Prototype set is here

[Penthode]

I have attached another version. The drawing is sideways: at least it is better than upsidedown.

It is interesting to put this paper on CPA in context with the development of the all-electronic compatible color. By mid 1952, much of the development of what would finally end up in the December 1953 standard was there. The constant luminance principle was there as was the chrominance quadrature modulated on a odd multiple of half line frequency to facilitate luma/chroma interleaving. But also the elements which would soon disappear including the burst pedestal as was assymetrical B-Y and R-Y bandwidths and the higher subcarrier frequency.

It is interesting that the original RCA RGB Dot Sequential sample frequency is fairly close to the later NTSC subcarrier of 3.58MHz. The original impetous for CPA was to extend the bandwidth of the chroma channel within the limited bandwidth of the US 6MHz wide channel. Hence the choice of 3.9MHz for the CPA subcarrier I suggest was for two reasons: to provide wider bandwith (vestigial sideband chroma) for both R-Y and B-Y and to extend the luma channel response. Remember that there is a luma 3.58MHz trap in early NTSC receivers which effectively cuts the high frequency luma response to 3MHz or less. A 3.9MHz chroma subcarrier would have extended the luma response a little bit.

CPA employed for PAF or PAL (Phase Alternate Field or Phase Alternate Line)was simply inverting the R-Y axis at either field rate or line rate. The reason for doing so was to eliminate differential phase errors which would result in hue errors and to cancel quadrature crosstalk between R-Y and B-Y at demodulation. Remember, crosstalk was eliminated with I and Q by ensuring both were double sideband to 500kHz and only I extended from 500kHz to 1.5MHz. CPA facilitated vestigial sideband R-Y and B-Y and the crosstalk was to be cancelled by averaging.

In 1952 only visual averaging of the PAF and PAL was available which led to the crosstalk and hue errors resulting in the annoying field flickering (PAF) and a vertical line crawl (PAL). The European engineers later termed the PAL line crawl as "Hannover Bars" and may be readily seen on what is referred to as "Simple PAL" decoders or PAL decoders without the electronic delay line averaging.

In 1952, the NTSC was exploring PAF as the field flickering was deemed less objectionable to the line crawl. I agree with the later NTSC assessment that moving away from the CPA to the I Q solution was one step beyond CPA in the development of compatible color TV. Successful deployment of CPA with PAL fifteen years later depended upon the one-line delay line which was not readily available in 1952.

We should look forward to finding out what remains of the CPA circuitry in this extremely unique prototype. I trust you will share with us the circuit as you trace it. I am particularly interested in the CPA transformer: the alternating R-Y phase I feel would have been derived from reversing at field rate either the chroma before demodulation or the regenerated subcarrier used for R-Y demodulation. I suspect the transformer is in the chroma signal path as per the diagram in the attached article?