I too favor a little better preamp than the the seduction. For me
a little better, is a tube design say using 12AY7's ( gain of about 33 per
stage) and then a driver tube other than the one they use.
This is 3 triode stages. Zener diode regulation[1]
and no CCS's.[2] PS. I am cheap too, I use 10 cent green leds for bias but...[3]
...
[1] Or tube regulation. Transistors require a heat sink that is a pain
to isolate from +300 volts.
[2] You want a load line 90 degrees from the bias point of your tube.
Horizontal is not always the best.
[3] Spend my money on big 100 uf poly-caps
Agreed that one can do better than Seduction, which was always intended as an entry-level design.
The 6DJ8/6922 has a low enough plate impedance to drive a moderate-length cable directly so that two stages can suffice. Most other tubes will as you say require an additional stage. For the Progressive Engineering phono preamp I did just what you have suggested - switched to 6N1P for the gain stages, and added a direct-coupled cathode follower for low output impedance, along with a shunt regulated power supply. In our experience, cathode followers are pretty decent if they have a good shunt-regulated power supply; otherwise they have been problematic. A 12AY7 would be a good candidate; like the EF86 it's a tube designed for small-signal audio.
An improved Seduction would have a shunt regulated power supply. We've done it experimentally and the difference is quite audible. It would require a new power transformer and supply of course. In fact we had planned to make such a thing, but it got replaced by the EF-86 based design after we listened to the Repro Amps.
Incidentally, we did build a test box with stereo cathode followers and six different shunt regulator circuits, including Zeners and gas regulator tubes, in order to chose what configuration to use in the Bottlehead products.
A couple points of information about the designs: In addition to lower distortion at large signal magnitudes (admittedly, not that significant for these tiny signals) a good current source plate load provides some 50dB of power supply noise rejection, so I continue to prefer that configuration. We use old-technology low efficiency red LEDs because their dynamic resistance is especially low, and they are very quiet.