This one had me dancing for a while.
The fault was no boot so you'd think it would be pretty straightforward!
When I first started this 'adventure' I didn't have a machine to test it in and had to run on the bench with just a PSU. I didn't have all the wires from the PSU connected at the beginning as I thought I'd only need the supply for the logic. That was the first mistake. I now know that there is a feed from the transformer that feeds the watchdog cct and without it it stays stuck in reset.
To compound that TR3 which switches the 100Hz feed was low in gain so that was replaced. I could now see it was attempting to boot but not getting there.
Point to note: I have seen intermittent reset faults caused by a fracture of this track under the power connector. The slightest movement of the PSU wires caused the board to go into reset. One to look out for.
After that it got a bit easier, the 6809 had a stuck address line so it was whipped out, socket fitted and chip replaced. Mind you I didn't know it was the processor so the RAM chip was also removed and socketed and the original put back.
Back on test and the reset LED would go out but sometimes it was rather sluggish, sometimes 10-20 secs before going out. The data and address lines were all active now so things were looking up.
Still nothing happening though so it was back to the logic probe.
Results compared with a good board showed no activity on the Prom select lines. I was now thinking there maybe a problem with the 11820 chip. These are rarer than hens teeth but luckily after removal and swapping one over from a good board it made no difference. The suspect chip also worked flawlessly in the good board so that was a relief.
Next step was to swap out the RAM chip, being as I'd already socketed it. Now I noticed the LED extinguished regularly after 1 sec on power up so the RAM was dicky!! Also the machine was now loading software so it was almost time to turn off the iron!!
By this time I now had my own machine so with the board connected and power applied things started happening. When it got to attract mode horror upon horror, there were no lights. The machine would play quite happily though.
So, back to the bench again. Looking at the schematics and the lamp matrix enable is driven from the sound chip, yep, I had to look twice at that. Pin 9 was stuck so I needed an AY8912 and I didn't have one.
After a bit of searching on the web I discovered that these chips were used in old Sinclair Spectrum 128's and as luck would have it my son still had his old one tucked away at the bottom of a cupboard. Being as the Sinclair doesn't use the data outputs I swapped them over.
Back on test and bingo, all lamps working and job done.
Time for a beer methinks!!
View the full article
