Tubes driven by BJT's

[kevinkr]

A 10 ohm current sampling resistor in the cathodes of each output tube is a good idea.

Despite what has been said unless the screens are also operated from a regulated supply, regulating the -60V supply is not such a good idea as every small shift in B+ will result in an attendent shift in operating point. When both are unregulated variations in B+ will be somewhat offset by a similar offset in the bias voltage, while not perfect a lot of the op drift will be cancelled this way.

You can also regulate the B+ in which case you should also regulate the -60V supply.

Incidentally I believe this circuit has almost no headroom beyond what is required to drive the output tubes, and would recommend something like - 70V to -80V supply voltage. The positive supply does not have to be regulated, and with a CCS in the tail you could comfortably get away with about a 12V rail here.

Better still take a look at some of the all tube designs out there on the net. I have heard the AMC hybrid, and at least one home brew of a circuit very similar to this one, and it lacked many of the virtues of good tube sound. The output and power transformers are the biggest % of the cost and would not differ significantly if you choose an all tube design, or even designed one yourself. That pair of transistors could be replaced by a single 6SL7 or even 6SN7 if you wanted something really simple, and don't need lots of gain.

[drago]

For the CCS, do you mean this?

[EC8010]

I'd echo everything kevinkr has said. Regulation could actually make the output stage's DC conditions less stable. A single transistor CCS isn't really up to snuff, a cascode would be better. Even then, there's precious little headroom and I see no virtues in this silicon drive.