With 325VDC B+, and 390-400 ohm, 2W cathode resistor, the 6S4As will idle at around 25 mA, and this is the tube's "sweet spot".
Cathode value should be 9.8VDC, +/- 0.2V, with NOS tubes, old beat up used ones will be "all over the place"
Figure on about 3X the tubes' internal P to K resistance, (3.6K) for the OPT primary impedance value, that is, 11-12K for a single 6S4A.
Always - use a grid stopper resistor, 1K, is good, foor each of the 6S4As.
I like to parallel two of the 6S4As, and use 5 to 6.6K for SE operation.
I've not tried more than two in parallel, no valid technical reasons not to try 3 or 4 of them in PSE
I prefer to use an SRPP driver stage, I use 6CG7s or 5963s with 1K RA/RK , and use at least 0.22 uF coupling cap for parallel operation, the 6S4As' internal capacitance starts getting "up there" when the numbers go above 2 in parallel. The grid resistor value may have to be reduced with 3 or 4 in parallel, try 150K or even 100K, that eliminates any vexingand annoying instability issues.
I look atthe 6S4A as being the apocraphyl "6BQ5, without the screen and suppressor grids" HaslowerPD thana triode strapped 6BQ5, 8.5 W vs 12, but I'd pushed them harder, just for the
halibut , but there is no valid reason to do so...
HTH
/ed B