my $0.02 for a WE300B
I started working on a 300B SET design over a year ago (not done yet) but early testing with a pair of WE300B has forced me to change the design for both the driver circuit and the filament supply.
A real Western Electric 300B has a center-tapped filament, meaning you have two separate filaments inside the grid structure which are lined up as left and right. They are tied together in the middle and brought to pin 4. The outer ends of the two filaments are also tied together and brought to pin 1. So you effectively have mirrored-phasing between the two filaments. This is what prevents using an AC filament supply, as it's impossible to null the hum level to an acceptable limit.
Based on the actual structure, a DC supply is mandatory for a quiet output stage. It also makes sense to fix the DC bias point to one of the tube pins, not floating between them (when using an adjustable null pot). If you can get a perfectly quiet supply you can also fix the AC return path, otherwise having a null adjustment pot could improve the hum level. Of course if you're not using WE tubes, whatever you are using should be examined to see if the filament structure is the same as WE's or not.
As for an acceptable hum specification, I spec everything to a 1-watt standard. For amplifier output hum/noise, it has to meet a minimum of -80dB below 1-watt. For an 8-ohm output, it's 283 micro-volts and for a 16-ohm output it's 400 micro-volts, or 0.000283v and 0.0004v.
Regards, KM
I started working on a 300B SET design over a year ago (not done yet) but early testing with a pair of WE300B has forced me to change the design for both the driver circuit and the filament supply.
A real Western Electric 300B has a center-tapped filament, meaning you have two separate filaments inside the grid structure which are lined up as left and right. They are tied together in the middle and brought to pin 4. The outer ends of the two filaments are also tied together and brought to pin 1. So you effectively have mirrored-phasing between the two filaments. This is what prevents using an AC filament supply, as it's impossible to null the hum level to an acceptable limit.
Based on the actual structure, a DC supply is mandatory for a quiet output stage. It also makes sense to fix the DC bias point to one of the tube pins, not floating between them (when using an adjustable null pot). If you can get a perfectly quiet supply you can also fix the AC return path, otherwise having a null adjustment pot could improve the hum level. Of course if you're not using WE tubes, whatever you are using should be examined to see if the filament structure is the same as WE's or not.
As for an acceptable hum specification, I spec everything to a 1-watt standard. For amplifier output hum/noise, it has to meet a minimum of -80dB below 1-watt. For an 8-ohm output, it's 283 micro-volts and for a 16-ohm output it's 400 micro-volts, or 0.000283v and 0.0004v.
Regards, KM