... circuits are designed according to the goals we wish to accomplish and from that, the limits of the tubes selected.
Given suitable heater current, bias and sockets, 6L6 will work just fine in the circuit ... they would put out a bit less power ...
... Maybe 6V6 is a good starting point if I can use the same power transformer.
EL84 draws 760mA of heater current per tube;
6L6 draws 900mA of heater current per tube. As long as your power transformer can supply an extra 280mA on the 6.3v winding, you can use 6L6s in your amp (obviously with octal sockets).
What pdf64 is saying is if you make no other changes (aside from cathode resistor value, to obtain the correct bias), you will get "EL84 performance" from your pair of 6L6s.
We
hope the B+ doesn't sag too much from extra idle current, though nothing says you actually have to idle the 6L6s at a current higher than you had with EL84s. That is, you don't need to idle the 6L6s at 100% of their dissipation rating (and therefore higher plate current than the EL84s). Also, the existing OT primary impedance will tend to prevent the 6L6s from drawing much more
dynamic plate current than the EL84s did.
A drawback will be the larger bias voltage needed for the 6L6s, which also implies they will need a larger drive signal for same-power/same-distortion. Without really investigating, I would assume the long-tail inverter has a high enough B+ voltage to be able to deliver the larger output (though that assumption should be tested).
________________________________
If you want "6L6 performance" from the amp, you're really looking to get a larger power transformer, larger output transformer (with a lower primary impedance to allow higher
dynamic plate current), and possibly some tweaking to your B+ voltage and phase inverter (including the amount of feedback).
These changes
could lead you to want some other preamp changes as well, but that's best figured out after the power supply & output sections are in their final form.