Thanks, Tone!
Don't hold your breath for a kit, though I might offer built units sometime in the future. And according to what I have heard from Doc Bottlehead, he will only offer completed units instead of kits. The parts cost for my unit was about $1400, so it wouldn't exactly be a cheap kit anyway.
You could always do it yourself if you wanted, though. There is enough information on my
headphone amp site to build it, if you know how to build tube components already. You would need to go to Gary Pimm for the current sources. I left out the things that an experienced builder would already know how to do, because I don't want to lead an inexperienced person to electrocute him/herself.
In my opinion, this triode design has great advantages over a pentode amp. Triodes inherently have the lowest distortion of any amplifying device. Other devices, whether tube or solid state, must have their distortion brought under control with negative feedback, which improves measured performance but is sonically detrimental. (Of course, a circuit that requires feedback will sound better with it than without it). The autotransformer steps the already-low output impedance of the triode down to even lower values. The resulting circuit has gobs of current capability, abundant linear voltage swing, a fast slew rate, and runs dead-quiet. The switchable output impedances allow matching-to-taste with any headphone.
The pentodes in the Espressivo are the active element in the current sources (one per channel). The current source feeds electrons to the plate of the triode at a constant rate, regardless of the signal swing. In the old days, this would be done (imperfectly) with a choke. The current source functions as a choke of nearly infinite impedance, allowing virtually no supply noise into the audio circuit, and preventing the audio signal from modulating the supply. Gary Pimm's extremely advanced current source circuit is what makes this approach especially successful.
I suppose the current souce weakens the "low parts count" argument, but the signal path still just consists of a few parts.
Czilla9000: Yes, tube amps can be built with fewer parts, though of course some tube designers take a complex approach. I prefer to keep the circuit path simple and direct. I believe in choosing the most linear parts, and trying to use them in a way that will allow them to perform to their best capability.
One way to compare tube and solid-state designs (in terms of parts count) would be to look at discrete SS designs instead of those using chips. I think the Gilmore amp is a discrete design (and a great-sounding one, from all reports!), so it might make a good example of my statement about relative parts count.