Futterman stage max input

[regal]

Generic technics varioation of the futterman output:




I assume in pactrical application one would need a cathode resistor/diode on each tube. Example here:




Now it would seem that for the bottom futterman triode the max ac in would be < the cathode bias, otherwise you are running up past 0 grid volts. So in the example above it would be 3xVf ~ 2.1 Vrms input ? (for discussion purpose forget the poor choice of biasing with 1n4002's.)

But this isn't the case in practice, so how does this work, would we say that when Vacin is > Vkbias we go into class B?


Would there be any advantage to biasing such that the kathode is always at a higher voltage than the ac signal input peaks?

In other words if my max expected from the gain stage is 15 Vrms, would it be worthwhile to find a tube with an operating point that is comfortable at a 15V cathode bias?

[Rod Coleman]

The Cavalli schematic looks like a line amp to me, so would not be usable for higher levels.

But Yes, if you want to use this circuit with 15V drive to the outputs, a different choice of output tube is best. Maybe a triode-connected PL84 (=15CW5??). The B+ would need to increase to 300-350V and the current increased to a very high 40-50mA - but it would then give the bias headroom you need.

Alternatively use resistors in the cathodes, making them followers. Then, as the grid voltage swings high, the cathode follows high too, and maintains the grid-cathode voltage (provided the anode current is neither cut-off, nor saturated). Care would be needed with this choice though - if the cathode resistors get too high in value, and the output sees a capacitive load, there could be stability problems due to this RC pole.

[regal]

Yes I was showing the Cavalli scheme as an example.

Thanks rod, if my ultimate goal is a 15Vrms output, would the bias requirement be half that (7.5V) since each tube is working with half the signal in a futterman output and combined ?

[Rod Coleman]

Dave, 15Vrms is tough - the top tube is a follower, meaning that both output grids would see about 15V rms too. 15Vrms means about 15 x 1.414 (22V approx) of peak swing.

Perhaps combination bias is worth looking into. Partly zeners/LEDs, partly resistors, so that the cathode has some padding. Maybe a 6.2V zener (these are the quietest) plus 5-7V from a cathode resistor - the 5-7V would increase during positive peaks, and decrease during negative peaks.

A small output tube is still a good choice, especially types from AC/DC radio & TV, where specs are given for <200V operation (since each tube sees half the B+).

[Rod Coleman]

The best tubes for this circuit might be Directly Heated types.

The RCA #71A needs -27V of bias for 135V and 17.5mA, which fits right into that circuit with a B+ of 300V or so, and lower current than other types, for the same bias.

[regal]

Originally Posted by Rod Coleman

The best tubes for this circuit might be Directly Heated types.

The RCA #71A needs -27V of bias for 135V and 17.5mA, which fits right into that circuit with a B+ of 300V or so, and lower current than other types, for the same bias.

But the low mu sort of kills the futterman output impedance advantage, I would just use classic P-P w/ OPT but the 15vRMS out into my 600 ohm load makes finding a suitable OPT difficult, the secondary DCR just kills the op impedance and getting a decent primary inductance means a winding ratio that has me swinging so many volts as to have speaker amp THD levels.

This may be one case where a hybrid just makes more sense.

[Rod Coleman]

Yes, the 71A will mean an open-loop output impedance of 500 ohms, but the circuit is using feedback to lower it further. It depends on how low you need it.... but if you add a power FET source follower, then <1 ohm - with no feedback - is easy.