Infinite Impedance detectors.

[Synchrodyne]

Bipolar transistors in what superficially look like emitter follower circuits appear to have been used within radio ICs for AM demodulation. Possibly this was easier than including diodes in ICs. Did those transistor demodulators have better linearity than diodes when working at the lowish signal levels found within ICs?

In respect of its CA3088 AM radio IC, RCA said: “The output of Q10 [the 2nd IF stage output emitter follower) is fed directly to Q12, an emitter follower operating at a quiescent current of approximately 100 microamperes. This stage becomes a detector by connecting the proper filter circuit to the emitter.” Here are the circuit details:



Another example is provided by the National LM1863, which was a high-performance AM radio IC:



This had a separate IF output, for example for feeding an AM stereo decoder. The signal level at that output was around 150 mV, which was also the level that was fed to the on-board demodulator.

And the Plessey SL623C communications subsystem IC, which had what was described as a full-wave AM demodulator:





Cheers,

[Synchrodyne]

Returning to infinite impedance demodulators as such, another (and quite late) commercial application was in the HMV 558 AM-FM tuner. That feature was the subject of a brief mention in Wireless World 1960 October:



I have not found any circuit details. One can imagine that given that it was an AM-FM tuner, an independent triode for AM demodulation could have been made available by using, on the FM side, an ECF80 where otherwise an EF80 might have been chosen, such as for a limiting stage. Both Leak and Jason did this ECF80-for-EF80 "substitution" in some of their tuners, in those cases to provide an independent triode for use as an output cathode follower. The ECF80 pentode was not quite the same as the EF80, but evidently delivered adequate performance in IF strips. That still leaves the question as to how AM AGC bias was developed, and did the AGC rectifier damp the AM final IF transformer.

Anyway, it suggests that late in the valve era, one maker saw merit in the infinite impedance demodulator for use in a hi-fi application.


Cheers,

[Synchrodyne]

The previous post prompted some pondering as to whether, in AM-FM equipment, a triode could be employed as an infinite impedance demodulator on AM and as an AF cathode follower on FM, without unduly complex switching. In the case where a cathode-follower output was considered to be desirable and so was going to require an extra triode anyway, it would be a way of including an infinite impedance AM demodulator without requiring still another extra triode. Conceptually, the switching required would seem to be no more complex than that involved when an ECH81 triode section is switched between oscillator service for AM radio and AF preamplifier service for a gram input and/or FM radio.

Taking that further, if the FM side included a full limiter (e.g. as required to drive a Foster-Seeley discriminator) then the limiter grid could be used as a source for AM AGC bias (as well as for FM). This would require that AM IF was also routed to the limiter grid, with implied damping of the final AM IF stage. But I think that it has already been determined that for hi-fi applications, this is not a major issue.

So far so good, but having gone this far, then why not also take the AM demodulated AF signal from the limiter grid, in which case the extra triode reverts to being a simple cathode follower in both the AM and FM modes. It has been shown, in the VHF case at least (see post #33), that limiter grid AM demodulation can also be low-distortion. And in the speculative case here, there is a cathode follower to provide high impedance loading for the demodulated AM signal further aiding and abetting the low distortion case. It looks as if we have talked ourselves into and then out of using an infinite impedance demodulator.

Now we need a precedent for the use of limiter grid AM demodulation in domestic AM-FM receivers. It certainly does not appear to have been commonplace, but I am aware of one example, namely the GE TI1000 group of AM-FM-FM Stereo table radios., which I think was the first from GE to include FM Stereo according to the Zenith-GE system adopted by the FCC in 1961. (Zenith’s corresponding model was the MJ1035W.)

GE used a 6AU6A sharp cutoff pentode as FM limiter and AM demodulator. As well as demodulated AM, AGC bias was taken from the limiter grid, and a single unswitched AGC line was used for both AM and FM. I’d say that GE was aiming to do the AM part as economically as possible whilst retaining performance comparable to that of a good standard AM receiver, and at the same time not materially detracting from FM performance. Obiter dictum in the context of this thread, but I think worth mentioning for general interest is that the 6AU6A 1st FM IF amplifier also doubled as the AM self-oscillating mixer, with AGC bias fed to its suppressor grid. Presumably GE did not want to compromise FM performance by using the relatively noisy (even by heptode standards) 6BE6 that was normal American practice in this position. Thus was the AM RF/IF/demodulation function handled entirely by the same valve types that most likely would have been used for a broadly similar FM-only receiver.


Cheers,

[Synchrodyne]

A quick search has shown that the use of the diode formed by the FM limiter valve grid and cathode for AM demodulation in FM-AM receivers was not particularly uncommon.

It was used at least by GE, Hallicrafters, Motorola, Packard-Bell, RCA and Sylvania.

GE made an early application in its X-415, an immediate post-WWII AM-FM receiver that covered both the old and the new American FM bands, and provided some SW band coverage. The FM limiter and AM demodulator was a 6SH7 pentode. For AM, its screen HT bias (supplied by a divider) was disconnected. However, that feature was not used on later implementations by GE and others. Limiter grid AM demodulation was by no means used on all GE receivers, some of which used the 6T8/19T8 triple diode-triode for both AM and FM demodulation; perhaps not so surprising considering that GE had originated that valve back in 1947.

Another early application was in the Hallicrafters SX-42 of 1946. This used a limiter grid AM demodulator at VHF (above 27 MHz, IF 10.7 MHz), but a conventional diode AM demodulator at HF (below 30 MHz, IF 455 kHz).

The valve most commonly used for limiter grid demodulation was the 6AU6/12AU6. But also used were the 6SH7, 7R7, 12BA6, 15EW6 and 19HV8. The last-mentioned was a 1961 GE release, and combined a sharp cutoff RF/IF pentode with a 70 mu AF triode. I’d say that it was intended primarily for use in FM receivers where solid-state diodes were used for demodulation, providing the final IF amplifier/limiter stage and the 1st AF stage, typically followed by a normal 12-watt anode dissipation output valve, such as the 6AQ5. And it also provided for AM demodulation in FM-AM receivers without requiring any additional active components.

This form of AM demodulator does not seem to have much coverage in the literature. For example, it was not mentioned by Langford-Smith.

A tentative conclusion is that the limiter grid AM demodulator was used in commercial equipment more often than was the infinite impedance type. It was of course effectively restricted to FM-AM receivers, where the limiter was present for FM purposes. But it does show that equipment makers were open to departure from the regular diode where this conferred an advantage.


Cheers,

[Synchrodyne]

A different, i.e. not AM audio demodulation, application for an infinite impedance detector was in the AGC circuit of the BBC UN1-585 VHF TV receiver unit.

This had a sync-cancelled black-level AGC system, in which a transistor infinite impedance detector was used to peak-rectify the sync-cancelled video, thus measuring the latter at blanking level.

In this case it looks as if this form of detector was chosen to provide appropriate loading for preceding sync amplifier stage and for the composite video signal mixed with it, and for easy matching to the following grounded base AGC amplifier stage. Highlighted copies of the schematic and the pertinent part of the circuit description are attached. The full document is available at: http://www.bbceng.info/ti/eqpt/UN1_585.pdf.







Cheers,