[John Caswell]

Hi Phil et al,
Well both are built, and after a bit of minor twincing frequency response is flat ±1/2dB from 3Hz out to about 80KHz. There is a bit of a rise from 80kHz up to 100kHz but as it is so far outside the "normal" bandwidth, I haven't bothered to look at it yet.
Residual hum and noise -70dB unweighted and -85 "A" weighted wrt to 1watt (2.83V//8R) ie can't hear it, which in absolute terms is well below 1mV. Sensitivity 575mv for 50 Watts max at clip. Distortion at 1Khz = 0.013% and about 0.15% at 20kHz
Both use ECC83, ECC82, 2X KT120, all ac heating of valves but heater windings are centre tapped, however there are slight variations on a theme.
Amp1 has 2X (cheap) 6d22S as rectifiers amp2 has a (b****y expensive) GZ37, but it looks like options for rectifiers are going to be 5U4,5R4GY, 5T4, GZ37, 2x 6D22S, 2x EY84 plus of course SS (not sure about that).
Obviously using SS and Directly heated rectifiers the capacitors are a bit of a problem and would need at least 800V ratings, so 2 X 450V in series is the answer here.
I am using standard Hammond 1650RA OPTxs with a variation of available power tx and chokes to give 500V HT.
The 6D22s heaters are run off a small 30VA toroid but would possibly run off the 5V winding according to Eric Barbour. Haven't tried that yet!
The basic amp cct is identical to the GEC cct with minor changes to C7/8 to flatten lower LF hump.
I cannot see any reason why you could not use a 6SL7 for the ECC83 and a 6SN7 for the ECC82.
I am pretty sure I am going to build these in anger for myself as monoblocks in which case I would probably punch the chassis for Octal/B9G valves (6SN7/6SL7/6D22S/EY500) and make some small sub plates to fit the B9A types (ECC82/ECC83/EY84).
Each amp power requirements are 500V/180mA, 6.3V/4A and 5V/3A.
I am passing the amps over to Matthew Snell to listen to, but frankly for a design first shown in Wireless World April 1957 I am seriously impressed.

John