🚀 go-pugleaf

Composite to RGB is similarly easy.  Chips exist that do most of the work.
The National Semiconductor LM1828 an LM1848 for example (called "Chroma
Demodulators") take a composite or chroma input signal and give you "color
difference signals" out at 3 pins.  Then you need to form three well-defined
linear combinations of these outputs, which requires hardly more than a
resistor network, to get RGB outputs.  You will also need a Phase-Locked Loop
circuit to recover the color subcarrier and maybe some amplifier circuits to
set levels and impedances, but the whole thing should be just a few chips.

There may be more sophisticated chips than the LM1828 and 1848 available as
these were 1980-vintage stuff.

Incidentally, a tidbit of info from my video-expert cousin: if you are AC
coupling composite or luminance video signals (as I plan to to in making my TV
set into a monochrome monitor), use BIG FAT capacitors like 50 microFarads or
you will get noticeable black level drift across the picture.  The right way
to do it is to DC couple, but you can get away with AC coupling if you use a
big enough capacitor.

The chroma signal, being centered on 3.58 MHz, is less critical of
low-frequency coupling and is shown in "typical application" diagrams for the
LM1828 as being coupled through a .01 uFd capacitor into the 1K-ohm chroma
input.

Hmm... I think this diagram is telling me that the correct mixing ratios
(which would be reciprocals of resistor ratios going to the summing point) are
as follows:

			LM1828          LM1848
		     -----------------------------
		R-Y     0.8             0.95
		G-Y     0.25            0.35
		B-Y     1.0             1.0


If anybody really wants to build converters for either direction, I could
probably wheedle schematic diagrams out of my cousin, who is a fellow
Atari owner and so should be sympathetic.


						-John Sangster
						jhs at mitre-bedford.arpa
				<anypath.to.ihnp4>!ihnp4!linus!mbunix!jhs