To be honest I have had no experience with visco clutches until now. I was lucky enough I had one spare I could open up and study. The clutch basically consists of the body (2 disks, one pressed in the other and either glued or siliconed), the visco fluid - very sticky, white to yellowish coloured. I have no idea how much of the fluid actually is supposed to go in there. There are 2 metal plates (disks) inside that rub against each other. The viscosity of the visco fluid between them does that the clutch gets 'rotated' by the engine but not with the same speed- there is some viscosity slip. It's like you had 2 waffles with honey in-between. When you rotate one the other catches up with some delay, owing to the nature of honey.

Then, apart from the central bearing, the only part I remember seeing is the rod and the valve. The valve is very simple - there is a hole through the front plate with a piece of metal strip laying slightly above over the hole. When left as it is the hole is open as the strip rests some tiny distance over the hole. When the rod presses against the strip the strip is forced to lay over the hole blocking it.

Now, when the temp is low, the bi-metal strip outside presses against the rod that presses against the metal strip that closes the hole. When the clutch heats up the bi-metal strip bends up releasing the pressure off the rod and the metal strip inside raises to its original position over the hole thus effectively opening the hole. I suspect then the visco fluid gets rotated thru the hole and causes additional friction not only between the 2 plates but also between the plates and the front of the body of the clutch.

Now to your questions about possible modes of failure :

This (only valve inside, there was no other one) is a very simple design and I can't think of any way how the valve could fail. The clutch doesn't get opened and the only part that gets stressed seems to be the metal strip that closes the hole inside. It travels up and down, being pressed by the bi-metal strip. The travel is minimal and I don't think there is anything that can go wrong with it unless some drunko in the factory welded the strip on improperly.

There are 3 parts on the clutch that are either dynamically stressed or stressed by heat. It is either the pin thru which the visco fluid leaks with time (this was my case) or the central bearing which can either fall apart or leak or both. On my visco clutch I saw no other openings thru which the fluid could escape. There were those 2 only - the pin and the bearing.

In other words - IMHO there is nothing that can go wrong with the internal parts of the clutch because if its design. The only reason a clutch would lock up, I could imagine, would be when somehow the 2 internal metal plates (disks) somehow get stuck with each other because of the heat and friction after the fluid escaped.

Also the bi-metal strip outside could somehow fatigue with age and not bend (that way the valve wouldn't open). Though this is tested very easily - heat it up with hot air and see if it bends. When it bends it bends very visibly.

One way or the other, if the bi-metal bends and the rod travels up by itself it means you have only lost visco fluid.

Drill a hole (the outter part at the bi--metal, where the reservoir is), pour some high viscosity fluid in, seal the hole and the clutch must work. I filled my clutch full with the oil additive. I had no idea how much should go there, was only a guess. An original clutch stops after half a rotation, mine does after a rotation and a half. This is because the oil additive isn't as dense as the original filling but it still does its purpose well and when the clutch locks I can finally hear some roar by the fan.

So, to be concise, I think, after having the clutch apart, that the reason for 99% of visco clutch failures is a worn out entrance pin thru which the fluid escapes.