A0' - A8' at expansion port

[dikdom]

Dear Martin,

this is an excellent (and tremendous) piece of work you made, congrats! What I would like to ask what was behind the decision that you changed the address lines on the expansion port compared to a stock ZX81? You routed there the A0' - A8' which are the address lines after a resistor instead of the A0-A8 CPU addr lines. The same applies to the internal RAM's address lines also...

I'm building an expansion cart for the ZX81 (it has 32kRAM, 64k EEPROM, 8M flash, SD card slot and a cursor joy port). On a stock ZX81 the WRX mode works properly. I'm thinking (unfortunately didn't get there yet to test) if would the WRX mode work with the +38 board, completely from the cart's RAM (yeah, my cart disables the internal RAM).

EDIT on 2023-10-31: we managed to test it and it doesn't work, even with the simple text display. Somehow there is a wrong character code is read from D_FILE - the external RAM (which is enabled on [$4000-] range access + (mreq or m1 or rfsh) + (rd or wr or rfsh) and in addtion to that the row counter doesn't get latched as in all the 8 lines the same character image is displayed for a character (different one for different characters - but not the proper character's first line)

Thanks!

[PPZ-00]

Hello, Is there any workaround for that? I have built the ZX81 Plus 38, in the beginning I thought that I did something wrong :) Now I read it's the design flaw.

[dikdom]

Hello, it turned out that my design was not a good one... So what I do is that I have a GAL chip on my cart and it

• disables internal RAM
• enables external RAM on $4000-$FFFF access, except $C000-$FFFF read on /M1 phase
• enables external ROM on $0000-$3FFF read access, except /RFSH phase. This is the case when the ULA reads the character info from the internal ROM. So originally I enabled this internal ROM access only when /RFSH and /MREQ both were active (low). That didn't work with +38. When I removed the /MREQ from the GAL's condition table (so the internal ROM was enabled for the whole /RFSH cycle) then the +38 worked well. It seemed that it works pretty well with stock ZX81, so I made ~40 pcs of cart for fellow ZX81'ers. The fun started here: it turned out that there are problems with specific ULA+ROM chip types. The problem was gone, when the GAL condition was reset to the stricter one (both /RFSH and /MREQ is low). So now I have two GAL firmware codes. I need to doublecheck with the TS1000/TS1500 which of the firmware they are OK with.

[dikdom]

Hello, Is there any workaround for that? I have built the ZX81 Plus 38, in the beginning I thought that I did something wrong :) Now I read it's the design flaw.

Uh-oh. Now I realized that you're asking about the +38. Well.. I built a +38 and it properly works for me. (without my stupid cart)

[PPZ-00]

Your cart is great, as well as Plus 38 :) Great designs, both of them. I built two of Plus 38, and 5 of your carts (I will keep two of them) but when using it with Plus 38, I cannot start some software, I get garbled screen. So I thought that I did something (I initially installed nmos Z80 CPU in one of them, but quickly realized that it will not work, even though it boots to a prompt). Now I can't run some of 32k programs from your demo dump for eeprom, and that is actually the only problem. So I thought naturally that I did something wrong in assembling the Plus 38. Anyway, even if this can not be fixed, it's a great project, and your cart is awesome. Although a bit expensive to build because of the two GALs (I used ATF22V10 and ATF20V8, it works great).