I don't believe your bias cap being a 100uf is causing what you're seeing as a lack of good adjustment range. 100uf takes a little longer to charge but should be more effective to filtering any ripple or noise in the bias circuit. On a twin this longer charge time shouldn't be an issue in my experience.
It's always a good practice to check your 220k resistors for accuracy, though I don't think that's what you're experiencing.
Since you're using the transformer shunt method the problem is you don't actually know what your mA draw is per tube, only across each side of the transformer. Keep in mind this will be showing you current for 2 tubes at a time, as opposed to the same method on a 2 output tube amplifier like a bassman.
Have you tested the tubes in a good tester? If you can read the plate current (octal bias probe is good here) on each tube the amplifier can serve as somewhat of a good tube tester as it will run them at operational voltage. You can see if one is drawing more or less current than the others this way.
I suggest the transformer shunt method only if it's a known good matched set of tubes. The transformer shunt method lacks certain data that can clue you in on if you have an amp problem or a tube problem.