There are so many issues being observed here that not all may be the same problem.
When I refer to wind seeking, the best example I can give is a 8ft machine with an offset of just under 3". This one refused to furl under any condition and the message was finally driven home when part of the tail mechanism fell off, leaving the machine with no means of steering it into the wind. The fact that it had an offset and there is thrust on the prop it should have turned round and stopped or faced the other way. In fact it held directly into the wind for about 15 minutes at full power.
Although it would not furl when the tail was intact it could be stopped by using a manual control to turn the tail at right angles. Using this manual control it held virtually full power until the tail was at about 75 deg and then suddenly the prop turned at right angles to the wind and it stopped. I tried a rope attached to the tail at one point with the tail fixed and it took a large force to pull the prop to about 45 deg to the wind then suddenly the force vanished and it then turned 90 deg to the wind and stopped.
The simple furling theory suggests that the thrust is proportional to the power extracted and so a scheme that breaks the tail away from the stop and maintains a constant restoring force should produce furling at constant power. Whatever you do in practice the thrust appears to be less than it ought to be and it always furls at a higher speed than predicted. This discrepancy varies drastically with the type of blade and to a lesser extent on the loading. Some furl at nearly constant power but at a higher speed than predicted. Others seem to hold into the wind until the blades are at a significant angle with the wind and then suddenly go into furl and the power drops dramatically. The first type will run with the blades at virtually any angle to the wind and the angle changes just sufficiently to keep the power near constant. This is surely how it should work.
The second type will drop power drastically when it furls and will not pull back into the wind until it drops speed considerably. This method of operation is safe and some may like it but it is a sure indication that something other than thrust is acting on the blades.
The disaster case is one where the offset is too small and the thing never does furl and power never drops, the blades never get at a sufficient angle to the wind to prevent the power rising with wind speed.
I am not sure about the described cases where things are claimed to furl at one wind speed and then loose control at a higher speed, I have never experienced this condition but I have never used a machine working stalled . I still believe that these don't really furl but stay near constant power from stall operation until the blades break stall. It's no good looking at tails, you need to see the angle of the blade disc to the wind and that is the only way you can see that it is furling and until the blade disc is at something well over 45 deg to the wind there is no reason to believe it is furling at all. I suspect that these large machines with drastically under powered alternators for the very high wind speeds can produce enough power to burn the alternator with the blade disc at 80 deg to the wind.
You really have to get the blades at this angle or greater in winds within half the maximum rating of the alternator.
Several people have said that they can't imagine how a tail can be at right angles to the wind and the prop be flying at high power into the wind. To be honest I can't imagine how this happens either but I have seen it and that is why I emphasise that you must compare the blade disc angle to the wind to be sure you are furling and forget looking at tails.
I really can't imagine what a mess the air flow is behind the prop and short tailed machines in particular may have their tails in an air flow that bears no resemblance to the true wind direction.
I suppose the conclusion from this long ramble is that what I call wind seeking is the difference in thrust along the alternator axis between the theoretical value and the real thrust. I am only considering the thrust as that which acts along the alternator axis the other forces such as drag and others from phenomena I don't understand don't produce rotation about yaw from the alternator offset. They certainly exist but are not those considered in the basic furling equations.
Flux