i DO like the sound of a UL amp--i like the harmonics it generates,
and i like the way it's halfway between triode and pentode
operation. it's got good output power, almost as good as pentode
ops for a given voltage, and low enough output Z so that you can
run it open loop.

dan, i was inside a fender 140 the other day, and it seemed pretty
open to modifications. i would remove the NFB and add some low
Z grid drivers, be they tubes or SS. once you get to that point the
amp just seems to keep going and going like the goddamned
energizer bunny... there's always a little bit more balls you can
squeeze out of the thing by slapping the strings harder. i like that
illusion of infinite headroom.

even if you just turn it into a clean machine, you can always use it as
both a bass amp and a guitar power amp, simply plugging in a high
gain preamp/stomp box if you so desire. 

the great thing about using some kind of follower grid drive is that
the actual phase splitter can be whatever you want, and you don't
have to pay much attention to how well it can drive things. for
example, you can use a concertina to great effect here (just keep in
mind its limited output swing), or a plain old ax7 as a differential
stage (i.e. schmitt). the grids/gates of the followers present a very
easy load for the voltage amplification/phase splitting stage.

also, i have never found a bias circuit that was as rock solid steady
as a direct coupled follower. you set it once, and it stays there. the
output tubes like this kind of treatment.

as far as your suggestions, i would say first try opening the loop.
hear what it sounds like. particularly at lower than cranked signals
the sound will not change if you go to follower grid drive, so it will
give you some idea of the sound of a UL output stage with a
minimum of effort.

after that, if you determine you like it, considering donating a tube
socket for a duo-triode cathode follower, or perhaps just finding
some space for two mosfet source followers. you'll need a negative
voltage rail, the more negative the better. if the PT is a center
tapped full wave then you can easily add a couple of reversed
diodes and generate a negative rail. if not you may need to add
another tranny, which is certain to work well. the negative rail
should be of decently low impedance itself since a few ma's of
current WILL flow through it. you might have trouble using a
capacitive divider off the main B+ winding, for example. 

kg