The Franklin ACE 1000 didn’t originally support color, due to an Apple patent on the skeevy color-burst generation circuit.
Franklin eventually worked around this by shipping a daughterboard that plugged into the character generator ROM socket and had a few flying leads to strategic spots on the motherboard. The daughterboard (aka “color board”) bypassed most of the motherboard’s video generation circuitry, and delivered a color video signal that was decidedly inferior to that of the Apple circuit.
Prior to the introduction of the color board, it was common to add the missing components to the monochrome motherboard to replicate the Apple circuit. The procedure is somewhat laid out on page 57 of the “Franklin Service 1 of 2” PDF available on archive.org, but it omits a few very important bits.
I have both a monochrome ACE 1000 (modified for Apple-ish color) and a color board ACE 1000. The video on the latter is, frankly, quite shitty in comparison to the former.
I undertook to convert the color board ACE to mono, with an eye towards creating a kit to make conversion mostly painless. This is a write-up of my notes on converting a color ACE to mono.
First, the color ACE removed a number of components from the motherboard — this makes sense, as the color board bypasses them and thus they were not populated. The following components will be needed:
- ZA4: 74LS166
- ZA9: 74LS257
- ZA10: 74LS151
- ZA11: 74LS194
- ZA12: 74LS74
- ZB4: 74LS194
- ZB9: 74LS194
- RJ6: 2k 1/4-watt resistor
- RJ7: 1.5k 1/4-watt resistor
- Q3 2N2222
- 2716 or 2732 EPROM
There will be small PCBs in the ZA4 and ZA9 sockets; remove them and replace with the appropriate IC.
The Q3 footprint will have two wires soldered into the pads; remove the wires and replace with the transistor.
RJ5 needs to be decoupled from ground. Don’t remove it completely, as the colorburst signal will eventually be fed into Q3 through it, just desolder that leg and lift it off the motherboard.
The color board’s chargen EPROM will not work when plugged directly into the motherboard, as it’s a 2532 and thus not-pin-compatible with the motherboard’s 2716 socket. Burn the bottom 2k of the stock color board EPROM into a 2716 (or twice into a 2732) and use that on the motherboard.
Side note: the lower half of the Franklin chargen EPROM data is byte-for-byte identical with the “Apple II+ – Lowercase Character Generator – 2716.bin” that Henry has made available at https://downloads.reactivemicro.com/Apple%20II%20Items/ROM_and_JEDEC/II_&_II+/. I should contact Henry and ask about the provenance of that binary — it would be interesting if it was originally sourced from Franklin gear. The upper half of the EPROM contains data that’s only valid in graphic mode (I haven’t reverse-engineered the color board to determine why that is) and can be safely omitted from the 2716/2732.
Pin 1 on ZB14 needs to be isolated from pin 2 and connected to the color burst signal generated by the clock circuit. The easiest way to do that is to lift that pin and solder a flying wire to pin 3 on ZD1. A cleaner method would be to cut the trace between pins 1 and 2 on the bottom of the motherboard and run the wire underneath.
At this point the ACE should be fully converted to monochrome-only video.
The new color circuit should look like this:
That circuit is a synthesis of the handwritten notes in the Franklin service manual, the color schematic in the Red Book, and what I found in my mono-modified ACE. I’m waiting on the 27uH choke and the variable capacitor to arrive before I can prototype this, but it’s relatively straightforward:
- the colorburst signal is picked up off pin 12 from ZB12, run through a LC filter, and fed into the output video amp,
- Q3 gates the colorburst signal by the state of TEXT/GR (picked up from pin 2 of ZB14), providing the “color killer” functionality added to the Apple ][+,
- the COLOREF signal from the clock generator is factored into the NOR gate
I’ll update this post with the color circuit procedure when I have this nailed down.