Posted By
Retroshire
on 2021-07-31
07:28:01
|  Re: Commodore 16: Black and white screen on Composite, color on the RF
@Mces solved the problem. In short: a broken capacitor trimmer + a broken crystal.
He wrote to me in an email:
Hi Iwan,
I tested your board (CVBS cable) with a little monitor and.... it's colour is working!
With a second monitor (CVBS cable) it has no colour....
Hypothesis: something is constantly on the border between "work" and "doesn't work",
First suspect: clock frequency (color coding is derived from the same clock), must be 17.734472 Mhz (+/- 200 ppm: ppm = parts per million).
Measured 17,727060 MHz (-418 ppm) TOO MUCH FAR!
Variable capacitor does not tune it, desoldered and tested: it is not good (too less capacitance).
Replaced the capacitor trimmer with proper 40pF, but it doesn't tune properly....
Replaced quartz, now clock can be tuned!
Switched on for 15 minutes (thermal stability reach) and tuned to 17,7344558 MHz (-1 ppm: better than frequency counter precision).
Board test: colour is present on both monitors!
Why on RF your monitor work?
Inside the color decoder there is a quartz which must be locked with the C16 color frequency,
at board powering on C16 frequency has to be inside the "capture range" of the colour decoder, and after to be locked the C16 frequency can remain inside the "lock range".
https://www.rfwireless-world.com/Terminology/PLL-lock-range-vs-capture-range.html
The "capture range" is narrow than "lock range", so my theoretical hypothesis is:
The analogical process of RF modulation (into C16) and RF tuning (into TV) generate a little bit of noise inside the signals,
so it's possible that the noise result in a little jitter
https://en.wikipedia.org/wiki/Jitter
that can permit to lock the decoder, so the decoder remain locked ("lock range" is bigger than ""capture range")
|
