[Synchrodyne]

On SSB, thanks for those comments RW, I can now see that the Costas Loop could not create DSB from SSB. It might work to create a mirror image sideband about a notional carrier, but it would have no way of finding the correct carrier frequency. Each sideband would end up with the same magnitude (but opposite sign) displacement error, which would probably derive from the happenstance initial relativity of the incoming sideband and the CIO.

As I recall both Macario and Gosling had an interest in demodulation. Certainly Macario wrote an article that was published in WW 1968 April that described what was a wideband quasi-synchronous AM demodulator. And was there not some interest in DSBSC and DSBDC (diminished carrier) systems for mobile radio around that time.

On the use of DSBDC at HF, the attached item from WW 1955 September about the GPO Rugby station suggests that it could have been used in point-to-point work the 1950s, along with SSB and ISB. The pertinent comment, in respect of the telephony transmitters, is: "When supplied with two separate telephony (or audio) inputs between 100 and 6,000 c/s, the drive equipment provides two independent sideband outputs or, if required, a double sideband output. It is also capable of providing an output of four telephony speech channels 300 to 3,000 c/s wide for operating the main transmitter as a four-channel independent sideband system."

Presumably the DSB output could have the pilot carrier at -26 dB, as with ISB, although it could be at any other level as well. In those days, DSB could have been received one sideband at a time using an SSB/ISB receiver, or by using an exalted carrier receiver that could handle diminished carrier levels.

The Crosby and TMC exalted carrier receivers could also demodulate PM (FM) by the exalted carrier method. In this case the reference carrier was simply phase-shifted by π/2, so that effectively it corresponded with the quadrature (Q) axis. This was thus quadrature demodulation with a narrow-band reference channel. Some other quasi-synchronous quadrature FM demodulators developed in the 1940s had wideband reference channels. These included the Zenith gated beam valve (6BN6) and the Philips nonode (EQ40, EQ80). Also appearing in this time period was the Philco locked-oscillator FM demodulator (FM1000 heptode), which belonged to the fully synchronous group. In the 1950s, RCA developed another quadrature FM demodulator based upon the 6DT6 valve. This acted as a locked-oscillator at lower signal levels, and as a quasi-synchronous demodulator at higher signal levels.

Synchronous techniques were also used – in fact required - for TV colour subcarrier demodulation. Back in the valve era, at the quasi-synchronous end of the spectrum, passive subcarrier regeneration by using the colour burst to ring a very high-Q crystal and then amplifying and limiting its output was probably an extreme example of the flywheel technique. At the fully synchronous end, both locked oscillators and PLLs were used, an example of the latter being the Quadricorrelator circuit.

Valve-era FM stereo decoders – and for that matter those from the discrete transistor era - seemed to be either quasi-synchronous, where the 19 kHz pilot tone was doubled and amplified to provide the 38 kHz reference, or of the locked oscillator type. As far as I know PLL was not used until the IC era, whence it quickly became the preferred horse for the course.

An interesting observation is that during the valve era, FM quadrature demodulation of the wideband type, and the two subcarrier examples mentioned, were widely used for consumer products. But the Crosby exalted carrier AM and PM demodulation system, and the two/three-channel SSB/ISB system, appeared to be confined to professional equipment, notwithstanding that they would have been very advantageous for shortwave broadcast listening. I guess that overall cost and complexity would have been prohibitive for consumer equipment, and I suspect that the very narrow carrier extraction filters were particularly expensive components. The simplest valve-era carrier recovery receiver that I am aware of was the Marconi HR22, which was a lot smaller and a lot less complex than its fully-fledged siblings such as the HR21 et seq. I think it was intended more for smaller-scale broadcast relay work, particularly shipboard. But it was still well beyond the sphere of domestic receivers.

Cheers,