As mentioned above, the Mark I Karna used a quartet of 45 tubes, but new production is limited at this time, with future availability an open question. Five years from now? Maybe, maybe not. Ten years? Maybe, maybe not. I can only go by popularity in the market and how many different vendors, in different parts of the world, are currently making them. Not just one factory in one country.
For a DIY hobby amp, availability of tubes not a real concern. But if people are buying a commercial product and expect service (warrantees are a legal requirement in the USA and Canada), availability of new-production matched pairs is a major concern. Don and I looked at the market and we both decided, nope, not yet, still a pretty specialized niche product, and NOS and old-stock supplies have all disappeared into the collector market. When the Mark I Karna was built, they were scarce, but could still be found. Now they are very scarce indeed.
As for sonics, for anyone that’s designed an amplifier from scratch, as I have, the sonics of the 300B are largely determined by the linearity and current delivery of the driver stage. Most of the commercial 300B amps I see at hifi shows have driver stages that are under-designed by a factor of 2 or 3, which results in not enough current to drive the Miller capacitance of the 300B grid. They aren’t as hard to drive as an 845, which is nearly impossible, but they are only second to an 845. A good driver needs high current, high linearity, high speed, and enough headroom to drive the grid 20 volts positive and then recover in microseconds.
Most drivers can’t do that, so the driver and output both clip at the same time, and then both have different recovery times, which can last an appreciable fraction of a second for RC-coupled amps. A typical SET will have RC isolation for the B+ supply between stages, and that RC filter has a characteristic recovery time once a stage saturates. So four things have to recover after saturation: the RC coupling between stages, the RC isolation B+ filter between stages, and cathode bypass caps for each of the cathodes. Typically, they all have different recovery times. Then we get into the subtler issues of capacitor coloration, which can overshadow tube coloration if the caps are not well-chosen.
Most of the "sound" of a conventional SET amp is simply a driver that has run out of linearity. The same is true of preamps, by the way. They run out of linear current to drive the interconnect cable capacitance and then the power supply folds down as it is saturated.
