Some newer power amp designs found here, by AK members, and the commercial high-end accept anything from the venerable 6L6 to KT-150s. A few very expensive offerings continue using KT-88s rather than the newer and more powerful "replacements".
Besides cost, is sound preference what usually drives the choice of output tube? Availability of the usual suspects doesn't seem to be an issue, unless the amp is one of more exotic design using less popular tubes.
I've read high praise for all the common types and heard good sounding amps of most every ilk. Clue me in, please.
Besides cost, is sound preference what usually drives the choice of output tube? Availability of the usual suspects doesn't seem to be an issue, unless the amp is one of more exotic design using less popular tubes.
I've read high praise for all the common types and heard good sounding amps of most every ilk. Clue me in, please.
For production designs, the main factor is availability. For one off designs, the latitude becomes much wider. I tend to look to types such as TV horizontal and vertical deflection finals. These don't come with the audiophool "tax", as these types never have any audio usage mentioned in the spec sheets. The horizontal finals aren't useful as single ended finals as the most linear parts of the plate characteristic are well within red plate territory. They do make for good sounding Class AB1, PP finals. (6BQ6GAs and 6BQ6GTBs both are essentially more powerful 6V6s) Of course, they're not making these anymore these days, and a lot are becoming Unobtainium -- types like the 6LW6. (The one saving grace is that a lot of these TV HD finals were made with a variety of heater voltages as it was a common practice to daisy chain the heaters to run directly across the AC mains with PTX-less DC supplies. The spec sheets often include brags about how much current they can pull with plate voltages of ~150V -- useful to kick start the horizontal deflection circuit when you're not including a PTX.)
Other choices that avoid the "tax" are the RF finals, both pentodes and power triodes. As for sound quality, you can always include local NFB to tame those types (6L6-oids) that like to make too much h5 and nasty sounding, higher order, harmonics. Local NFB is also useful for making high-u "zero bias" RF power triodes look more like low-u audio power triodes. Types like the 6V6-oids and the 6BQ6-oids make mainly h3, but not much of the higher order harmonics.
Keep "jack of all trades and master of none" in mind. O/P "iron" has to be optimized to a specific type, in order to obtain max. performance. Because of a common basing configuration, multiple O/P types may "work". However, only 1 power type will yield max. performance and performance does not equate to power.
Two good points, Eli. The tube rolling capability of some commercial amps may be more of a sales attraction geared towards the average consumer. Still, a "mismatch" may actually sound better to some listeners and with some speaker/amp combos.
My other wonderings pertained to using KT-88s vs. KT-120s. The 120s were supposed to be a superior sounding and more powerful replacement for the 1954 upstart. Some expensive modern amps are furnished with and, I assume, designed for the 88s. If the 120 claims are mostly correct, plus they're plentiful and competitively priced, what might be some reasons to stick with Grampa's Tube?
My other wonderings pertained to using KT-88s vs. KT-120s. The 120s were supposed to be a superior sounding and more powerful replacement for the 1954 upstart. Some expensive modern amps are furnished with and, I assume, designed for the 88s. If the 120 claims are mostly correct, plus they're plentiful and competitively priced, what might be some reasons to stick with Grampa's Tube?
JMO, New Sensor would have done better to work up close copies of early production EI KT90 type 3s. I've seen the KT120 data sheet and it's distressingly non-linear. 
You have to bludgeon the KT120 with NFB to bring it up to HIFI distortion standards. I regard the KT120 as a guitar amp type, in which distortion is actually desirable.
New Sensor does make KT90s, but they are no match for EI's better efforts. AAMOF, New Sensor's 1st KT90 iteration was pretty darned bad. Their product has since improved, but still has quite a way to go.
The current "Gold Standard" in big beam power tetrodes is New Sensor's "reissue" Gold Lion KT88. The variant exhibits excellent sonics and the only possible issue lies with durability. Talk to Jim McShane about the "reissue" GL's track record.
For those whose bank balance simply can't stand the cost of "reissue" GLs, New Sensor's ElectroHarmonix (EH) KT88 gets the nod. It is a tad shy in the bass extension dept., but is otherwise a very solid performer. The EH variant costs about 1/2 as much as its GL "stable mate".
You have to bludgeon the KT120 with NFB to bring it up to HIFI distortion standards. I regard the KT120 as a guitar amp type, in which distortion is actually desirable.New Sensor does make KT90s, but they are no match for EI's better efforts. AAMOF, New Sensor's 1st KT90 iteration was pretty darned bad. Their product has since improved, but still has quite a way to go.
The current "Gold Standard" in big beam power tetrodes is New Sensor's "reissue" Gold Lion KT88. The variant exhibits excellent sonics and the only possible issue lies with durability. Talk to Jim McShane about the "reissue" GL's track record.
For those whose bank balance simply can't stand the cost of "reissue" GLs, New Sensor's ElectroHarmonix (EH) KT88 gets the nod. It is a tad shy in the bass extension dept., but is otherwise a very solid performer. The EH variant costs about 1/2 as much as its GL "stable mate".
The nonlinear characteristics of the larger tubes make them more suitable for class AB - the increase in transconductance at higher current fills in for the opposite tube which has stopped conducting. Look at the plate curves - more current change for a given grid voltage delta at higher currents. Depending on operating point, there generally will be a null in distortion at some higher power in addition to a low distortion region at low power with 180 degree conduction. Of course negative feedback reduces distortion further.
VAC likes those GLs too, and uses eight of them in their new $150k 452iQ integrated. That beast is two inches taller than my 46" Paradigm towers and almost twice my weight. Crazy. Just imagine the size reduction and shed pounds going the SS route. Although it wouldn't look as pretty and maybe wouldn't sound as beautiful. Glow tubes!