inexperienced that placing 5 lbs of ballast in a compartment sized to hold 10 lbs should be done with some understanding of where the 5 lbs should be placed in order that the CG not be displaced. If often does occur to even grade schoolers that the cargo might shift, so they block it to one end or the other. etc etc