What does that trimmer cap on the NES (and famicom) board do

[wyatt8740]

If you look at a toaster NES, down by the expansion port, there's a little green trimmer pot. You don't even have to open the system to see it. Does anyone know what this thing does?

On the Commodore VIC-20 I have in a corner of my basement, there's one that relates to the color output. Does the pot do the same thing on the NES?

[lidnariq]

Frequency trimpot. It'll slightly change the frequency that the master clock runs at. Should speed up or slow down everything (CPU, sound pitches, video) by some trivial amount, maybe 0.1% at most. This could cause a visible color shift. My understanding of NTSC is that you should get some amount of rainbow shift between the left and right ends of the screen, except that the same design in the NES is also used in the original IBM CGA, where it apparently just rotates the hue across the entire screen without any difference from left to right.

[blargg]

SNES also has one near the power switch, and in one SNES I have its slight misadjustment was the cause of having colorless video output.

[l_oliveira]

The purpose of that adjust is allow a service technician to finetune adjust the frequency of the color carrier for the CVBS output.

[mertkara]

@wyatt8740, I agree with the last comment, it doesn’t change the frequency of the whole controller and circuits running inside it, it only adjusts the carrier the carrier for CVBS. The designers had to place an external pot at that time in order to ensure that technicians can quickly change the CBVS based hardware to adjust according to various available video terminals for this output.

[lidnariq]

No thanks for the necropost that adds nothing.

And you're wrong: there's only one clock source in the entire machine, so it changes the frequency of everything.

Maybe you're going to get pedantic at me in response, so let me cut that off: crystal oscillators can not be detuned by more than about 1000ppm, which is a small enough amount (2 cents) that it's less than the JND for pitch (6 cents). Furthermore, NTSC isn't guaranteed to correctly decode color unless you're within 3ppm (although in practice it's not that strict) of the correct frequency. So it only seems as though it changes only the colorburst frequency.

[rainwarrior]

...get pedantic at me in response... it's less than the JND for pitch (6 cents).

Ultra pedantic note: the 6 cents noticeable difference is for pitches played in isolation. It you play two pitches simultaneously, you can determine much finer tuning than 6 cents by the beating they produce together. If you don't have an external reference pitch (e.g. maybe you're running two NESes at once) it is correct that you should not be able to notice, though.

[l_oliveira]

Well, the pot is there to allow for adjustments for the Chroma carrier, it has been mentioned before that if it deviates too much the TV will lose the color and the image will be black and white. Anyway, any variation on that trimmer would result on a clock fluctuation small enough to have zero difference on the frequencies output by the 2A03 APU.

So in the end what are we discussing ?