Sinclair ZX80 / ZX81 / Z88 Forums

I'm happy soldering, and I've got a 16K expansion board... but I'm after the holy grail of the 32K

This is the one I'm currently looking at - http://www.zx81.de/english/32k-rame.htm

OK - so reading things... there's no real gain by going for the 32 and I should stick with teh 16k because it'll be a lot easier. Yes?

BUT... I'm reading about Hi-Res graphics... and ... well... if there's going to be no real benefit, then I won't and stick at the 16K... but I'd be interested to hear thoughts please.

Why not go full 64k with battery backed up RAM? It’s even possible to replace ROM area with RAM, so you can load your own custom ROMs.

https://web.archive.org/web/20210224003 ... mories.htm

msknight wrote: ↑Sun Dec 24, 2023 10:47 am This is the one I'm currently looking at - http://www.zx81.de/english/32k-rame.htm

That design is slightly misleading, in that the 100Ω and 10kΩ resistors, and the diode are only needed if going for the battery back-up option.

The two NAND gates on the left of the schematic carry out the extra address decoding.

Internal RAM is already HRG capable.

Mark

mrtinb wrote: ↑Sun Dec 24, 2023 10:58 am Why not go full 64k with battery backed up RAM? It’s even possible to replace ROM area with RAM, so you can load your own custom ROMs.

Let me learn to walk first before I set off on a marathon

msknight wrote: ↑Sun Dec 24, 2023 10:52 am BUT... I'm reading about Hi-Res graphics... and ... well... if there's going to be no real benefit, then I won't and stick at the 16K... but I'd be interested to hear thoughts please.

For what it's worth (maybe not a lot), here's what I did with the 2nd non-contiguous 16K of the 32K version of the ZX-Team mod.

I have 3 favorite programs that all install themselves above RAMTOP: TK (a ToolKit), ZXAS Assembler, and ZX-Bug. I wanted to be able to swap between the 3 of them on the fly as I develop a program, ie. without having to SAVE my work-in-progress, LOAD the new utility, and then LOAD my program again. (Such a pain using actual tape,...)

It took some engineering, but I now have a "master" program in 2 parts. The first is a fairly long LOAD, which carries in all 3 machine code images as the contents of a string variable, and does a machine code copy to cache them into the 2nd bank of 16K. It also sets RAMTOP to a value large enough to hold the largest of the 3 tools.

Then I load a skeleton program I call "MODEL", which has 3 different BASIC routines above the line 9000 mark, out of the way. Each routine does a ML copy of just the utility I want at the moment, to the prime area above RAMTOP. (It also runs any init logic required by the target utility.) And of course "MODEL" is my ongoing model for development, with glorified SAVE subroutines and miscellaneous bells and whistles.

Somewhat to my surprise, it all actually works quite well. I can dynamically swap in whichever of the 3 utilities I desire at the moment, without rebooting. When I'm done my development, I swap in the TK toolkit one last time and and use its DEL command to delete all the lines above 9000, and then call for a final SAVE. All that is left to be SAVEd (in the prime area, below RAMTOP) is the program I wanted to create.

Engineering this entire mechanism was a bit of a fun project challenge for me at the time. I'm ashamed to admit I've done very little development actually on the ZX81 since. I guess it was anti-climactic, as well as being eclipsed by slicker options. But it was still the best bit of "90% BASIC" programming I've ever done on the ZX81.

Another use of the 48-64K area I just remembered.

My "model" program contains a routine called FLASHOUT, which copies in its entirety the normal 16-32K program area to the higher one. Why?

Sometimes during development, programs crash. (I know I know,... this may come as a shock,...) Murphy's Law dictates it will crash BEFORE I've saved my latest changes. (I like working with actual tape, remember.)

A NEW command does not clear the upper memory area. So if, while working on a still dangerous program, I FLASHOUT the code and the computer subsequently crashes to a "K" prompt, I can LOAD a stub program (a quick load), and FLASHIN the code back to the active area. Then SAVE, as I should have done the first time.

(Obviously any use of the FLASHOUT routine clobbers the 3 utilities cached in the upper area. A choice has to be made there.)

Again this is only advantageous if, like me, you like working with, (but not waiting for) actual tape. Obsessive Compulsive Disorder, I think its called.

TMAOne wrote: ↑Wed Jan 03, 2024 2:14 pm My "model" program contains a routine called FLASHOUT, which copies in its entirety the normal 16-32K program area to the higher one. Why?

:

A NEW command does not clear the upper memory area. So if, while working on a still dangerous program, I FLASHOUT the code and the computer subsequently crashes to a "K" prompt, I can LOAD a stub program (a quick load), and FLASHIN the code back to the active area. Then SAVE, as I should have done the first time.

Great rutines.

It sound a lot like RAMDOS2 or SIMDOS2, which creates a RAM disk in upper memory, and you only need to save your work to tape before you turn of the power. With Wilf Rigters 64k RAM you have the memory battery backed up, so no worries here.