Now that the CRT board (housing the CRTC, SRAM, and character ROM) is finished, it’s time to serialise the dot and colour data. This next stage involves the "Colour, Video & RF Circuit" board. Its job is to convert parallel data into EGA-style signals, composite video, and RF.

While the final build will include an RF modulator, I’m leaving it out for now. I haven't been able to find a new one yet, so I may have to salvage one from a C64. An authentic, static-filled picture will have to wait!

The board features a 15-pin D-sub connector with HSYNC, VSYNC, and two pins for each RGB channel. Although I haven't found official documentation, it’s a safe assumption that this is EGA (specifically 6-bit 'Enhanced' RGB). My plan is to hook this up to a GBS-8200 (EGA-to-VGA converter) first, as getting the composite video working adds a layer of complexity I’d like to avoid for now.

• Clocking: At the bottom of the board, a 74157 divides the main board’s clock frequency and switches between low and high resolution.
• Serialisation: A shift register handles the pixel data serialisation, while an octal buffer drives the 6331 Colour PROM.  The 6331 acts as a Colour Look-Up Table (CLUT). Since these are one-time programmable and no longer available, I’ve replaced it with an ATF16V8 (GAL).
• Timing: A 17.734 MHz master clock and 74S04 inverters form a Pierce oscillator for video timing. This provides the high-frequency "dot clock" and is divided down to create the precise PAL colour burst signal. This frequency is exactly four times the PAL subcarrier (4.43361875 MHz).
• Logic: The top row of ICs contains 74156 dual 2-to-4 decoders/demultiplexers with open-collector outputs. These handle the colour mixing, functioning as a DIY Digital-to-Analogue Converter (DAC). They take digital data from the EGA lines (Ra, Rb, Ga, Gb, Ba, Bb) and decode them into specific voltage levels. A resistor ladder is connected to their outputs to create the various voltages representing different colours for the composite and luminance signals.

Installing that ladder was a massive pain—so many different values to keep track of! But finally, the signal passes through an amplifier and buffer stage before exiting via a dodgy-looking RCA connector (I accidentally chose the wrong footprint on the PCB 🫤). So what’s Next?

Next up is testing the EGA output. I didn't actually need to install half of the components for the EGA to work, but the board looked so bare without them! In the next update, we’ll see what happens when I run it through the GBS-8200. I’ll just turn it on and squint my eyes a little—I'm sure it’ll be fine.