Matt what do you think of the above schematic and what is wrong with DC coupling? I have read where it is not always a good thing.
Before I say anything else, let me say that, in general, I am rather conservative in my design philosophy. This translates to usually allowing experience rather than dogma guide my design choices. As a consequence, I categorically reject any claim that any one topology will always sound better than another. I have always believed that regardless of the topology chosen, good design is what makes an amp sound good. Not the dictates of a certain topology or set of components.
That being said, I have nothing against direct coupling per say, but the topology does impose significant restrictions on the design as a whole. By using the plate voltage of the first stage to provide the output stage upper bias point, then, in general, significant voltage drop has to be inserted into the cathode circuit of the output stage to arrive at the desired operating point. Because the output stage is the high current stage, this generally means that significant power is wasted in the output tube cathode circuit. This translates into a bigger power supply with more required ripple reduction and a build that has to contend with significantly more heat.
This design was supposed to be small and use a back to back transformer PS design. At the chosen output stage operating point (which is near ideal for this triode) the quiescent power lost is about 11.8W. So the total power that the supply has to deliver is on the order of 14.5W (11.8+2.1W/0.8). If we were to use the same operating points (Sec 1: [175v, 1.55mA] and Sec 2: [215v, 45.5mA]) in the direct coupled mode, the supply power requirements increase to approximately 400v @ 45.5mA or 18.2W plus the output, so approximately 21W. And this is all for one channel. The total power supply load goes from ~30W to ~42W; a 40% increase.
The impacts to the build, all to eliminate a single capacitor in the signal path, are I believe, not warranted. Especially when the capacitor is chosen such that the dominant zero is placed outside of the audio band (note the 7.8Hz rolloff frequency) where the impact of the transfer phase characteristic in minimized. There may be good reasons to direct couple, especially in instrumentation amplifiers where near DC response is critical, but I generally wouldn't apply it in an audio amplifier without a very specific reason.
As for the battery bias of the small signal stage, I'm just not a fan. I'm a stickler for choosing my bias points and I don't like the forced effects that a fixed potential imposes (whether battery or diode) including quantized bias potential and no grid drift auto adjustment. I've also found both battery and forward biased diode circuits to have significantly higher noise than good quality modern metal film or wire wound resistors. And seeing as how the noise in the cathode circuit of the input stage significantly affects the noise figure for the entire amp, minimizing noise here is of critical importance.
So these are just my thoughts on the matter. I don't intend to offend anyone's design or design philosophy, it's just that after 35 years of design experience (nearly 25 of it as a professional Electrical Engineer), these are the opinions I've formed. The opinions and practices of others are of course their own and not to be discounted out of hand.
I'll get off my soap box now.