Those Magnificent Television Tubes

[smoking-amp]

Oh,
I guess if all the plate outputs are individually AC coupled, then X*X + (-X)*(-X) cross coupling will work. But just using the X*X output alone should be fine here.

(the two BDT tubes over the CCS are used to get the grid 1 functions linearized, otherwise one would get X*(X^3/2) due to non-linearity of grid 1. Some cathode degeneration resistors could also be included for even better linearization here. Or, for extreme linearity, a third BDT could be used in place of the CCS in a split cascode like arrangement, so not using grid 1 control at all then. )

[Tubelab_com]

I would also suggest getting some 6197 tubes...

The 6197 is a 6CL6 with a cathode coating formulated for "computer use" where long periods of operation (days) with zero cathode current are possible. I have several hundred of each tube, over a thousand 5686 tubes, and several boxes full of other military surplus and computer tubes from the 50's and early 60's. All are well used and have not seen power since the 1960's.

I did experiment with all of these during the development of my vacuum tube VCA (voltage controlled amplifier). The VCA started with a clone of the famous Fairchild 670 compressor circuit, build with genuine 6386 tubes (real expensive now) from the 1960's as my benchmark. I used the same Edcor transformers in each build to remove that variable from the testing. My vacuum tube VCA works as good as the 670 design, but uses a pair of $1 tubes and a pair of mosfet followers for buffers. The vacuum tube Moog ladder VCF (LPF) is not quite up to Moog specs yet though.

If I were to revisit the squarer circuit with opamps, it would be trivial. We (Motorola) developed a circuit in the early 80's to recover a data stream from the discriminator of an FM radio. We called it a "center slicer" and it used a single stage of an opamp or comparator chip. The raw DC coupled discriminator output fed one input, while the LPF (a few Hz corner) filtered signal fed the other input. The output of the chip is the data stream. The corner frequency of the LPF has to be high enough to track frequency drift between the receiver and the transmitter, but low enough not to follow some pretty slow data (around 100 baud). Max data rate was 2400 to 4800 baud.....it was the 80's and we were using 25 KHz RF channels.

[spinch]

I build valve power amps, and I use non - microphonic EF183 and EF184 in the first stage with excellent results. The EF183's - frame grid varimu - sound a lot better - anyone know why?

[Tubelab_com]

The variable mu remote cutoff characteristics of the EF183 (6EH7) tend to generate considerable second harmonic distortion. The level of 2H goes up as the signal level increases. Many listeners find this effect pleasing.

I use similar tubes in my variable gain amp. The 6EH7 is one of the possible choices and it is quite cheap. Like the old Fairchild 670, the even order harmonics are reduced by using balanced push pull operation.

[Kay Pirinha]

6EH7 is equivalent to EF183, see:

EF183 and 184 are quite cheap tubes here also, as are their Russian equivalents, 6SH51P and 6SH15P.

Best regards!

[spinch]

As balanced operation from XLR outputs of many sources (eg.CD players, etc) are already built in, there is no need for additional phase splitters from unbalanced RCA sources-it's already done for you. So I use two EF183's each driven from their two XLR outputs in a fully balanced p/p arrangement driving a p/p output pair of EL34's - thus completely minimising even harmonic distortion.