I did not suggest a design but only said that high efficiency can be reached. This is the same for both SE and PP. However it will not happen if the impedance of the output transformer and the operative conditions are the only subject of the discussion....
How can you comment on something you did not check first; am I missing something?
About elaborating: check the differences between anode- and cathode follower output stages. Broskie (TubeCad) is a very good read; I don't pretend to teach here.
Just joking, but it seem not so easy to get it.
Nobody expect you to "teach" here, but with your statement, that contains even more details than the first one, the usual level of bla bla is undermined.
I don't think you contributed anything worthy or constructive to this thread. Just the usual bla bla.
P. S. Me, too can point to hundreds of tube amp books. Very easy thing to do. Just go on and study for yourself.
P.P.S. Check for this and for that. The web is a good read on this topic, you know.
Nobody expect you to "teach" here, but with your statement, that contains even more details than the first one, the usual level of bla bla is undermined.
I don't think you contributed anything worthy or constructive to this thread. Just the usual bla bla.
P. S. Me, too can point to hundreds of tube amp books. Very easy thing to do. Just go on and study for yourself.
P.P.S. Check for this and for that. The web is a good read on this topic, you know.
Last edited:
I downloaded the entire magazine article from 1955 and read all of it. There are some novel features explained in each of the individual test amps. It is a worthwhile read, but some of these technique are not necessary today....They didn't have mosfet followers in 1955.
The amp in figure 12 extracts 65 watts from a pair of triode wired 807's into a 2800 ohm CT OPT on 420 volts by driving them deep into AB2. I know from experience that this is not very far from tube arc territory. A tube arc will flash through a 6L6GA (same guts as an 807) at about 75 watts in triode mode in AB2. You can get to a little over 100 watts in pentode mode before the arc happens, again on 3300 ohms.
75 watts per pair in triode is easy with 6550's or KT88's in AB2 on 3300 ohms. I did not push the expensive tubes further:
75 watt per channel triode mode tube amp prototype - YouTube
The push pull 300B amp that I made nearly 20 years ago makes 28 to 30 watts into a 6600 ohm OPT in AB1. It made more power when I first built it, but despite the claims right on the outside of the Sovtek 300B box many of those tubes will run away at 450 volts, and stability was only achieved well below 400 volts of B+. Many of these tubes had a portion of the filaments exposed on the top where there was no grid between it and the plate. No amount of negative voltage on the grid could cut those tubes completely off.
I have been selling TSE and now TSE-II boards for about 15 years. They run a 45, 2A3, or 300B in single ended A2. A Chinese 300B can make about 12 to 15 watts in this manner. I have driven the the two channels out of phase and wired a P-P OPT from plate to plate to make an AB2 test amp.
A pair of old 45's can make nearly 20 watts into a 3300 ohm OPT on 300 volts of B+. ]
A pair of 15 year old Chinese 2A3's made 38 watts on 380 volts into a 2500 ohm OPT.
Do I believe a good set of 300B's can do 50 watts into 3300 ohms, Yes, but the key words are GOOD set. My old Sovteks are not up to the task so I did not try them.
Merely driving the tube CLEANLY beyond AB1 into AB2 WILL extract more power out of most tubes compared to AB1 under the same B+ and OPT load conditions. How much more power can you get? That all depends on the tube.
Tubes like the 6L6 / 807 can only pull their plate down to the 70 to 100 volt range before saturating at zero volts on G1. Drive G2 to +20 volts and that number is more like 20 to 40 volts. Less voltage dropped in the tube means more across the OPT, making more power.
The increased plate current capability afforded by further saturating the tube also allows for a lower OPT impedance offering an even greater power gain.
TV sweep tubes developed from the 1950's through the 70's have much larger cathodes with far more emission capability, so there is no need for AB2 operation, and many of them do not like it. When the tube can pass 1.4 AMPS with zero volts on G1 and 150 volts on G2, AB2 is not needed.
All of these are important, especially if you want to build an all tube design.
The "kicker" boosts the B+ to the cathode follower a bit under positive grid conditions in the output tube. This is to offset the Rp of the driver tube when the impedance of the output tube grid abruptly drops from hundreds of K ohms to hundreds of ohms. Swap the cathode follower for a mosfet, with a few ohms of Rdson and the kicker is not needed.
A stiff power supply is required for a low distortion design, especially when high dynamic tube currents are involved. A CLC supply with a fat output cap is often OK for music, but the amp may test poorly under constant sine wave operation. A regulated G2 supply is required for any pentode stage, and some means of bias control is needed for high powered amps. No OPT is perfectly symmetrical, and no two output tubes have matched Gm over all operating conditions, so some degree of mismatch in tube current tends to generate the lowest distortion. Bias servos and even microprocessor controls are available today, and beneficial in a big tube amp with multiple output tubes. I'm working on a 500 WPC tube amp.
It has become far to easy for an amp designer to pile on lots of GNFB to make the numbers look good, thus chasing the life out of an amp. On the other hand too much Inter Modulation Distortion makes an amp sound "hazy" "veiled" or whatever term is in vouge today, and contributes to listener fatigue. For every amp there is a combination of feedback systems that will make it sound good and perform well when driving a speaker or a pile of test equipment. Often these "systems" involve multiple local feedback loops instead of a big GNFB loop. Finding the right combination of feedback takes some work that often involves listening, testing, and more recently computer simulation. Given good vacuum tube models, simulation can be a valuable tool, just don't blindly trust the results....model, build, test, listen, repeat.....until you are happy with all the results.
The impedance of the OPT is one of many "knobs" that an amp designer can play with to reach his performance goals. In a push pull amp lowering the OPT impedance generally increases power output and distortion. There is a point where the output tube can't comfortably drive that impedance, and lowering it further rapidly increases distortion and risks possible tube damage. I will try different load impedances and measure the plate efficiency at each load value. The point where the efficiency is maximized is usually very close to the optimum for a given tube and B+ voltage. This point will be different between AB1 and AB2 as explained previously.
The amp in figure 12 extracts 65 watts from a pair of triode wired 807's into a 2800 ohm CT OPT on 420 volts by driving them deep into AB2. I know from experience that this is not very far from tube arc territory. A tube arc will flash through a 6L6GA (same guts as an 807) at about 75 watts in triode mode in AB2. You can get to a little over 100 watts in pentode mode before the arc happens, again on 3300 ohms.
75 watts per pair in triode is easy with 6550's or KT88's in AB2 on 3300 ohms. I did not push the expensive tubes further:
75 watt per channel triode mode tube amp prototype - YouTube
The push pull 300B amp that I made nearly 20 years ago makes 28 to 30 watts into a 6600 ohm OPT in AB1. It made more power when I first built it, but despite the claims right on the outside of the Sovtek 300B box many of those tubes will run away at 450 volts, and stability was only achieved well below 400 volts of B+. Many of these tubes had a portion of the filaments exposed on the top where there was no grid between it and the plate. No amount of negative voltage on the grid could cut those tubes completely off.
I have been selling TSE and now TSE-II boards for about 15 years. They run a 45, 2A3, or 300B in single ended A2. A Chinese 300B can make about 12 to 15 watts in this manner. I have driven the the two channels out of phase and wired a P-P OPT from plate to plate to make an AB2 test amp.
A pair of old 45's can make nearly 20 watts into a 3300 ohm OPT on 300 volts of B+. ]
A pair of 15 year old Chinese 2A3's made 38 watts on 380 volts into a 2500 ohm OPT.
Do I believe a good set of 300B's can do 50 watts into 3300 ohms, Yes, but the key words are GOOD set. My old Sovteks are not up to the task so I did not try them.
Merely driving the tube CLEANLY beyond AB1 into AB2 WILL extract more power out of most tubes compared to AB1 under the same B+ and OPT load conditions. How much more power can you get? That all depends on the tube.
Tubes like the 6L6 / 807 can only pull their plate down to the 70 to 100 volt range before saturating at zero volts on G1. Drive G2 to +20 volts and that number is more like 20 to 40 volts. Less voltage dropped in the tube means more across the OPT, making more power.
The increased plate current capability afforded by further saturating the tube also allows for a lower OPT impedance offering an even greater power gain.
TV sweep tubes developed from the 1950's through the 70's have much larger cathodes with far more emission capability, so there is no need for AB2 operation, and many of them do not like it. When the tube can pass 1.4 AMPS with zero volts on G1 and 150 volts on G2, AB2 is not needed.
All of these are important, especially if you want to build an all tube design.
The "kicker" boosts the B+ to the cathode follower a bit under positive grid conditions in the output tube. This is to offset the Rp of the driver tube when the impedance of the output tube grid abruptly drops from hundreds of K ohms to hundreds of ohms. Swap the cathode follower for a mosfet, with a few ohms of Rdson and the kicker is not needed.
A stiff power supply is required for a low distortion design, especially when high dynamic tube currents are involved. A CLC supply with a fat output cap is often OK for music, but the amp may test poorly under constant sine wave operation. A regulated G2 supply is required for any pentode stage, and some means of bias control is needed for high powered amps. No OPT is perfectly symmetrical, and no two output tubes have matched Gm over all operating conditions, so some degree of mismatch in tube current tends to generate the lowest distortion. Bias servos and even microprocessor controls are available today, and beneficial in a big tube amp with multiple output tubes. I'm working on a 500 WPC tube amp.
It has become far to easy for an amp designer to pile on lots of GNFB to make the numbers look good, thus chasing the life out of an amp. On the other hand too much Inter Modulation Distortion makes an amp sound "hazy" "veiled" or whatever term is in vouge today, and contributes to listener fatigue. For every amp there is a combination of feedback systems that will make it sound good and perform well when driving a speaker or a pile of test equipment. Often these "systems" involve multiple local feedback loops instead of a big GNFB loop. Finding the right combination of feedback takes some work that often involves listening, testing, and more recently computer simulation. Given good vacuum tube models, simulation can be a valuable tool, just don't blindly trust the results....model, build, test, listen, repeat.....until you are happy with all the results.
The impedance of the OPT is one of many "knobs" that an amp designer can play with to reach his performance goals. In a push pull amp lowering the OPT impedance generally increases power output and distortion. There is a point where the output tube can't comfortably drive that impedance, and lowering it further rapidly increases distortion and risks possible tube damage. I will try different load impedances and measure the plate efficiency at each load value. The point where the efficiency is maximized is usually very close to the optimum for a given tube and B+ voltage. This point will be different between AB1 and AB2 as explained previously.
Please re-read the circuit. The transformer in the cathode is a power transformer to heat the filaments. The audio power comes out the plates to the audio transformer, just like 99% of all tube amps do.
The special advantage of class 2 (grid current) operation is Increased Current. There should then be no question "Why?" the load impedance was reduced.
Yes, the 300B is not specified for grid current; its special claim is high plate current without entering grid current operation. It might burn-up. At least too often for theater or home use, perhaps hours per day. However this application is Disk Cutting. Less than 20 minutes per side, much prep and post-work so maybe two sessions per day. If that day is billed at $1,000, and new 300Bs sold for $50 at the time, occasional trouble is affordable. (That's another reason to double-up: if a tube dies mid-side, the amp will "work" to finish cutting the side, and the part-power situation may not be noticed unless a MAX peak comes along.)
Attachments
You are right PRR, I read it like the cathode windings are part of the output transformer.
There seems to be cathode feedback (6BQ7), right?
And yes, not an audio application.
There seems to be cathode feedback (6BQ7), right?
And yes, not an audio application.
A disc cutter is an audio application, and one that requires good fidelity as it is cutting the master disc for pressing vinyl records.
Many many years ago I got an old monophonic Rek-O-Cut cutting lathe. I was a kid at the time, so I did what curious kids did, I took it apart.....this was before I learned how to put stuff back together. The only things that I remember was the name, and the 2A3 tubes in the cutting amp. There were at least two 2A3's.