Tjingeling,
dusted off some old back issues of RCM today and found an old article from August 1976 where it is suggested that ailerons(or rudders) should be made thicker than the TE of the wing(or stab) to reduce flutter AND drag...! The article seems to be well researched and is supported by wind tunnel data. The author(W.D. Mitchell) summarizes at the end of the article:

"...make all your control surfaces about 12 to 15% thicker than the trailing edges of their respective wings or stabs. It´s the cheapest, easiest and most effective way to decrease drag and flutter, and increase control surface efficiency......"

Attach a figure from said article to illustrate and now I wonder why we have seen so little of this approach since it appears to be almost embarrasingly simple to apply. Cannot recall having seen any kit or design using this approach - why is that....?

The author also discusses the "usual" remedies such as sealing the gaps and mass balancing and he maintains that ".....mass balancing cures flutter if exact..."
............Cheers/Harald

>>So what's left? Play due to worn ball-links and/or kwiklinks. <<

That's why I avoid z-bends.. Almost impossible to make them truly square and slop-free.

Okay, everyone.

Another easy method.

You may remember hearing about an engineer named Kamm. He did some research on blunt trailing ends, and concluded that a taper less then 4:1 was wasted, a completely flat trailing edge was every bit as efficient as a steep taper in reducing air drag. Remember the Chevrolet Vega wagon? The one GM called the "Kammback?" That's what that was about.

One of his results was to find that an UN tapered moveable control surface was almost flutterproof, and only slightly more air drag than a shallowly tapered surface. A current example of this is in the Ultra Stick line of airplanes, although they do cheat a bit in having he trailing edges rounded.

So, if all else fails, rsieminski, go back and replace the elevons with flat plates and try that!

Ailerons that are flat?
. Might as well try that.

Bill.

Trailing edge "thickness".. Really fast planes seem to have missed the benefits of blunt edges.
Practice over theory.
The flutter most of us encounter is due to poor systems geometry and installation.
Tight systems don't flutter.

Paul"
You missed the point. Most of the "Really fast planes" have a taper ratio far greater than 4:1. But there were exceptions. Check the tail feathers of the X-15. No trailing taper at all.
This is true.
Wrong, sir. Every system has a point at which it will flutter. Effort is always made to have that point outside the normal operating range, but it's there.

Viewing the crash:
It fluttered,
. He stuttered.

Bill.

Are your trailing edges rounded off. I read somewhere in a RCU post someone describing that a rounded trailing edge induces flutter even though every instruction manual you see tells you too.

I'll see if I can find that post.

Flutter
Caused by flex in the structure
cures:
Reduce flex in the structure. Flutter is when the wing and the aileron vibrate out of phase with each other.
a. Sheeted wing is stiffer than an open structure.
b. Dual aileron servos with straight pushrods are stiffer than a single servo with torque rods
c. A tight hinge gap is stiffer because thee is less exposed hinge to flex. Otherwise, gap decreases response because of air leakage but has no effect on flutter.
d. A sealed hinge gap is just mechanically reinforced by the sealing strip. If sealing your hinge gap stops flutter, the hinges were installed sloppily in the first place.
e. linkages should be attaced as far out on the control horn as possible. For the same torque, there is less force in the pushrods/clevises
f. Larger-stiffer control horns help
g. Built-up sheeted control surfaces have a much higher natural frequency than solid surfaces. Also they are stronger because the grain direction always is aligned with the local forces instead of being unidirectional as in solid wood.
h. Sharp trailing edges allow less force to be transmitted to the control surfce by the alternating vortex sheets as they depart the control surface. When ONE sharp trailing edge is impractical, two sharp trailing edges can be used (Squared trailing edge). Rounded trailing edges are worst by this theory, both in terms of flutter and control response.
i. The control surface should be slightly thicker than the aerodynamic surface to which it is atached. This provides energy to reattach the flow that has to separate to jump the hinge bevel. Also improves control response at small deflections that would otherwise be submerged in the separation bubble.
j. Mass balancing is effective. Mass balancing is also awkward. If a model airplane needs mass balancing, nine times out of six it's because the other elements of the system are still not adequately stiff.
k. Aftermarket heavy-duty servo horns are stiffer
l. E-Z connectors are asking for trouble. Use clevises.
m. Reinforce the area of the control surface where the control horn attaches. If solid balsa, at least pierce the area and soak it with thin ca. Even better is a hardwood insert. This gives a firm support for the control horn.
n. Any bend in the pushrod must be eliminated. This includes z-bends.
o. Install a hinge no farther than 3" apart.
p.Hinges should be within 1" of the ends of the control surface.
q. Install the control horn through a hinge. This transmits the forces directly to the structure without springing the control surface.
r. Monokote or oracover are stiffer than low-temperature films. Tight covering adds a lot to the stiffness of a structure.
s. Mount the control horn in the middle third of the control surface-not at the end. This reduces twist in the control surface.
t. Ball-links put an off-center load on the control horn. Avoid ball links unless there is some weird geometry that makes them necessary. Then use the stiffest control horn that you can find.
U. Mount the servo in a reinforced mount that cannot easily deflect.

If you do all this and you STILL have flutter, send me pictures.

>>Remember the Chevrolet Vega wagon? The one GM called the "Kammback?" That's what that was about. <<

No, that was just marketing. The rear glass of the Vega was still inclined at about 45 degrees AND it had a 2" high spoiler formed in the hatch above the glass. It was Kamm in name only.

You wanna' see a Kamm profile? look at a .50 cal projectile.
Kamm also demonstrated that the squared-off end could be about 50% of the frontal area, and he did some work about the radius between the taper and the blunt end.

.
My bad!
I keep neglecting the applicability of Mach 3.5 flight to Mach .035 flight.
I assure you the very next plane I make from Inconel X with an XLR rocket motor will have a wedge shaped horizontal.
And I'll make every effort to keep the plane flying slower than it possibly can in any eventuality, to keep below the flutter threshold.

Bill,

I have heard of Kamm, and the Kammback. There is no mystery about the drag vs. the taper at the back end of a body. It is simply another to way to say that the shape of the body behind the separation line doesn't have much effect on drag. The result that mention regarding flutter is new, and counterintuitive to me. Could you provide a reference?

Thanks,

banktoturn

Banktoturn (You model guys never use the rudder!):

Here's a reference right on the thread. Been too long for me to give a quick text reference, sorry. Maybe Jim has one?
And in spite of Tall Paul's implications, it works at Mach 0.035 too. But you'd probably have to get up to 0.1 Mach (60 mph) before you started noticing much difference in air drag.

It is, of course, speed dependent. At lower air speeds the boundary layer will follow a steeper taper, as the air speed increases we reach the point where boundary layer separation is almost guaranteed. This was the case with the X-15.

A later Mach 3.5+ example that did not use the blunt trailing edges was the A-11/YF-12/SR-71 series airplanes. This series was originally designed as high speed high altitude fighter/interceptors, capable of acceptable maneuverability at sub-Mach speeds as well as the high speed cruise. And they had to take off from the ground under their own power. Even so, while the public name was "Blackbird" (on the '71 planes) the private name was "Sled." It was a hell of a ride. How about 30,000 ft/min climb at 1400 mph?

But back to the flat/blunt aileron and flutter resistance. By having the parallel upper and lower surfaces the aileron reverses the taper of the airfoil, tending to compress the air flowing from the wing, and in doing so the boundary layer is more firmly attached to the surface. Then at the blunt trailing edge there is a sudden separation, but it is away from the control surface, and with any speed on, the turbulence generated wont affect the control surface at all. When the surface is deflected by control command input the air stream force itself resists the deflection, and that force prevents the onset of flutter.

Put the churn on the surface with flutter.
. Fly a whille, then you'll have butter!

Bill.

MY information on this came from doing some vortex flowmeter engineering.

If you stick a strut of relatively unimportant but uniform cross section (in this case, a wing) into a flowstream, it will shed alternating vortex sheets. These are caused by the flow separating at the trailing edge. As the vortices shed they impart a reaction force onto the structure. The frequency of the vortices varies directly with speed:e.g.: twice as fast at 60 as at 30.
Once the vortex frequency hits the resonant frequency of the structure/control system upstream of it, all hell breaks loose and pieces start coming off.
If you want to SEE alternating vortex sheets, go watch a flag. The ripples in the flag are the alternating vortices from the flagpole. In a brisk wind, the flagpole goes into resonance and you hear the flag rop slapping against the pole as it shakes.
In Vortex flowmeters, we shape the strut to MAXIMIZE the force transmitted back to the strut so we can read it electronically and infer flowrate. In airplanes it's important to MINIMIZE the forces from the vortex sheets, so you make the control surface sharp-edged so the mass of the vortex sheet is minimized, or square edged so that when a vortex forms there is no reaction component perpendicular (normal) to the separation surface.

Rounded trailing edges provide a surface for the vortex sheets to push against as they jump off, so the excitation force on the structure is greater.

Kamm's work showed that the blunt afterbody dragged along a separation bubble and the free-stream streamlines were very similar to a fully tailed shape. I extrapolate that to mean that a squared-off aileron would act as if it has a greater chord than actually exists.

IN PRACTICE: Trailing edge shape is probably about 3 orders less important in preventing flutter than having a tight system. As Paul said: Tight systems don't flutter. The obsessive compulsive way to say this is to say that tight systems have a resonant frequency that corresponds to an airspeed greater than terminal dive speed. With a tight system, You can't GO fast enough to get flutter.

"The obsessive compulsive way to say this is to say that tight systems have a
resonant frequency that corresponds to an airspeed greater than terminal dive speed. With a tight system, You can't GO
fast enough to get flutter."
Commandment #11!