Hi Jens,
Fascinating option, however, this is a very different tube to the ECC88/6ES8/ECC189 family. Some quick data on the Russian 6H6p are Ri=1k8, Va=120V Vmax=300V, Ia=28mA μ=22-26, S=11,5mA/V dissipation 4,8W/tube or 8W both sections powered.
The GK1 uses a frame grid 6ES8 tube which is run at 9mA. The 6H6P needs 28 +7-8 mA, so its minimum would be closer to 21mA, more than double. The power supply on the GK1 is high impedance by design (to promote interaction with sonics) and this would lower anode voltage from the design spec of 140V down to around 80V. This could be brought back up, but every resistor in the supply and the bias chain would need swapping out, and the transformer would be pressed pretty hard to supply the additional current.
The GK1 does not use a tube in gain; it's a unity gain situation, a cathode follower. Thus the mu is not the critical factor it would be with a plate loaded output, so this at least is forgiving. Furthermore, the 6H6P is a constant mu tube, where the 6ES8 is variable mu, and this property is exploited in the GK1 to give a very finely controlled H2/H3 spectrum which is very musically pleasing. The Russian tube would have no such spectrum, though it is rash to give categoric statements. My feeling is that even with the necessary supply and biasing mods this tube would not sound as good as the supplied 6ES8, but I'd have to say that tube rolling should do no damage, though the trafo would get warm!!
Cheers,
Hugh
