The 300B ... all of them ... quite happily accept at least 20 volts of positive grid drive. This is not secondhand info gleaned off the Internet, I’ve seen it for myself on a Tektronix scope screen back in the Nineties. I was frankly surprised, because there wasn’t even a trace of a kink or a glitch as it went from negative to positive grid drive. I was expected more drama from the Big Bad Positive Grid Drive, but nothing, no drama, and no signs of grid or plate overheating, either. The 300B is a pretty tough tube, even the bargain-basement Chinese tubes of the day, back in the Nineties. The only reason I stopped at 20 volts is I lost my nerve at that point, and rolled back the gain. I actually have no idea how much power can be pushed into the 300B grid.
That’s when I realized why this amp, or rather the Amity precursor, sounded like a 60-watt tube amp, or a 150-watt transistor amp (I compared it to a Crown Macro Reference and it played just as loud). It just acts like a compressor when things get hot and heavy, and the separate B+ supply for the driver sails right through the heaviest output stage overload. Best of all, the interstage transformer recovers instantly from overload, nothing like RC coupling which requires the cap to re-charge once grid current flows.
An interstage transformer is certainly a benefit for a single-ended amp, thanks to the efficient power transfer and smooth entry into the A2 region, but the benefit is much greater for Class A push-pull. You see, in "normal" or Golden Age amplifiers, only one driver plate is available when the power tube grid goes into A2 and current starts to flow, or if heavy current is needed to overcome Miller capacitance.
The power tube grids never demand A2 at the same time; they take turns. This is important when a push-pull driver, coupled to a balanced interstage, needs to deliver A2 current into a 300B grid. Thanks to summing in the interstage, both sides of the driver circuit are available to push current at every moment, and not only that, because it is a symmetric circuit that always remains in Class A, it is far more linear than a single-ended driver. It really is a small, very linear power amplifier in its own right, so it shrugs off the demands of the DHT grid.
The key principle in a non-feedback amplifier is the lowest possible inherent distortion, coupled with immediate recovery from overload. A balanced, low-distortion driver that uses a well-balanced interstage transformer offers the greatest voltage swing, greatest linearity, and greatest immunity to a reactive load. And the requirements aren’t trivial: one grid needs to swing up 100 volts while the other grid swings down 100 volts, and this needs to happen at 30 kHz with no slewing, transformer saturation, or power supply sag.
This is far beyond the drive requirements of any pentode (35 volts typical), and double anything seen in a SE amp, unless it uses an 845 transmitter tube. (If you’re wondering why so many 300B amps sound mushy and dull, there’s your answer. It isn’t the 300B. It’s the driver falling short.)
That’s also why I discarded the SE driver -> PP output topology. I built an amplifier that used this approach, but the demands on the interstage transformer were too severe, and it never had the clarity or the sense of unlimited headroom of the balanced driver approach. That’s also when I switched from 5687/7044/7119 driver tubes to push-pull 45’s ... actual power tubes with 25~32 mA of current going through each one. Since the 6V6 was specifically designed to replace the 45 (they bias up pretty much the same) in an era when feedback was not universal, the performance of the 6V6 had to be good enough to replace the 45 in hundreds of thousands of radios in the mid-Thirties. The guitarists adopted it and the rest is history.
