Goldstar FC-200 VDP Oscillator replacement

Par BignellaMSX

Resident (42)

Portrait de BignellaMSX

19-05-2021, 11:37

On testing the machine I noticed the screen had a very pink/purple look rather than MSX blue.

So I opened up for a general inspection and clean and noticed this:

Looking through the service manual I found that the oscillator was at 10.68(2 or 7)5 MHZ!!!
The closest I could find online was 10.7MHZ so I ordered a few of them, just in case, you know.

As I was desoldering it became apparent just how bad the oscillator was (bad pic but you can still see how degraded it was):

And then I installed the new one:

I'm happy to report that the machine works fine and the start screen is nice and blue as it should be!

!login ou Inscrivez-vous pour poster

Par Grauw

Ascended (10010)

Portrait de Grauw

19-05-2021, 12:17

The normal clock frequency for the VDP should be 10.738636 MHz (3 * 315 / 88 MHz), but on PAL machines the clock frequency is often tweaked to get a more in-spec PAL signal frequency with a 15625 Hz horizontal scanning rate. I presume to improve compatibility with 80’s era TVs. The crystal frequency used is therefore 684 * 15625 Hz = 10.6875 MHz (where 684 is the nr of master clock cycles per line).

At the normal clock frequency the VDP outputs lines at 15700 Hz. Note that this is also slightly out of spec for NTSC (expected 15734 Hz) so Texas Instruments took a middle ground value that was “good enough” for intents and purposes with both PAL and NTSC outputs. p.s. Note that progressive scan itself (“240p”) is also out-of spec, liberties were commonly taken with video signals in 80s game consoles.

Since on this machine the CPU clock frequency is derived from the VDP clock, it also means the CPU runs slightly slower at 3.56 MHz rather than 3.58 MHz.

Btw nice hand-drawn schematics in that service manual!

Par BignellaMSX

Resident (42)

Portrait de BignellaMSX

19-05-2021, 12:42

Thanks for the explanation.
The scanning of the service manual wasn't great so I couldn't tell if it was a 7 or 2 Smile
And yes, the hand drawn schematics were great!

So by using a 10.7MHz crystal I've slightly upped the CPU speed?

Par PingPong

Prophet (3754)

Portrait de PingPong

19-05-2021, 19:44

i also suspect that even the psg sound is affected in tone frequency by the slower clock rate

Par Grauw

Ascended (10010)

Portrait de Grauw

19-05-2021, 22:16

BignellaMSX wrote:

So by using a 10.7MHz crystal I've slightly upped the CPU speed?

Very slightly Smile.

PingPong wrote:

i also suspect that even the psg sound is affected in tone frequency by the slower clock rate

Yes slightly, sounds play 8 cents lower (there are 100 cents per semitone).

Par Pentarou

Champion (389)

Portrait de Pentarou

19-05-2021, 23:52

But the crystal in the OP's first pic is not the system clock, it's the video modulator clock.
In the WIKI there's a couple of pictures, I can only read the value of the French version (the usual PAL 4.43 MHZ), so probably the PAL version is the same value.

Par BignellaMSX

Resident (42)

Portrait de BignellaMSX

20-05-2021, 09:49

From the service manual:
The VDP generates CPU clock (3.58MHz)
A fundamental frequency, parallel-mode X-TAL is used as the frequency reference for the internal clock oscillator, which is the master time base for all system operations.
This master clock is divided by two to generate the pixel clock (5.3MHz) and by three to provide the CPUCLK (3.58MHz).

So yes, it is the VDP crystal but that runs the whole show in terms of frequency.

Par sdsnatcher73

Prophet (2143)

Portrait de sdsnatcher73

20-05-2021, 10:16

Whilst that is not wrong (it is true for every MSX), it is not what Pentarou was saying. The oscillator that was replaced is not the actual source for the VDP master clock. It is used solely on the video board (which just takes the video signals from the VDP and creates the RF/RGB/composite depending on the model), I am not sure what for exactly it is there. The main oscillator that drives the VDP is just above the VDP and is marked as 10.6875MHz. So that the 10.7MHz oscillator is working on the X-TAL location on the video board seems like pure luck.

Par BignellaMSX

Resident (42)

Portrait de BignellaMSX

20-05-2021, 12:34

Ok, yep, I see exactly what you're saying now looking inside again

This is the VDP crystal