Finished my G14-S - the selectable-tube-or-solid-state version of our good old G14 - added only a extra step marked "S" to the bypass switch..
Has taken me forever though: I thought it would be relatively simple, so I announced working on it more than a year and a half ago. Which instantly and to no surprise dropped sales of the "old" G14. Had serious and unexpected trouble getting it behaving predictably, for the longest time I simply didn't get why - extremely non-intuitive err pattern, nothing like I've seen before (and I've been around some). I just work my way through this sort of trouble, knowing that I will eventually crack it. But it didn't resolve, even after a year of regularly returning to it. At some time I set aside one day per week to that project - still was not right ever, or even moving in the right direction.
While at NAMM in january, I and my US distributor Frank "OneF" went to a dinner with Ruari, Dave Collins and Ron ‘Spider’ Entwistle.. Across the first fifteen minutes of pre-dinner drinks I realized that this was probably some of the sharpest company I've ever found myself in - and I've never, ever, been in a situation concentrating this much hands-on-knowledge on the topics I find most interesting.
So somewhere underway I mentioned the ongoing trouble I had with G14S.. explained the basic circuit, the troubling symptoms, what I tried already.. Then follows a quick negotiation around the trouble, and out comes the verdict: "..that tube, the 6922, wasn't that made for UHF television tuner front-ends?" - "..you probably have oscillations going on at some very-high frequency.."
Way too simple, I already checked for that, repeatedly, off course.. ..but my equipment won't go over say ½GHz tops, I never, ever want to do that sort of signals or work..
Then I had a chance of borrowing a GHz analyzer setup - and to my horror; sure enough; in not-tube-mode I saw clear oscillation building just under a G, spreading to supplies, messing up ground plane currents and whatnot
They knew that all the time, didn't they? The reference to tv-tuner front end and all :-[
..from that point it has more or less just been manual work to try out different types of tubes and silicon to find combinations that we perceive to be suitable. The only slightly tricky thing was to find a solid-state amplification strategy that mixed well with our traditional tube circuit - we didn't want the unit to change character (or the feel of the actual equalizing) too much when flipping between tube and solid-state topologies..
..that dinner with Ruari and friends - that's something I won't forget..
/Jakob E.