One additional point to remember in determining actual plate dissipation when it comes to the designs that power the small signal tube heaters from output stage quiescent current: First, subtract out the cathode voltage.
Larry gave an excellent run through on determining real plate dissipation, by also accounting for the screen current that flows through a pentode tube as well. But as a basic point of theory, plate dissipation -- being the product of plate voltage X plate current -- is more specifically the voltage that exists
between the plate and cathode X plate current.
Applying this to the worst case example that Larry laid out in your unit then (V4), and using your approximate cathode voltage of 43 vdc, the new exercise becomes (440 - 43) X .0466 or 397 X .0466 = 18.50 watts.
This shows that this particular tube is actually operating within its designated limits -- if only barely. Plate dissipation will be reduced somewhat when the quiescent current is reduced -- but remember, as quiescent current is reduced, plate voltage tends to increase, causing the two actions to somewhat counteract each other. Plate dissipation does decrease, but not as fast as a simple reduction in current alone would imply.
These particular designs of Fisher were really threading a needle so to speak, in that the output tubes need to pass enough quiescent current so as to be able to light the heaters of the preamps tubes effectively, but not pass so much current as to cause the plate dissipation rating of the output tubes to be exceeded. The unit also had to be able to compete in the market place, so maximum practical B+ voltage was also required to produce a competitive amount of power output from the scheme as well. All of these requirements then work to make the window of acceptable quiescent current rather small in these amplifiers. A little too much current and the output tubes cook. A little less, and you can pack a lunch waiting on the small signal tubes to warm up. So all of these designs operate the output tubes at the high end of what would normally be considered as safe operation. But there's one other thing as well.
Because the tubes operate so much closer to their limits, the importance of using a
well matched quad of tubes in these types of designs cannot be overstated. Using the best case example (V5) in the OP's unit, actual plate dissipation becomes 15.60 watts which is quite reasonable, but represents over a 15% difference in dissipation that V4 is incurring over V5. In other words, V4 carrying more than its share of the work load (lighting the heaters), while V5 is comparatively loafing. Besides the degradation in amplifier performance this causes, it also causes the tubes to wear unevenly, which only is accentuated with use as time goes on.
My intent is hardly to trash the design, as for the price point of these amplifiers, the scheme produces the huge benefit of operating the small signal tube heaters off of DC voltage, which is one reason why Fisher equipment was considered to have such a low noise floor back in the day (and still does), where so many other manufacturer's pieces did not. However, in today's vacuum tube audio environment of higher AC line voltages and modern manufactured tubes, these particular types of designs require close monitoring and the types of modifications as already done here to ensure proper operation, and the long term health of all the components involved. There are modifications that can be done to completely circumvent these problems, which also produce a significant improvement in amplifier performance as well -- but they are extensive, and only for the advanced hobbiest to consider. An example of one such effort can be found here, offered primarily to reinforce the level of modification required, but also to show what the performance potential of the amplifier truly is:
http://www.audiokarma.org/forums/index.php?threads/improving-the-fisher-x-101c.582379/
Good luck with your X-101B!
Dave