Let's clear some things up.
We'll start with safety. Please (I'm not begging), do not use 2×4's to test the fan clutch. If we are testing the fan clutch then we don't know if it is engaged or not. If it is unknowingly engaged or engages automatically, and the heavy metal blades grab that piece of wood, whether or not the clutch works will be the least of your worries. Roll up some newspaper or a magazine and you'll achieve the same results.
The NHC250 uses a fan clutch made by Evans. They were the sole manufacturer of this clutch. The 8.3L used a different design that was manufactured by several companies: Evans, Horton, etc. Comparing the A0/A1 clutches to the A2 clutches is comparing apples to elephants. Sed6 has a NHC250 so let's work off that. Pictures below for those that like them.
In the first picture you can see the massive rear bearing and the small front bearing of a NHC250 clutch. These are sealed bearings. By design they have more drag then other types of bearings. Bigger bearings have bigger drag.
The second picture shows the back of an 8.3L clutch. While not disassembled, you can easily see that there are no massive bearings. The big bearing in the back of the NHC250 clutch is the main reason the fan turns when not engaged, bearing drag. A NHC250 clutch, completely rebuilt from front to back with new parts, will spin the fan when not engaged. I like this aspect because it's like having a 2 speed fan.
The last picture compares the two. As I said...apples to elephants.
Operation is simple: immediate on and immediate off. There should be no delays either way because a clutch receives more wear when it is slow to activate. There should be no air leaks when the clutch is engaged.
To test clutch operation, remove the short output line from the shutterstat. Then remove the AN×1/8" NPT fitting from the shutterstat and install it back into the line. Connect shop air to the line @ 120 psi to engage the clutch. The clutch should immediately engage. Pull on the blades with the clutch engaged. The belts should slip before the clutch gives. Remove air and make sure the clutch immediately disengages.
Check for axial play and runout with the clutch engaged. Axial play can be anywhere from 1/8"-1/2" when measured at the blade tips and still be acceptable. By design there will always be axial play with the clutch engaged. There should be no runout when everything is properly torqued. There should be no runout with the clutch disengaged either.
The small Allen cap screws on the thrust plate are a weak point given their recommended torque value compared to size. This was the main focus of Brasco's thread that he linked. They can be felt from the back when installed on the truck. The main retaining nut is behind the front nose cap. If it loosens then it can cause noises too. The fan blade bolts can also create noise if loose, as mentioned above. Ron's warning on the bolts is important. There is a lot of weight sitting in that heavy metal fan.
Everything I have discussed here pertains to the NHC250 clutches only.