An Extended Discussion on the Design of V-Tails

From: “Don Stackhouse “at” DJ Aerotech” djarotec “at” bright.net Date: Tue, 01 Sep 1998 08:26:05 Subject: Re: [RCSE] V-tail design questions

Dave, this is a common question on this exchange. The short answer is make the total area (NOT the projected area) of your new V-tail the same as the total area of the conventional tail you’re replacing. Make the angle from horizontal on each panel equal to the arctangent of the vertical tail area divided by the horizontal tail area of the conventional tail you’re replacing.

If you have a phobia about trigonometry, a lot of folks just use an angle of 110 degrees between the vees (in other words, 35 degrees from horizontal on each side). Be careful though, when you get into MHLG’s a lot of these types of shortcuts don’t work very well anymore.

There are several different ways to size a V-tail, based on different criteria, but this is about the simplest. Most of the other methods come out with nearly the same numbers.

There are a number of articles in the “Ask Joe and Don” section of our website:

http://www.bright.net/~djwerks/

that go into this question in more detail. You might also find answers there for some of your other questions as well. Start in the “Design” category, but take time to look at some of the other categories too. I think you’ll find it time well spent!

Don Stackhouse “at” DJ Aerotech djarotec “at” bright.net http://www.bright.net/~djwerks/

=====================================================

From: “Don Stackhouse “at” DJ Aerotech” <djarotec “at” bright.net>

The reason SOME V-tails have poor rudder response is the same reason why SOME conventional and SOME T-tails have poor rudder response: BECAUSE THEY’RE DESIGNED WRONG !!!

The biggest problem with V-tail design is some old misconceptions and a lot of mythology that frequently ends up causing them to be sized improperly. I posted some details on this just a short time ago on RCSE, as well as in an article in NSP’s new catalog. A given T, V, or conventional tail will all have essentially the same control authority if they have the same total area. Too many people still have the idea that you can give a V-tail the same projected area as the supposedly equivalent conventional tail, which results in an undersized V-tail. Let me repeat that, just in case you missed it: The response is poor NOT because it’s a V, a T, a W, a Q, or any other wierd configuration you dream up, it’s response is poor because it’s TOO SMALL !!!!

The Monarch hlg has been available almost since the beginning of it’s history with either a conventional or a V-tail. Both versions have the same total area, and essentially the same stability and control response. BTW, most folks who’ve flown them tell us that control response (including “rudder”) is very good. The vast majority of Monarchs are built with the V-tail, and I believe their contest record speaks for itself.

I’ve flown both versions of the Chrysalis and they both fly the same. The first V-tail prototype had a different area and dihedral and it didn’t fly the same. Nothing mysterious to any of this.

With both the Monarch and the Chrysalis we first developed the V-tail version, then designed a conventional tail equivalent to it. There was no significant change in handling or stability.

If your model is insufficiently stable in pitch, make the tail bigger. If the yaw stability is insufficient, make the tail bigger. Change the tail’s dihedral as necessary to make the extra area help pitch or yaw as appropriate. If the model dutch rolls excessively, reduce the dihedral of the wing and/or change the dihedral angle or area of the tail to increase its yaw influence. If it tends to slip into a spiral dive in steep turns (spiral instability), reduce the dihedral and or area of the tail, or increase the dihedral of the wing. It’s really no different in concept than the way you fix problems in a conventional tail.

BTW, remember that there is a tradeoff between dutch roll tendency and spiral stability. Generally speaking, if you improve one you hurt the other. If you want really nice turning behavior, it may be a worthwhile tradeoff to put up with a little dutch roll. It all depends on the design priorities of that particular model.

If your airfoil has hysteresis or deadband (and BTW, thank you Cameron for your excellent explanation of the difference between the two), don’t blame the tail shape, fix the airfoil.

If your stability is good but the rudder response is inadequate, increase the size and/or travel of the ruddervators or increase the dihedral of the wing (particularly at the tips). How to decide which parameter to change? Just watch during flight tests. Does the model yaw but not roll when you give it a big rudder input? In that case the rudder is doing it’s job, but you need more dihedral in the wing. If it doesn’t yaw noticeably, then you need more rudder effect. Remember that whenever you make a change like this, you will probably have to make additional changes in the other parameters to get the handling in balance again.

Note that none of these problems are due to the configuration of the tail. They can occur just as easily in a T, a V or a conventional tail.

There are a lot of good reasons to use a V-tail. Properly designed they can have better spin recovery than a conventional or T-tail. They keep the major mass of the tail assembly close to the tail boom for minimal torsional loading during groundloops. They also keep themselves up out of the grass during groundloops and landings on rough ground. They usually have less joints, less interference drag, less stress concentrations, less parts count, and less weight. Their Reynolds numbers are usually better, a very significant factor for our models. They also keep most of themselves clear of the wing wash without being subject to the blanking and deep stall tendencies that can plague T-tails.

They do need a mixer, but there are a number of fairly simple ways available today to deal with that. About the only big disadvantage to V-tails is that many people don’t understand them, and might therefore be afraid of them. Hopefully I’ve helped that situation a little.

Don Stackhouse “at” DJ Aerotech

============================================================ From: Joe Wurts JoeWurts “at” compuserve.com Date: Mon, 31 Mar 1997 00:53:13 -0500 Subject: Re: [RCSE] V-Tails

After watching Gavin and others bat back and forth the V-Tail saga for a while, I had to chime in.

I am going to come down (mostly) in agreement with Gavin. Also, Djaerotec has a lot of good comments.

Now, for my nits, and corrections to my version of the “truth”:

From DJaerotec: > the inverted >V is structurally more tricky (which also means probably heavier), and >suffers from increased tip losses in most steady state flight conditions due >to the extreme anhedral.

Extreme anhedral, extreme dihedral, the losses are quite similar. For that matter, just how nose heavy are you flying? Most airplanes are set such that the tail loads are low, and the trim drag (a function of tail Cl squared, kinda), is very low. So a 5% change in a number that is 2% of the total drag, gives a .1% change in the total drag. IE, inconsequential.

From Gavin: >Unfortunately all of the fixes result in loosing the >benefits from the V-tail. The fixes include large surfaces, thicker >airfoil sections, longer tail moments, more throw.

And from DJaerotec: >As far as thicker sections, this is definitely not true, and we just might be >getting to the root of your problem here. A thick section generally (although >at our Reynolds numbers, not always) has a lower “lift curve slope” than a >thin one. This means that for a given increase in angle of attack, a thin >section sees a greater increase in lift coefficient than a thick one. For a >tail seeing a given disturbance in pitch or yaw (resulting in a given change >in its angle of attack), the thin section will generate a greater correcting >force than the thick one….

I think ya got that one backwards. Thicker tends to have a higher Cl/alpha. For the thickness (thinness…) range that we play with, call it 2*pi, and be done with it. The change in the lift curve slope with thickness only becomes apparent above 12-15% t/c (generalization warning!). Ref: I’m looking in McCormicks book as my wife has our copy of Abbott & Van Doenhoff (sp?).

From DJaerotec: >In fact, if you’re talking about only a yaw correction, or only a pitch >correction, you should theoretically have a thinner airfoil on the V-tail, so >that the effective section in the plane of that particular control axis is >the same as the conventional tail.

Huh? Thickness should be measured normal to the plane that the LE and TE are on. Anything else is voodoo aerodynamics.

From DJaerotec: >Improving the lift curve slope will stiffen the response of just about any >tail design, not just the V.

If you can find an airfoil that does more than a couple percent better than 2pi, let me know. In general, the lift curve slope delta between airfoils (at least, ones that do not have “problems” such as bubbles popping, separation, etc) is basically nil. From a pure stability standpoint (linear region only), a 8020 is just as good (or bad) as the MB253515, neglecting the non-linearities in the Cl/alpha curve that happens at low Re. And, I’d strongly recommend picking a tail airfoil that is pretty linear, otherwise your handling qualities are kinda “fun”.

From Gavin: > While the lack of rudder is perhaps not as important to the average >thermal pilot or HLG, who can use a ton of differential and does not bother >with the rudder, it is very important to F3B. It is important to fly in >straight lines, and rudder response is vital for tow… >…Another concern for >F3B is the V-tail wobble after speed turns. This is NOT a dutch roll >problem, or a poor design as someone pointed out (Dutch roll is a combined >roll/yaw motion), but has to do with V-tail yaw damping

From DJaerotec: >That it’s a yaw damping problem I agree. That it’s inherent to V-tails, >sorry, I don’t buy that. As I said before, in my experience the equivalent >V-tail has the same area and airfoil as the representative conventional or >T-tail. This isn’t just anecdotal evidence, there are published mathematical >derivations by well known designers (Stan Hall, the noted designer of full >scale sailplanes, for one) that reach this same conclusion.

Aaah, theory! Yes, in theory what you say, is true. That is, linearized theory, where each v-tail is in air that is not affected by the other v-tail. In practicality, a v-tail in yaw, is kinda similar to a bi-plane, and suffers from the same maladies. The two tails are not as effective in concert, as they should be by the derivation mentioned above. The net result is less stability (and damping) than is predicted by standard linear theory. Go play in a wind tunnel for a while and collect some data, I suspect that you will agree. BTW, this effect is most noticeable for low aspect ratio tails, and also for tails that have less included angle betwixt them than they should. WACO should be congratulated for using high AR tails!

All that said, I still feel that there is a strong place for V-Tails in model design. But, they are not the solution to all problems. And where (strong) authority in yaw and/or good yaw damping is very important, you might wish to consider going oversize, another tail configuration, or learning to live with the reduced characteristics in yaw inherent in V-tail configurations. I remember reading somewhere that to get an equivalent yaw stability, take your conventional vertical tail area, divide it by the sine squared of the dihedral angle, and then divide by 0.7. The 0.7 is to make up for the destructive interference between the tails. Of course, this is over-simplified, but gets the point across.

===================================================== From: Tord tord.s.eriksson “at” swipnet.se Date: Mon, 18 Aug 1997 02:55:47 +0100 (GMT) Subject: Re: [RCSE] V-Tail vs. Conventional

On Fri 15 Aug, Ron Quintana wrote: > What are the advantages of a V-tail over a conventional elevator and > rudder, assuming you wish to use a seperate rudder and elevator > servos and channels, if any? Are there basic guidelines on how to > convert one form to the other, in regard to span, chord, aspect > ratio, total area, etc.?

As far as I’ve learned from the text books the advantage is theoretically small, at best and very seldom used on real planes, as the stabilator there usually is set so that the turbulence over the wing hits the stabilator only when the plane is on the verge of stalling, as to give the pilot ample warning. Modern fighters have them under the wing, usually, where it will be effective at high angles of attack. T-tails are the heaviest, but aerodynamically very clean, while V-tails are mechanically simple and lighter, while a low-set elevator is prone to damage during landing, if you don’t land on tarmac and use a landing gear. So, for being simple mechanically (but you need a mixer radio or a mechanical mixer!) V-tails are often used on models, but next to never in the real world.

The size needed is quite easy to find: If you have a conventional tail think of each surface as sides in a box, the stabilator being the bottom and the fin being the side, the width being the average width of tail and stabilator. To get a workable V-tail, you draw a V from the middle of this box up to the box’ upper corners and voila! you have your V-tail. The enclosed angle should not be less than 90 degrees, preferably 110 -120 degrees to minimize interference.

There have been a guy in the UK who, after having half a stabilator torn off during competition flown successfully with that half missing, even in later competitions. The idea of having just two surfaces is to minimize interference drag, half a T-tail probably being the best and by far the heaviest and most awkward structurally.

The Williams VjetII, designed by Burt Rutan, with ideas from Mr Williams, is one of very few jets with a V-tail, the Fouga Magister jet trainer being another one. Again, T-tails or tails with very low-set stabilators, like the F-4 Phantom fighter-bomber that have stabilators canted downwards to lower it’s centre of lift, is probably the aerodynamically cleanest, but heavy, while the V-tails never are out of the wing wake, thus not ideal, but at the same time they are very difficult to stall.

I think the classic Learjets have the best design: a short T-tail with delta finsat the bottom. After that was introduced the jaw-dampers and a lot of other artificial stabilizing could be scrapped, as that tail, which without all the hydraulics is lighter than the old, without deltas, and much more efficient.

Propeller aircraft often tries to keep the stabilators out od the slipstream, as the vibrations will be noticable in the cabin, as sound and/or vibrations, and the vibrations also decreases the stabilator’s design life. But, as fitting the stabilator to the tail cone is the lightest solution, they instead use stabilators canted upward, like they were part of a V-tail with a centre fin.

You can even find a handful of passenger-carrying modern jets that has this arrangement, including the 747. The ideal position on a powerful jet aircraft is in line with the jet pipe, but then you need to employ odd tricks, like that on the Phantom.

So building-wise, V-tails are easy, weight-wise light, but setting up and efficiency is at the best doubtful! Some V-tailed aircraft are said to have nastier spins than their normal-tailed sisters, while cruxiform (stabilator halfway up the fin and forward of rudder) promises to be the best overall performer, inverted or not! Not the lightest, not the most efficient, but next to fool-proof and used by many STOL aircraft and gliders, too! Promises excellent inverted control, not too demanding on tail strength and looks very nice! Even an old Bergfalke has it’s stabilator halfway up it’s rudder - a real bonus when you train spinning!

========================================

The following is an anthology from RCSE’er Mick.

V-Tail Experiences (FYI)

Pitch up tendency:

The most likely cause of the tendency of the V-tail converted Synergy to pitch up when large rudder deflections are used is – differential in the V-tail’s rudder movement. It may not look like there is differential, but measure it. If you find some you’ll have to get rid of it by control system geometry adjustments. The computer radio probably won’t be able to correct it in the program only.

Unintended Spins:

The things that promote spins (there are others - these are the big ones) - Heavy wings, Aft CG, small horizontal tail, and small vertical tail.

The things that reduce spin tendency - the opposite of all of these things.

With the bigger tail, the spin tendency will be reduced. Most planes modified to V-tail benefit if the tail moment is increased by lengthening the fuselage- - as well as providing an appropriately sized set of surfaces.

Joe Wurts provided an explanation to this effect a while back. It bears repeating (and a little paraphrasing). The upward-going V side is a little more effective because the opposite surface is acting as a large “winglet” for it. The upward V not only increases the pressure on the top of its own surface, but also (to a limited degree) on the top of the other surface. So rudder inputs generate a slight amount of up trim as a by-product.

Unless you have some funky inverted V thing. Computer radios can most definitely (and easily) tweak out the differential. On Futaba radios (such as the Super 7) you adjust the control throw endpoints by servo, not by function. Simply set the “down” travel for each V servo to be a bit more than the “up” travel.

On the Stylus with glider card, the V tail setup has a built in differential adjustment. It allows you to tweak the mixing (up to +/- 20%) so that the throws are matched side to side (even if the geometry is off a little) and the desired differential is achieved. The standard Stylus can do this with a general purpose C Mixer.

I have found that with V tail HLGs (Monarchs, etc.) that about 3 - 5% more down than up travel was required to get no pitch up with rudder input.

V tails- DECALAGE

Michael Pitoniak asked about how to set the incidence of a V-tail. I just built a Mirage and here is what I did. The instructions call for the tail to be set at -3 degrees to the best of my recollection. My first thought was to use an incidence meter directly, but that proved to be pretty hard.

Then I got the idea of making a jig to hold the tail from the shucks. That seems to work pretty well. Basically, I laid each tail in its lower shuck and used the incidence meter to make sure it was dead level. Use tape or rubber bands to hold it in place. I then made a triangular foam block with the correct angle to give 100 degrees between the tails and cut off the corner that would be near the fuselage. This needs to be 4" to 6" thick. Next, I propped the boom up and shimmed it until it was level. The last thing you need to make is a shim that sets the incidence angle. Lay that shim on the building table, set the triangular block on top of it, and put the V-tail in its cradle on top of the triangular block. You should now have the V-tail set to the proper angles. Sliding the V-tail up and down on the jig block and sliding the jig block in and out will give the vertical placement of the V-tail onto the boom. When satisfied, tape it all together and use the jig to sand the boom’s contour into the root of the tail by wrapping sandpaper around the boom and sliding the tail along it. Then you can use the jig to hold the tail in position while gluing it.

This seems like a lot of work to complete the instruction “glue the V-tails to the boom at a negative 3 degree incidence angle”, but it’s the best I could come up with.

Larry Hardin

V tails- DECALAGE

This seems like a lot of work to complete the instruction “glue the V-tails to the boom at a negative 3 degree incidence angle”, but it’s the best I could come up with.

Larry Hardin

V TAIL DECALAGE

Someone posted a question yesterday (?) about the proper alignment of the v-tail on his ship, but I don’t recall seeing an answer. So here goes. Some of you may have things to add to this, too. (No one around here seems shy about expressing their opinions anyway. ;) )

The simple answer is that you should use the alignment shown on your plans/instruction sheet. The proper angle is established during development by calculation and little tweaking, so it is best take advantage of the designer’s efforts. Your main worry should be in putting everything on straight. If the fuselage is of a molded type with a mounting platform built in, use it. If you’re building it up, take the time to get it right.

The long answer is that what we’re interested in is the angular difference between the wing and the tail, called decalage. This angle depends on the airfoil used and the speed at which you want to fly. It is entirely possible for two ships to have different amounts of decalage, even though they use the same airfoil–it depends on how fast you intend to fly under normal trim. This is what the polars for an airfoil section, which show lift and drag at different angles of attack and speeds, are used for. (This is another reason why we should support the work of Micheal Selig & crew, and Herk Stokeley, his publisher. Using the data from these charts takes a lot of guesswork out of establishing the decalage for a particular design. Generally speaking, decalage for our sailplanes will be in the vicinity of 2-3 degrees. Fast slope soarers will have less. Powered planes that are supposed to go fast or fly inverted will often be set to 0 degrees.

There was also a question about whether an incidence meter is useful in setting up the tail. The answer is yes! Since we’re really interested in the difference between the wing and the tail, set the plane in one place and don’t move the fuselage while taking your measurements. Take readings from both sides to check for warps or misalignment, just like you do for wings. You can also place a carpenter’s level across the tips of your v-tail to make sure you don’t have one side higher than the other.

Mark

V-TAIL SET UP

I use ball joints on the V-tail of my F3B model which works very well. Each tail half plugs onto a steel wire joiner and incidence peg, which makes transport easier. One piece (bolt on) V-tails can be surprisingly difficult to fit into model boxes… For the actual horn I use a piece of wire (around 3/32" I think). The wire is recessed and glued into the root end leading edge of the control surface, with the end nearest the tip bent 90 degrees and glued into the surface (standard torque rod type setup). The other end (at the control surface root) is bent down at the appropriate angle to form the horn. Onto this end I solder the ball part of the ball link. Dubro make a nice unit that has a female threaded part instead of the conventional bolt. The end of the wire is pushed into the hole and soldered making a very secure connection.

| |
| |
|_| End of wire horn    
    |  Push wire into hole and solder
    V
__   __
|  | |  |
|  | |  |
|  |_|  | Dubro ball link
|_     _|
\   /
| |
/   \
|     |
\___/  

The plastic cup part of the ball link is then screwed onto the
pushrod allowing it to be snapped on and off. This task is easier
if you cut the rear of the fuse back slightly to allow better
access to the ball joints.  

    |                  |     | Control surface
    |                  |    /
    | V-tail surface   |  /   
_____|__________________|/
                    / | <-- Wire control horn
    Pushrod           /  |
=====================/===o <-- Ball link
____________________/
                    \
                    \ Rear of fuse cut back to
                        allow access to ball joints