The MX-110 uses a combination full wave bridge rectifier in conjunction with a conventional full wave rectifier circuit. The bridge rectifier produces twice the voltage output that the conventional full wave rectifier does, and powers the audio circuits and local oscillator. The conventional full wave rectifier power the remainder of the tuner circuits, the MPX circuits, and provides the positive heater bias for the DC powered heaters.
Within the four diodes making up the bridge rectifier, D401 and D402 strictly serve the positive output of the bridge rectifier. But D403 and D404 make up not only the negative side of the bridge, but also serve as the rectifiers for the conventional full wave rectifier circuit. The conventional full wave rectifier may not be that apparent to see because it is typically configured with the center tap of HV winding grounded, with the positive side taken from two diodes -- one connected to each end of the HV winding (tube or SS). In this case however, the polarity of the two diodes that serve in the conventional full wave circuit are "turned around backwards" from how they would normally appear, because they have to be to provide the negative output of the bridge rectifier circuit. What to do? Mac simply provided a CT to the HV winding that then becomes the positive output of the conventional full wave rectifier circuit, so that the two diodes D403 and D404 are really doing double duty. But here is the point: because they represent the negative output of both of these supplies, it means that any resistance placed in series between these two diodes and ground will lower (proportionately) the output of both the bridge and conventional full wave rectifier circuits simultaneously.
As a result, be sure to check the B+ voltages to the tuner circuit as well. Those circuits draw by far the most current in the unit, so any adjustment should be based on the voltage going to it. If you find that after adjustment that the voltage to the audio circuits is still higher than you would like, then you will need to place an additional resistance between the output of the positive rectifiers, and the first filter cap. Any resistance added there will only affect the voltage supplied to the audio circuits. I suspect that if you have been making the adjustment based on the audio circuit voltages, then the tuner voltages will now possibly be too low.
If you want to go the thermistor route, watch that the AC and DC heater voltages don't drop too low as a thermistor will affect all voltages in the unit. As for the turn on surge voltage, the conversion to silicon rectifiers definitely makes for a higher surge voltage. A properly sized 1A CL device for this unit will usually only curtail that slightly (but only slightly) -- but also make the surge last longer because the heaters are warming up more slowly. As long as the filter caps can handle the surge voltage wise, the surge is really of little issue.
The rectifier design is not anything that Mac developed, but is a configuration that became possible on a practical level due to the development of appropriate SS rectifiers (selenium or silicon). The configuration could be populated with vacuum tube rectifiers as well. But then you would also need the attending heater windings -- three separate heater windings -- with the tubes employed being a convention full wave type tube, and two separate half wave rectifier tubes (think damper tubes). From a production cost standpoint, this would be prohibitive in a product like the MX-110, but becomes very practical if not desirable when non-thermonic rectifiers are used.
Mac used the exact same approach in the original MC275, using a bridge rectifier circuit to provide the B+ to all of the amplifier tubes except the 12AZ7 cathode follower driver tube in each channel. The B+ for that tube was derived from a center tap on the HV winding, in the exact same way that the tuner circuits are powered from the CT on the HV winding in the MX-110.
Dave