Tube Physics Question


At least I think it's Physics. 

Short version:  Is it logical that going to a less powerful power tube in an amp can yield better dynamics?  What's the Physics behind tube design/operation and their ability to reproduce the swing of musical dynamics?  What part does bias play?

Longer version:  A couple of years ago I acquired a used tube amp for a second system.  It came with three sets of power tubes - the original KT88s, a set of Gold Lion KT88s and a set of Tung Sol KT120s.  The amp had been biased for the 120s, and those are what I used and have been happy with.  More power = better, right?

Recently I swapped out the speakers in that system for a more efficient and transparent pair.  While the change was nice and an obvious speaker upgrade, I started hearing a little edginess in the treble.  Since I'd never checked the bias I put a meter on it and found it low (avg ~12mV vs. a recommended 45-50mV for KT120s).  Running the pots up to max only reached ~35mV, but it did improve the sound.  Called the mfr and he said not to worry about a lower bias setting, as it made little difference.

Not content, I swapped out the KT120s for the Gold Lion KT88s, checked/reset bias to 35mV (recommended for 88s) and wow, what an improvement they made.  Without getting into a bunch of hyperbole, what was immediately and most noticeable was the increase in dynamics.  So my question is, what about the design/production and implementation of a tube affects its ability to reproduce dynamics?  Or is it more likely that running a lower power tube at relatively higher bias causes the effect?
tmcclintock
Kevin Hayes of VAC has also preferred the KT88s to the KT120s.   I believe his thoughts on the tubes, which include a technical discussion, are in a blog on his website. 
While my manufacture, Bob Carver, feels the KT 120s are a sonic improvement in his VTA180 design I've returned the KT 88s back into service while my 120s are collecting dust. 

There is definitely an audible difference between the tubes but nothing like what you've described. To me the much older 88s simply sound right to my taste. Possibly another amplifier design might showcase the 120s.

At some point I may consider the upgrade needed to use the KT 150. Good luck with it.
The question from the op is really interesting; can anyone address it? Does the KT120 compress the signal, while the KT88 does not?
Thanks for the link, rcprince.  A lot of what Kevin Hayes says went over my head, but a couple of his points maybe speak to what I heard.

"3) You should expect more distortion with the KT120, possibly particularly at moderately low listening levels and with highly reactive loudspeakers"

"6) You’re likely to give up a few kHz of high frequency response"

And one of his main points is as mlsstl suggested earlier, an amp designed for KT88s will not sound as good with tubes having a different spec.

I'd still like to hear from anyone who knows what differs mechanically in tubes.  Like what is physically different between KT88s and KT120s - bigger plates? - thicker metal? - different metalurgy?