The 6U8 was designed to be a VHF radio/TV oscillator and mixer (though later superseded by lower noise dual triodes). A mixer is operated in a non-linear mode to efficiently produce intermodulation distortion. The sum or difference frequencies produced are part of the principle of a superheterodyne circuit, almost universally used for radio and television receivers since the 1930s. The triode is non-linear since that works best for oscillators, limiting output amplitude without clipping.
Of course other uses of this tube are common - many tubes have found alternate uses. The 6AW8 (another triode/pentode) was designed as a TV sync separator and video amplifier according to GE - but GE later used it as voltage amplifier and phase splitter as the 6U8 was. Fun fact: the Mosley CM-1 amateur receiver was an excellent performer with five tubes - all of them 6AW8. Used for every stage of the radio - oscillators, mixers, IF and audio amplifiers.
Of course other uses of this tube are common - many tubes have found alternate uses. The 6AW8 (another triode/pentode) was designed as a TV sync separator and video amplifier according to GE - but GE later used it as voltage amplifier and phase splitter as the 6U8 was. Fun fact: the Mosley CM-1 amateur receiver was an excellent performer with five tubes - all of them 6AW8. Used for every stage of the radio - oscillators, mixers, IF and audio amplifiers.
Sure they can. I'm not saying it's useful, but they can...
Attachments
For the last schematic in post 20 could onevconnect R3 to the cathode of U3 and let u3 catode-resistor go to negative voltage for biasing 6v6? Cathode resistor of 6v6 to be set to small valve like 10-ohm just for measuring idle- current.
Am I missing something or would this be a feasable DC coupled tube-amp with slightly more output compared to cathode biased 6v6??
Am I missing something or would this be a feasable DC coupled tube-amp with slightly more output compared to cathode biased 6v6??
You could triode-strap the pentode, and then it is just two triodes in parallel. That's may be what @artosalo suggested.
Mercury99,
Experiment, just like you are doing.
But it is helpful to know some of the factors that affect the results you get.
You used an 832A with a g1 grid resistor of 330k.
The 832A data sheet calls out a maximum of 25k (of course that is for use as an RF amplifier that has lots of drive voltage to g1, and that draws lots of grid current. Fortunately you are using a self bias resistor, R1, 800 Ohms. Hopefully there will not be thermal run-away; hopefully you are using less power to the screen, g2, and less power to the plate, versus the dissipation ratings of the screen and the plate.
For audio service, and driven from most small input tubes, you do not want to draw grid current (*) . . . it will disturb the self bias voltage, it will cause distortion, and it will cause blocking).
Because the 832A and input tube uses global negative feedback, much of the "sound" of the tubes is smoothed out, and also just changing to different tubes, the global negative feedback can hide much of the sound differences between tubes of different types, and multiple tubes of the same types too.
But, because you are using the 832A in beam power mode, it needs negative feedback, otherwise there will be too much distortion, and the damping factor will be too low.
Yes, the 832 was meant to draw g1 grid current for RF amplifier service. But that required fixed negative bias from a 25k Ohm grid resistor. And, self bias (cathode bias) was not recommended (dis-recommended to make it clear).
Cathode followers have low impedance. That means for very high frequencies, it overcomes the capacitances of the 832A.
But at 20kHz, a driver tube rp, in parallel with the driver tube plate load resistor RL, should be able to take the output tube to 50kHz or beyond.
A cathode follower with RC coupling to the output tube . . . if there is grid current, causes the problems in (*) above.
Whenever the signal voltage from either a driver tube, or from a cathode follower makes g1 of the output tube near to, or the same as the cathode volts, g1 will draw grid current.
g1 near to the cathode volts, the contact potential will cause grid current to be drawn, usually a very low current.
When g1 goes slightly above the cathode voltage, grid current will be drawn, and this can be enough to be a problem, especially when RC coupling is used.
The 2A3 usually does not need global negative feedback, the distortion and the damping factor are reasonable without it; much of that is dependent on the quiescent current (enough current helps the plate resistance to be low), and the primary impedance of the output transformer, if high enough, gives low distortion and good damping factor.
A 2A3 typical plate impedance, rp, is 800 Ohms.
If output transformers were lossless (they are not), the damping factors would be:
2400 Ohm primary / 800 Ohms = 3, a Damping Factor of 3;
and 3200 Ohm primary / 800 Ohms DF = 4.
Expect DF of about 2.5 for the 2400 Ohm primary, and about 3.5 for the 3200 Ohm primary (output transformers do have insertion loss, much of that from primary DCR and secondary DCR).
2A3 tubes are not like RF beam power tubes. 2A3 were not meant to be operated with any grid current.
Again, a medium impedance from a driver tube, or from a cathode follower, should make the 2A3 work to 50kHz and beyond.
Since there is an RC coupling, and no grid current, what is the need for a cathode follower? again, RC coupling, and any grid current combine to cause (*) just as noted above.
Happy designing, building, and listening!
Experiment, just like you are doing.
But it is helpful to know some of the factors that affect the results you get.
You used an 832A with a g1 grid resistor of 330k.
The 832A data sheet calls out a maximum of 25k (of course that is for use as an RF amplifier that has lots of drive voltage to g1, and that draws lots of grid current. Fortunately you are using a self bias resistor, R1, 800 Ohms. Hopefully there will not be thermal run-away; hopefully you are using less power to the screen, g2, and less power to the plate, versus the dissipation ratings of the screen and the plate.
For audio service, and driven from most small input tubes, you do not want to draw grid current (*) . . . it will disturb the self bias voltage, it will cause distortion, and it will cause blocking).
Because the 832A and input tube uses global negative feedback, much of the "sound" of the tubes is smoothed out, and also just changing to different tubes, the global negative feedback can hide much of the sound differences between tubes of different types, and multiple tubes of the same types too.
But, because you are using the 832A in beam power mode, it needs negative feedback, otherwise there will be too much distortion, and the damping factor will be too low.
Yes, the 832 was meant to draw g1 grid current for RF amplifier service. But that required fixed negative bias from a 25k Ohm grid resistor. And, self bias (cathode bias) was not recommended (dis-recommended to make it clear).
Cathode followers have low impedance. That means for very high frequencies, it overcomes the capacitances of the 832A.
But at 20kHz, a driver tube rp, in parallel with the driver tube plate load resistor RL, should be able to take the output tube to 50kHz or beyond.
A cathode follower with RC coupling to the output tube . . . if there is grid current, causes the problems in (*) above.
Whenever the signal voltage from either a driver tube, or from a cathode follower makes g1 of the output tube near to, or the same as the cathode volts, g1 will draw grid current.
g1 near to the cathode volts, the contact potential will cause grid current to be drawn, usually a very low current.
When g1 goes slightly above the cathode voltage, grid current will be drawn, and this can be enough to be a problem, especially when RC coupling is used.
The 2A3 usually does not need global negative feedback, the distortion and the damping factor are reasonable without it; much of that is dependent on the quiescent current (enough current helps the plate resistance to be low), and the primary impedance of the output transformer, if high enough, gives low distortion and good damping factor.
A 2A3 typical plate impedance, rp, is 800 Ohms.
If output transformers were lossless (they are not), the damping factors would be:
2400 Ohm primary / 800 Ohms = 3, a Damping Factor of 3;
and 3200 Ohm primary / 800 Ohms DF = 4.
Expect DF of about 2.5 for the 2400 Ohm primary, and about 3.5 for the 3200 Ohm primary (output transformers do have insertion loss, much of that from primary DCR and secondary DCR).
2A3 tubes are not like RF beam power tubes. 2A3 were not meant to be operated with any grid current.
Again, a medium impedance from a driver tube, or from a cathode follower, should make the 2A3 work to 50kHz and beyond.
Since there is an RC coupling, and no grid current, what is the need for a cathode follower? again, RC coupling, and any grid current combine to cause (*) just as noted above.
Happy designing, building, and listening!
How about wire the pentode as a triode and then place it on top of the 6U8 triode section as a cascode.
A cascode section will give a significant increase in gain, increased bandwidth (better stability) as the Miller effect is eliminated in the input section and perhaps with less distortion depending on signal levels.
It may limit your maximum voltage swing on a 300V rail. Would have to spice it to see.
A cascode section will give a significant increase in gain, increased bandwidth (better stability) as the Miller effect is eliminated in the input section and perhaps with less distortion depending on signal levels.
It may limit your maximum voltage swing on a 300V rail. Would have to spice it to see.
Most of todays signal sources, CD players, phono preamps, control preamps, tuners, etc. have medium output impedances that can drive a 25k, and even a 10k volume pot.
With those volume pots, almost all input tubes have plenty of bandwidth to drive the miller effect capacitance without changing to a cascode stage.
And, cascode output impedance sometimes are fairly high, so the output stage miller effect load on the cascode output can reduce the bandwidth.
It comes down to calculating the performance of a single triode, versus cascode, and how well they drive the output stage.
High frequency bandwidth; distortion; and maximum signal voltage to drive the output stage.
That is one reason so many designers use software simulation programs.
I do not use such software, but calculate and estimate by Longhand and experience.
But I have to be well energized to do so, and in the mood.
With those volume pots, almost all input tubes have plenty of bandwidth to drive the miller effect capacitance without changing to a cascode stage.
And, cascode output impedance sometimes are fairly high, so the output stage miller effect load on the cascode output can reduce the bandwidth.
It comes down to calculating the performance of a single triode, versus cascode, and how well they drive the output stage.
High frequency bandwidth; distortion; and maximum signal voltage to drive the output stage.
That is one reason so many designers use software simulation programs.
I do not use such software, but calculate and estimate by Longhand and experience.
But I have to be well energized to do so, and in the mood.
Do nothing. The pentode has no influence on the triode if not used even if pentode filament is active.
True, its not necessary to use it at all. And if it goes bad, so what. But us OCD folks will find “suggested” uses for it.
Connect the pentode section in parallel with output (6V6) tube.
You'll get slightly more power for free.
You'll get slightly more power for free.