Hi,
Just wanted to give an update. I have not given up on this project (repro ALI main board) but did take some time out from working on the actual design (and other projects). During that time, I have still thought about the project and considered several factors:
The cost to produce this size pcb is high;
The cost to populate this board as a 100% repro is very high;
Several IC's are expensive and very hard to source.
Given these factors there would probably not be many (only a few very committed real keen) takers for the current pcb design. It would be expensive for me to prototype and test these given what I think would be a very low uptake when the cost to populate a complete board is considered.
Remember, for the ALI compared to other pinballs this pcb is the cpu, switch, display, lamp and solenoid driver all in one pcb. On top of that this pcb does not use any matrix drivers for the lamps or switches, so the parts count is higher. Given this I feel a repro that involves some changes might be more useful. This is without changing the game code, as I feel that should stay original.
So, I have added the following objectives for changes to the pcb (let's call it repro Mk1):
1. Replace the 6504 with a 6502 (still allow an alternative overlayed socket footprint to fit a 6504 as a build option);
2. Replace the three 6530's RRIOT's (IC3, IC5, IC6) with one 2732 EPROM and 2 x 6532's RIOT's (and some discrete decoding logic);
3. Replace the expensive hard to get 75492 driver IC's with currently manufactured and available 75452 IC's;
4. Look for a suitable (current) replacement for the MPS6531 transistors;
5. Consolidate unused gates in buffer and driver IC's to reduce total IC's (I can already see about 7 IC's that are candidates);
6. Replace the setting jumpers, use jumper blocks or dip switches to make it less messy (but still large board space);
6. Reduce board size as much as possible;
7. For this version it would stay with original firmware (no software modifications).
For a further improvement it will need changes to the firmware (let's call it repro Mk2):
1. Change cumbersome jumper settings for another solution, possibly using NVRAM and settings via coin door switches;
2. Add some basic self-test on power up utilizing the on-board LED's and displays for indicating self-test faults;
3. Add saving high score if NVRAM used;
4. Look at better I/O protection circuitry for inputs where needed.
5. Further reduce board size (with jumper setting area mostly removed).
I am still pondering about the 75452's. They are currently manufactured and available from reputable supplies like Digikey. Each IC has only 2 drivers, but they are rated better than the 75492's. There are 32 lamp drivers (used) in the final revision ALI PCB. So, this would require 16 of these alternative drivers IC's. I think it is still an easier solution than the hard to get original 75492's or the other option I am considering of using 32 buffer gates and 32 transistors along with associated resisters.
Things completed since last update posted:
Added the 6502 IC;
Changed the connectors for J4-J6 to the correct (original) pcb footprint (which are pricey but still available from Digikey);
Reviewed the code firmware (for I/O and memory addresses actually used) and chip hardware decoding to work out initial decoding for the replacement 6532's.
Added 2732 EPROM to schematic;
Commenced schematic changes to replace 6530's with 6532's.
In summary, Mk1 would be a little smaller, use available parts, cheaper to build and complete pcb and still be 100% firmware compatible with an original pcb. Mk2 is about changing the way settings are done and adding some features and requires changes to the firmware.
Comments on any of the above are most welcome. I will continue on Mk1 and hopefully post more info shortly.
Edit: I almost forgot. I have had the first prototpye pcb's made to plug into the edge connectors to help test the lamps and switches.