|
Its been a while but last week I got a bee in my bonnet to try to fix the
last color problem in my CT-100.
This was caused by a mention of old color bar generators in another forum here.
I realized that I could make one using a computer program to generate an
image, screen grabbing it, and putting in a file with Photoshop. I did so
both with the "real" ones that RCA generators made back in the day, and
a different one that showed the whole 360 degrees of color. I also made
one of each with black bars and one where the Y was constant over the whole
image.
This meant that I could now do the "main official" color adjustments rather than
the "second best" ones they describe using essentially what became SMPTE bars.
It turned out that the angle I had set for the phase difference between I and Q was
90 degrees, exactly! I set the Q phase as described, and the I looked perfect.
Then I tried setting the I gain using their procedure which involves looking at the R, G, and B grid signals. It would not go high enough
for the scope traces to look perfect. So I tried increasing the Y gain by putting
a 2700 uF cap across the I demodulator cathode resistor. This allowed getting the
three signals ALMOST right .... but red and blue looked clipped! Turning down
color gain ALMOST fixed this. But ... the bars LOOKED worse!!! Too much blue on the right center part of the screen. And still, the top of one sine curve clearly
was clipped. So I removed the cap. Still a bit clipped! Turning down either
contrast or color gain or I gain didn't help.
Turning down video gain on the modulator did get rid of the clipping. Then I checked
the I and Q waveforms for correct 90 degrees ... still correct. Adjusting I gain by eye resulted in a much better color rendition. It was about 80% of full scale, rather than
100% as before. The RGB waveforms were good but not perfectly as in the manual,
but no sign of clipping.
Then I looked at SMPTE bars .... as good as I had them before, but good, still
not perfect, but slightly differently not perfect.
Then I look at my standard test pictures, including Dorothy first seeing the Scarecrow, some flowers (the ones I posted) and, and, most important,
a bunch of Christmas cards I got off the net that had balls the exact color violet
that was not reproducing well before. Better. Maybe the
blue gain and screen were not set quite right ... too high a color temp.
Changing that a tiny bit helped the color too. FINALLY!
Then, GO CUBS!, I watched the Cubbies beat the Sox by one run, and the colors
matched my Sony HDTV even better than before, especially yellow and yellow-green,
which had never been really right. Note that the real RCA bars don't show yellow and yellow-green while my 360 degrees ones do.
Morals: don't overmodulate, and use the old style color bars that have the color phase
rotating 360 degrees over the whole horizontal cycle.
Also, the color bars with the constant Y background make a good test for old B&W sets as well as color ones: there should be no brightness difference across the whole
screen. My CT-100 was good before, now its near perfect. The TRK-12 really has
absolutely no luma variation, though up close you can see the color signal if
you move your eyes at the same rate the diagonal bars move.
|