Part of this hum arises from the fact that the grounding scheme on the Mark III is cruder than the star grounding scheme on the ST70. I ended up moving the power ground and the filament center tap to the driver board ground and this helped a lot. In addition, Roy supplies the board with the triodes in the input 12AT7 paralleled. When I removed the traces that tied the two triodes together, and used a single triode, the hum dropped to an inaudible level. I had also increased the feedback a bit, which drops the gain and noise. After all this, the gain is still plenty high.
The 1 watt square waves at 1 khz and 10 kHz are near perfect. I get ~55 watts of output power between 30 Hz and 20 khz at less than 1% THD, dropping to 28 watts at 20 Hz (power line at 117V AC). The frequency response is flat to +/-0.2 db between 20 Hz and 20 kHz.
The amp sounds excellent, as good as the best I've built. The tested performance and sound is a notch above what I got with the Triode driver board on a previous Mark III, although some of this might be due to the separate bias controls on the VTA board. Proper balance has a big effect on distortion.
The VTA driver circuit has a minimum of parts and complexity, but it has the essentials including a two triode phase splitter and individual bias. Definitely recommended for someone willing to fuss a bit. It may change your mind about the Mark III.