The buffer is a hybrid cascode with IRF820 MOSFET's and a section of 5687 running at 15 mA. This faux pentode has some seriously flat "plate" lines, so I went with a single stage of B+ filtering.

Output is through a 1:1 80% Nickel TX by Intact Audio. .004" gap for 15Hy of primary L.

I ran a $2.99 RadioShack Alps pot for the volume control. It sounds quite good. The low gain removes the principle complaint with those cheap pots: tracking mismatch.

Compared to the previous active design( with gain ), this one is a bit cleaner. Slightly less euphonic perhaps. The active loaded 12B4 is not very colourful tube, and that's part of the reason I like it so much. The buffer circuit is a step cleaner. How much from the cascode buffer circuit, how much from the high perm, bifilar output TX's?

I have got to say I am pleased with it. If not for the Nickel outputs, this would be quite inexpensive. No DC filament requirement, single stage L-C filter for B+. It is quiet...on AC no less. That one was a very pleasant suprise. The 6754 allows a full warm up of the 5687 before bringing on the B+ slowly enough to make a damper diode jealous.
cheers,
Douglas