Notice the AND gate in the schematic fragments I've uploaded. If PN junction diodes can be used to make an AND gate, then (with proper voltages) a 6AL5 should make an AND gate too. Heaven knows that reverse recovery spikes will not be an issue, as is the case with PN junction diodes.
3, perhaps 4, IF stages feeding a 6BN6 gated beam limiter would seem to provide the hard limited signal needed.
Comments/suggestions are very welcome.
3, perhaps 4, IF stages feeding a 6BN6 gated beam limiter would seem to provide the hard limited signal needed.
Comments/suggestions are very welcome.
Attachments
Last edited:
A 6BN6 can do the gating too, so no need for extra diodes.
I suspect the difficulty with using just diodes for a quarature detector is that the two paths will interact and distort the characteristic. Transistors or valves have better isolation between their outputs and inputs. Hence FM using diodes is almost always the ratio detector, or sometimes the Foster-Seeley detector.
I suspect the difficulty with using just diodes for a quarature detector is that the two paths will interact and distort the characteristic. Transistors or valves have better isolation between their outputs and inputs. Hence FM using diodes is almost always the ratio detector, or sometimes the Foster-Seeley detector.
FWIW, I've seen reports of unsatisfactory distortion performance, when a single 6BN6 is used to amplify, limit, and demodulate. Interestingly, Stu Hegeman's H/K Cit. 3 uses 2X 6AU6s and 2X 6BN6s in front of a germanium diode Foster-Seeley discriminator.
If somebody can suggest a better AND gate built with triodes, instead of diodes, I'd very much appreciate it.
In ICs, PLL and quadrature/coincidence are the "in vogue" methods of FM demodulation. PLL is impractical in 100% tube circuitry, but quadrature/coincidence is quite doable.
I scrounged a "pulled" Ratio Detector transformer off EBone, at reasonable expense. Being somewhat self limiting is a plus for Ratio Detectors. However, a Ratio Detector yields only 1/2 the O/P signal level of that yielded by a Foster-Seeley Discriminator, when fed the "same" I/P signal. Of note is the fact that many, including the Cit. 3 and McIntosh models, "high end" tube tuners employ the Foster-Seeley setup.
If somebody can suggest a better AND gate built with triodes, instead of diodes, I'd very much appreciate it.
In ICs, PLL and quadrature/coincidence are the "in vogue" methods of FM demodulation. PLL is impractical in 100% tube circuitry, but quadrature/coincidence is quite doable.
I scrounged a "pulled" Ratio Detector transformer off EBone, at reasonable expense. Being somewhat self limiting is a plus for Ratio Detectors. However, a Ratio Detector yields only 1/2 the O/P signal level of that yielded by a Foster-Seeley Discriminator, when fed the "same" I/P signal. Of note is the fact that many, including the Cit. 3 and McIntosh models, "high end" tube tuners employ the Foster-Seeley setup.
It is very difficult to have a and gate made with triodes capable of working properly at 4.7 or 10.7MHz.
In the other hand, the un-decouplig resistor at the plate of the 6BN6 can control the amount of RF signal coupled to the limiter grid and its tank circuit (signal level, phase and Q), then from it you can control BW of the detector and eventually distortion. In any case, what are the amount of distortion these sources argument to have?
In the other hand, the un-decouplig resistor at the plate of the 6BN6 can control the amount of RF signal coupled to the limiter grid and its tank circuit (signal level, phase and Q), then from it you can control BW of the detector and eventually distortion. In any case, what are the amount of distortion these sources argument to have?
I doubt if the 6BN6 was intended for hi-fi. TV sound more likely, so a little distortion would not be noticed by most people.
Foster-Seeley and ratio detectors work in quite different ways, although some books try to claim that they are similar. In essence, the F-S derives two IF signals by adding a quadrature signal to the input - but the angle varies according to frequency so the two resultant signal amplitudes vary too. Rectifying both in two separate envelope detectors and taking the difference gives the audio output. The ratio detector works by gating a sample from near the zero crossing of the IF - the exact timing depends on frequency as it is set by the peak signal in the second (quadrature) tuned circuit.
Foster-Seeley and ratio detectors work in quite different ways, although some books try to claim that they are similar. In essence, the F-S derives two IF signals by adding a quadrature signal to the input - but the angle varies according to frequency so the two resultant signal amplitudes vary too. Rectifying both in two separate envelope detectors and taking the difference gives the audio output. The ratio detector works by gating a sample from near the zero crossing of the IF - the exact timing depends on frequency as it is set by the peak signal in the second (quadrature) tuned circuit.
I read in the LS book and in an application note in the blue RCA book about FM, that the main difference between them is that one keeps difference voltage constant, and the other, the sum of the voltages constant, in which the modulated FM signal is rectified by the two envelope detectors.
Correct. Check the 6J10 compactron out, where a gated beam section and a power pentode are packaged in the same bottle. That's the entire audio section.
Back to the vacuum diode AND gate, as triodes are (presumably) too slow at 10.7 MHz. A 6AL5 seems to be fast enough. I'm uploading a fellow's homebrew PN diode AND gate schematic. Please suggest parts values and voltages to adapt the topology from "sand" 5 V. logic to "hollow state" FM. Would buffering the gate prior to RC integrating the pulse train produced make sense?
Attachments
Yes, you will find that many books lump them together as though one is merely a half-inverted version of the other. It took me quite a while to work out how the ratio detector works, and decide that most of the books were wrong.
May be, in those eras also exist Ctrl-C Ctrl-V, but in a more elegant way.
I know a guy who is sure that the theory of class C oscillators may be quite different from those explained in all text books. His hypothesis is that class C oscillates amplifying noise in brutal amounts by regard of the extremely high Q that a positive feed-backed amplifier can give, and filtered by the tuned circuit itself.
Perhaps you and he are true, each one in his shoes.
I know a guy who is sure that the theory of class C oscillators may be quite different from those explained in all text books. His hypothesis is that class C oscillates amplifying noise in brutal amounts by regard of the extremely high Q that a positive feed-backed amplifier can give, and filtered by the tuned circuit itself.
Perhaps you and he are true, each one in his shoes.
There are two standard explanations of oscillators. One is found in most amateur and technician level books, about an ampliifer fed its own output; OK for newbies but doesn't actually get you very far. The other is found in serious textbooks, especially those specialising in the theory of oscillators and oscillator noise etc. The latter basically assumes that oscillation is highly filtered noise, which is what your friend says.Osvaldo de Banfield said:
- Status
- Not open for further replies.