Longer springs on the F. By means of converse proof, by shortening the spring, Ripster made one of the keys on his keyboard stiffer. My guess is that because there is more spring to buckle, and more spring to buckle back, this leads to the lower actuation point, and greater bounceback.
Besides spring length, has anyone compared other attributes of the Model F springs vs. the Model M springs. For example, are they made from the same kind of steel, the number of coils per inch, the gauge (thickness) of the wire, etc.
I removed all the keycaps from one of my PC AT keyboards for cleaning and while the barrels and springs were exposed, the F7 spring got snagged on the towel I was using and was subsequently pulled hard enough to deform the spring. I was not able to "recompress" the spring back into its orginal tightly coiled form.
I sent the keyboard to Unicomp for repair. They replaced the spring with a spring I presume they use in their current boards. The repaired F7 key once again buckles, but both the sound and feel are different from the all the other keys. The click is not as pronounced and it is a tad bit softer (quieter). It almost, but not exactly, like a Model M key!
This would support the theory that the spring is the most important, but not the only component that differentiates the Model F from the Model M.