Here's some practical modelling experience on the flutter issue. I've read all the posts and the subject has been talked all around in circles and I'm sure lotsa people are quite confused. I provide simple and practical expalanations to club members when they ask about flutter and the solutions do work.

Flutter requires a driving force. It is a result of oscillating pressures at the wing trailing edge - an alternating vortex stream. This is the same driving force which causes flexible steel power poles to sway from side to side in a strong wind. The frequency of vortex shedding increases with increasing airspeed.

On a wing the result is oscillating torsional forces (wing twisting). On a control surface, the surface is forced to flap up and down against it's eleastic restraints. Forget the wing for now and consider the control surface. When the frequency of the oscillating force (airspeed dependent) matches one of the natural frequuencies of the control surface/linkage system, then resonance will occur. What is resonance? In simple terms, the energy from the airstream is admitted freely to the control surface/linkage system with little if any resistance - the result is usually very spectacular. The control surface motion effects the wing in interesting and complicated ways, mostly resulting in wing structural failure - quickly.

How do we fix it? Well we can't. It will always happen - at some speed. What we want is for resonance to occur at some speed outside the expected flight envelope, ie. a speed that we will never attain. This is most conveniently a very high speed, especially a speed that the airplane cannot possibly attain. At high speed the vortices are shed at a high rate. A solution is to make the natural frequencies of the control surface/linkage system high so they match the vortex shedding rate only at high speed.

To increase the natural frequencies of the control surface/linkage system the following need to be done;

1. Increase control surface stiffness to weight ratio.
2. Reduce the moment of inertia of the control surface by redistributing the mass more evenly about the torsional axis (or axis of rotation - hinge line).
3. Reduce the mass of the linkages.
4. Reduce the moment of inertia of rotating linkage components.
5. Increase the stiffness of the linkage components.

Other things that work;

1. Remove all free play. Free play does not change the resonant frequency but it reduces the effectiveness of any dampening.
2. Make trailing edges thin. This reduces the magnitude of the driving forces by reducing the size of the vortices.
3. Make wings stiff in torsion and bending. This delays the effects of resonance (the "interesting and complicated" stuff I referred to in para 3).
4. Make the tail boom stiff in torsion (to resist rudder effects).

Fortunately, these things are easy to do at the small scale of model aircraft. High stiffness to weight ratios come with the territory. Of course, despite nature being on our side, there is always one person in every club who manages to get it wrong, to the considerable amusement of the other club members as his model transforms itself into a balsawood shower.

On my jets and pattern models I use the following;

Ailerons - Titanium pushrods with MK ball race links both ends. Aluminium servo arms. Digital high speed servos. Ailerons are lightweight, stiff, and have many hinges (at least 5). Trailing edges are inlaid hardwood and faired to 1/2mm.
Elevator - CF tube - large diameter, thin wall - titanium threaded ends, MK ball race links, etc. (same as for ailerons).
Rudder - same as elevator or cables with tensioning system.

With these I cannot get any flutter. If I do then the next move is to consider mass balancing (reduction of control surface moment of inertia).