View Single Post
Old 06-23-2011, 06:49 PM
Penthode's Avatar
Penthode Penthode is offline
VideoKarma Member
Join Date: Jun 2009
Location: Syracuse, NY
Posts: 659
History of CPA

I have been reading some interesting IEEE articles on early CPA. An especially interesting one is by Bernard Loughlin who worked for Hazeltine Labs in the early 50's. He presented a paper at the Seventh Chicago Spring conference on Broadcast and Television receivers in Chicago in 1966 on the PAL TV system. He makes reference to the early CPA and he puts CPA in the context of what later transpired with the evolution of world standards.

CPA or color phase alternation appears to have originated to help offset the problem of limited bandwidth of the RMA US TV channel eg 6MHz. It is interesting to think that 6MHz is still with us for the ATSC system. But it seems that by the early 50's, after the constant luminance with restricted bandwidth chroma channels has been accepted that somehow interleaving the chroma with he luma as cleanly as possible was the ultimate challenge.

The original NTSC CPA approach was to vestigially modulate B-Y and R-Y with a subcarrier of 3.89 MHz and phase alternate B-Y at either field rate or line rate. (Europe ultimately chose alternating R-Y at line rate). The vestigial sideband approach was to extend the chroma bandwidth and the B-Y alternation would minimize quadrature crosstalk between R-Y and B-Y.

It appears of the two options, the 1952 field test used field alternation rather than line alternation. Is this correct? Slight phase errors would introduce flicker on image edges or chroma transitions due to the crosstalk not being properly integrated or cancelled. It appears that in 1952, a one line delay line was impractically expensive for consumer use and so the true benefit of CPA Phase Alternate Line had to wait a decade and a half.

It appears that the failure of the 1952 CPA test was due to color subcarrier to sound carrier beats (600kHz) and vestigial sideband quadrature crosstalk. Lowering the subcarrier to 3.58MHz and the IQ working overcame the most of the problems.

With regards to the prototype, it appears that if it is indeed still configured for 1952 field CPA, the conversion to IQ NTSC should be relatively straight forward. The crystal frequency of 3.89MHz should indicate whether it is remains the CPA prototype. Chroma demodulation would be restricted to equal band R-Y/ B-Y demodulation since there is no bandwidth differentiation between I and Q in the original CPA prototype. That is, unless it was later further modified.
Reply With Quote