exxos's Atari, Amiga & retro forum

Made a schematic in EasyEDA real quick.

• Pin 24 on both EEPROMs is pulled up from the socket, as this is A16 on original ROMs.

• Pin 24 is connected to GND for Output Enable.

• A16 is connected to pin 2 of EEPROMs from unused sockets to the left.

I would do 10K instead of 33ohm.

D1 is in backwards.

Thanks for the advice guys! I have found some 10K resistors that I wired up to the toggles.

I corrected the Diode on the diagram, and I think I have the real one wired up the same way as in the diagram. The forward bias on my multimeter shows 688 when I put the red probe on the CPU side, black on GAL side. This seems a bit high.

I have some new/unused diodes that read low 600s/upper 500s. I also read that a Schottky diode is best for signals like this due to faster switching speed, but I'm not sure this is one based on that reading (or maybe it is bad). Seems all my new ones are standard diodes.

I updated the diagram with a few more details. the only real change to the circuit was the resistors at this point.

EPROMs check out. The dump I sent to PP are verified OK (exact match to his local files he used to burn them).

TOS 1.04 on $FC0000-$FEFFFF is working as it would in a stock ST.

The other three versions in the $E00000-$E3FFFF address space just sit there with a white screen. No bombs (no bus errors) which I am told signals that the diode is probably OK?

So really it would have to be the GAL chip or my connections going to it. (Maybe I baked the GAL chip soldering it too long, maybe something happened when it got programmed, maybe GLUE isn't getting bypassed because I made a wiring error, maybe I missed an address line, etc etc) PP programmed the GAL, so I don't know what the design is for it and don't have the .JED file.

Luckily I have a few spare GAL16V8 chips I found in my workshop... I programmed one with the file I found here:
http://atari4ever.free.fr (Under TOS section: TOS2_06R.ZIP (G) A TOS 2.06 board which fits in all STs. Includes GAL listing)

I notice it has more connections than the guide I followed for the specific pre-programmed GAL PP sent me. I will install it tonight and see if I can get it to cooperate. I think this is better, the GAL chip won't be a complete black box. (I assume it is easy enough to reverse engineer the JEDEC file to figure out the logic diagram from here, though I am not experienced with this. Would be nice to have so I can fully fill in the circuit diagram - I am happy to do so and post it if someone can point me where to start on that)

I will post an update on how that goes once I have a chance to solder it in, hopefully tonight.

Yeah, it sounds like the GAL is not decoding correctly.

Good news! With that new GAL implementation I found, and adding in a few prototype boards I was able to fix issues with the motherboard layout and also get it mostly working. Got the ROMs moved out from under the PSU, and it all should fit with the shielding back in place still. I also added two flip switches near the joystick ports to avoid drilling holes in the case:
It flipped - it successfully boots everything except for original TOS 1.04 (which runs in address space of original GLUE). It actually tires to boot, repeatedly reading from the floppy but never getting to desktop while on a white screen. Stuck in a loop, but no bombs.

The other three TOS versions work perfect. (Atari original TOS 2.06, and two versions of TOS Improved 1.04) These all run from the additional address space the GAL added.

My theory is the new GAL implementation outputs the different address ranges on different pins... so may only be able to do one or the other. Or maybe I could take an additional CE output signal and combine?
Sorry for such a slow reply. Took a while to go through it, I am also very inundated with school, and I accidentally put the LO ROM in backwards while exhausted one night and baked it so I had to order some new ones.. (If anyone is wondering why the ROM chips look noticeably different now lol)