FM stereo decoder

[wa2ise]

I've had thoughts (but insufficient motivation) of using a combination of the PLL stereo decoder ICs with tubes. Like tubes used in color TV chroma difference signal demodulators. The PLL circuit to find and create the 38KHz suppressed carrier, and the tubes to work directly on the audio (where tubes are at their best). A tube like the 6HS8, a pair of dual control pentodes where g1 and the cathode are shared, but g3 and plates are separate. Feed the MPX audio to g1, feed one g3 with one phase of the 38KHz suppressed PLL recreated carrier, and pick off say the left channel audio off the associated plate. The other g3 gets the inverted phase of the 38KHz carrier, and the plate has the right channel audio.

Reason why I never broke out the soldering iron is that the stereo PLL decoder chips don't let me have access to the 38KHz (it's buried deep in the silicon). One could just use generic PLL chips, but I don't know enough about PLL design to do a decent job of it. And you need to get the phase right or else the separation of left vs right channels will be poor.

 

[Osvaldo de Banfield]

I saw basically two circuits: one using diodes (6AL5 or germanium), single and double, and the other using active balanced modulators, using triodes, pentodes or beam tubes. I have some 6BU8, with the common grid as you say, and a common cathode and screen, but isolated anodes and suppressor grids. What I don't know the advantages of using 4 diodes in the balanced (de)modulator, or only two.

The FM set I'm imagining is:

6FQ7 cascode RF amp;
ECF801 Mixer and oscillator;
EF183 FI with AGC;
EF184 FI and limiter;
6AL5 Foster Seeley detector;
ECL80 or 82 as reactance tube (AFC, pentode section) and the triode with a neon in the grid to generate +60V from its cathode to the oscillator section.

The FM Multiplex circuit still not defined.

Opinions and suggestions welcome!.

 

[DF96]

To avoid valve ageing affecting stereo separation, make sure the mixer(s) switch rather than act as multipliers. This is why many valve decoders use semiconductor diodes.

 

[Osvaldo de Banfield]

OK, but it ultimately depends on the amplitude of the 38KHz signal, the greater, the more switching characteristic, or not?

 

[DF96]

Yes, but consider the voltage needed to turn on a germanium diode versus that needed for a thermionic diode or triode switch. A beam deflection mixer needs many volts. Getting volts is easy, but not with only one or two valve stages (to keep power consumption down to what a typical FM tuner can spare) and with low Q 38kHz circuits (to avoid continual retuning). With limited gain a low voltage diode is best.

These days I think I would use silicon diodes, as they have much better reverse isolation and are less affected by temperature than Ge.

A fairly simple design would be:
Cathode follower buffer
Tuned full wave rectifier to double 19kHz to 38kHz, followed by 38kHz (pentode?) amplifier
Diode switches
Cross-mixing to correct for 4/pi factor arising from switching mixer
Maybe a pair of CFs to finish

Somewhere in there you need to include an MPX filter to avoid tweeter frying.

 

[Eli Duttman]

Opinions and suggestions welcome!

Use a 6922 in the cascode TRF section. It's what the type was originally designed for. 😀

Consider forgetting about AGC. A clipping IF strip actually improves AM limiting. Also, use at least 3 stages of IF amplification, before the AM limiting circuitry. While it can be used as a combined limiter/detector, the 6BN6 gated beam tube also makes a very effective stand alone limiter. Read Dr. Adler's 1949 paper on the 6BN6 here. To see a practical use of the 6BN6 as a limiter, refer to the Harman/Kardon Citation III, which is yet another Stu Hegeman masterpiece.

 

[Osvaldo de Banfield]

Eli: Yes, I had the stuff you speak about some week ago, and I printed it and have it in a folder with other rare printed material.

DF: I'm wanting to use only vacuum tubes. My circuitry isn't still defined, but it may be as follows:

A SRPP at input,
A double T tuned circuit to reject SCA.
A small pentode to amplify the 19KHz pilot
A double diode-pentode (like EBF90) as frequency doubler and 38KHz amplifier,
A 6AL5 as demodulator (I also have some 7360 as SSB carrier generator from some SSB scrapped rig).
A 12AU7 as cathode follower.

Is a "may be" only.

Thanks to all replies.

 

[kevinkr]

I've used small signal Schottky diodes as a replacement for GE diodes in some balanced stereo demodulator designs - seemed to work quite well.

 

[Osvaldo de Banfield]

Kevin: I recently downloaded the 7360 and 6BU8 datasheets and appears to be interesting to test them, by the moment I want to avoid use of solid state in this current project.

Thanks for your suggestions.

Have a nice 2013!