What's wrong with a cathode follower? Use it and you don't need to hunt for tubes with low plate resistance in the front end. Too much gain? Use a step down input transformer and you can flip the polarity to make the output in phase with the input signal. And, of course, you also gain galvanic isolation.
6T4 is an example of all the US made tubes that pretty much stayed in the USA. Nothing used I can see in the UK. For us Europeans who don't want to be dependent on importing from the USA it would be useful to have more widely distributed choices. 6V6 for instance. The 6AH4 is probably better but again I had to import mine from the USA. The 6W6 is over here in tiny numbers, 6Y6 more common, A2293 is available. Choices here are quite different from the USA.
Ale Moglia published a good table of triode-strapped Russian pentodes, with measurements, including ra
https://www.bartola.co.uk/valves/2013/02/23/russian-pentodes-in-triode-mode/
https://www.bartola.co.uk/valves/2013/02/23/russian-pentodes-in-triode-mode/
Don't put the type number in print for the robots to find. Just read the quick reference data in the upper right, then look up the prices. While Stan (ESRC) was still alive he sold me a box of 100 new tubes for $75. Nobody seems to want them. Running in push pull UNSET mode I get about 30 watts from a pair into a 3300 ohm OPT. Don't remember exactly how far I turned up the power supply to get there though.
The octal output tubes from many old AA5 radios were also odd heater variants of the 6W6. The 12L6, 25L6. 25W6, and 50L6 are all 6W6s with odd hearer voltages. The 35L6 is not the same. It has the same guts but uses a lower voltage heater to support the two extra tubes needed for AM-FM radios. Attempts to squeeze more than about 25 watts per pair results in red plate meltdown.
The problem with a triode line stage is gain.
If you start with a phono stage or digital source that puts out 1V peak, and a power amp that reaches full power output with 1V peak input, you actually don't need any gain at all, or at most 2x or 3x gain. To the best of my knowledge, there is no triode that has a mu of 3. Maybe 4? But usually 10 to 20. So you put a triode with mu of 8 in between the source and the power amp and you need a voltage divider or transformer to knock down the signal voltage, which brings in noise, frequency losses, etc. And around and around we go.
I had built a 2-stage push-pull 2A3 amp that needed 3.5V peak input signal to reach full power. I put a twin-triode with mu = 16 in front as a sort of 'single ended line stage'. The volume control went at the input to the SE line stage (right after the source and selector switch). That worked well. Very low noise, inaudible hum, even with AC heaters for the 2A3s.
Another way to do it would be to take a twin triode with mu of 30 or 60 and make it an 'anode follower' (shunt feedback, plate to grid) and set the gain where you need it to be. Put a MOSFET source follower on its output to unload the NFB loop. Super low THD, low Zout, ticks all the boxes. But yeah, it uses NFB.
Good project for cheap, neglected tube types.
If you start with a phono stage or digital source that puts out 1V peak, and a power amp that reaches full power output with 1V peak input, you actually don't need any gain at all, or at most 2x or 3x gain. To the best of my knowledge, there is no triode that has a mu of 3. Maybe 4? But usually 10 to 20. So you put a triode with mu of 8 in between the source and the power amp and you need a voltage divider or transformer to knock down the signal voltage, which brings in noise, frequency losses, etc. And around and around we go.
I had built a 2-stage push-pull 2A3 amp that needed 3.5V peak input signal to reach full power. I put a twin-triode with mu = 16 in front as a sort of 'single ended line stage'. The volume control went at the input to the SE line stage (right after the source and selector switch). That worked well. Very low noise, inaudible hum, even with AC heaters for the 2A3s.
Another way to do it would be to take a twin triode with mu of 30 or 60 and make it an 'anode follower' (shunt feedback, plate to grid) and set the gain where you need it to be. Put a MOSFET source follower on its output to unload the NFB loop. Super low THD, low Zout, ticks all the boxes. But yeah, it uses NFB.
Good project for cheap, neglected tube types.
The big ugly 6AS7 has a Mu of 2. With a pair of 13 watt plates it's a bit of overkill for a line stage. The smaller 7233 has a Mu of 4 and works quite well, but it was designed for regulator service in expensive test equipment, so It's hard to find and expensive now. I haven't checked recently, but I paid $10 each about $15 years ago. As mentioned, the big half of the twin dissimilar vertical sweep triodes has a lowish Mu. The 6CY7 has a Mu of 5, and the 6DR7, 6FY7 and 6EW7 are 6.
Yes exactly - you don't need gain in the preamp, but people are still using preamps and they like using tube preamps to get the "tube sound". Reality or myth, there's still a demand for tube preamps. Commercial amps are designed around a typical input signal of 2V, but even then have extra gain as a rule because they're 3 stage PP designs with 12A*7 input tubes. So though a passive preamp is enough for most systems, the demand is still there for a tube preamp without much gain. Like you say, a gain of 3 or 4 would be good and all those dissimilar triodes are around 5 or 6, and so is the 6W6.
Very pleased with my 12B4 PP preamp. Combined with its output transformer, the total gain is less than 1, and output impedance somewhere in the 50 ohm range. Pick a different ratio of the transformer and you can get half or double the output voltage depending on your amp needs. Distortion of 0.005% of almost exclusively 2nd and 3rd harmonic. Too many negative posts over the years about the little tube that could. Want more gain, go 12A4.
I also purchased a boatload of six-dubya-six at maybe $3 each years back. Brand new in box GE, you could tell the boxes were likely never opened. I find these need to be run at high current to get in the sweet spot, and coupled with the heater demands their use in a preamp seems overkill to me. I save them for drivers of 2A3/300B. They also don't mind driving interstage transformers in the least.
I also purchased a boatload of six-dubya-six at maybe $3 each years back. Brand new in box GE, you could tell the boxes were likely never opened. I find these need to be run at high current to get in the sweet spot, and coupled with the heater demands their use in a preamp seems overkill to me. I save them for drivers of 2A3/300B. They also don't mind driving interstage transformers in the least.
Most everyone knows about the 6W6. That's not what I was trying to say. Look at the spec sheet for the XYZ123 that I included in post #29 to get the real number. There you will find a relatively unknown tube that may be of interest. The reason for the interest is stated in the quick reference.
I built a cathode follower amp many years ago. I used some 6EM7's loaded with a CCS chip that ran from about 500 volts to get enough drive to feed a paralleled 6336 tube as a CF output stage. It took a large box of tubes to find a pair of 6EM7's with similar gain in the first stage and low overall distortion. I guess that anyone can make a short 9 pin tube with a high gain triode, but the tall ones are all over the place. The 6LU8 and 6LR8 have similar issues, but not as bad as the 6EA7/6EM7s.
I built a cute little 14 WPC P-P 10EM7 amp, and the big triodes were well matched. Didn’t matter that the little ones weren’t, since one was a gain stage and the other a concertina. I’ll end up doing the same thing with the 6LR8’s at some point - the little triodes aren’t well matched at all. And I was using a sand CCS. I ended up giving up and using 10JA5’s (same pentode) and 6BQ7 for the LTP splitter to salvage that particular project. I was hoping to save a bottle per channel and use those vintage-look RCAs, but no dice.
The Novar version of the EM7 that I subsequently found fared better in terms of matching. I made up a set of 4 socket adapters so I can run either or. The Novar version isn’t in such short supply, either.
The Novar version of the EM7 that I subsequently found fared better in terms of matching. I made up a set of 4 socket adapters so I can run either or. The Novar version isn’t in such short supply, either.