View Single Post
Old 10th Dec 2014, 12:54 am   #22
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: Quasi-Synchronous Demodulation

Thanks for the nice summary of AM demodulator types.

Regarding the PLL type fully synchronous demodulator and its handling (or not) of PM, that came up in the discussion I had with the Liniplex designer. The Liniplex demodulator was designed to track relatively slow carrier movements (e.g. Doppler shifts) but I was advised that it would not follow phase modulation, such as data transmissions piggybacked on LF and MF transmissions, or AM stereo transmissions. Thus any PM on the signal would be demodulated in the Q channel. As well as transmitted PM, this applied also to oscillator phase noise. That was of no consequence in the DSB mode, which used only the I-channel demodulator output, but it did show up in the LSB and USB modes, which used both the I and Q channel outputs. The Liniplex was actually very tenacious once locked on to a transmission. A good test of HF receiver PLL demodulators was how they handled the Big Ben chimes on the BBC World Service. The Liniplex F2 stayed locked, but the Sherwood SE3 often unlocked briefly. The Sony ICF2010 almost always unlocked and wandered around for a while. I suspect that it unlocked when it saw the Big Ben chimes coming...

As an example of “horses for courses”, the ICs used for C-QUAM AM stereo decoding used an interesting mix of techniques. Envelope demodulation was wideband quasi-synchronous, using limited but unfiltered signal as the reference, so that none of the (stereo) PM came through into the envelope audio channel. But a PLL was used for I and Q demodulation, presumably because precise phase and phase relativity were required.

When PLLs were used for TV vision demodulation, a wider PLL bandwidth of around 500 kHz was suggested by National Semiconductor in respect of its LM1823 IC to ensure that incidental phase modulation (ICPM) was tracked by the reference and so did not transfer to the output. Interesting here is that the BBC, for its rebroadcast TV receivers, moved from PLL fully synchronous demodulation (in the RC5M-502) to quasi-synchronous demodulation (in the RC5M-503) in which the carrier was extracted by a narrow band filter and conditioned in a low-phase shift limiter, with the objective of minimizing the effects of ICPM. Again a “horses for courses” choice, I think.

The problem of AM-to-PM conversion during limiting in the carrier path for the quasi-synchronous case prompted the thought that this problem might have been eased by the availability of integrated circuit differential pair limiters. In the late 1960s and early 1970s, RCA often made the point that such ICs provided symmetrical limiting – and lots of it - over a wide range of input signal strengths. That might explain why ICs that incorporated wideband quasi-synchronous demodulators, that is with the reference obtained by hard limiting unfiltered input signal became quite common from the late 1960s. Previously, with valve or discrete solid state technology, wideband quasi-synchronous demodulators without carrier filters seem to have been scarce.

Cheers,
Synchrodyne is offline