Replacement ROM

[Moggy]

If push comes to shove lass, and you get stuck, find a few 2764s, more than one needed in case of duff ones, from wherever ( ebay for me no hassle all worked Y.M.M.V ) post the buggers up to me and I'll fire up the Spectrum and put what ever ROM code want on them ((Big Bang is favourite. )

As for EEPROMS only having a ten year life, that's no worry to me as I don't think I have that long left!!

[msknight]

As for EEPROMS only having a ten year life, that's no worry to me as I don't think I have that long left!!

I hear you ... I'm trying to be self sufficient and plan ahead. I know I've got to take this bios chip out and clean the pins, but it only takes a moment to bend a pin I didn't mean to, and I'm in trouble. I've cack handed enough chippery to its untimely doom! I'd like to know enough to dig myself out of the simpler snow drifts, if you know what I mean

[Lardo Boffin]

I have a few spare originals if disaster strikes with cleaning this ROM. I normally keep them for replacing 550s so I can fit a Wespi but can spare one if required.

[1024MAK]

That does bring me to the ultimate question

Well, we know the answer is 42...

of whether I should be using eeproms at all and instead look at eproms. From what I'm reading, eeproms have about 10 to 15 years before they start to forget the data on them.

I'm using eeproms because it was the easiest way to program them, at the time, on Linux.

There are EPROMs that are far, far older than 15 years old with there original programming that still work

EEPROMs are just developments of EPROMs. And just like manufacturers gave conservative estimates of how long EPROMs would retain data, it's likely the same with EEPROMs.

Mark

[msknight]

I found an interesting article on how they determined the longevity of memory - https://hackaday.com/2023/12/21/how-do- ... 100-years/

The AT28C256-15PU is rated at 10 years on the datasheet. I guess if I could find an alternative EEPROM chip with a longer period, then I'd be happy sticking with eeprom. It's going to be an interesting hunt.

[msknight]

There are two that came up which (to my novice eyes) appear to be variants of the same chip. Not sure if they're suitable, but they have retention of 100 years on their datasheets..

https://www.digikey.co.uk/en/products/d ... 5/18160055
https://www.digikey.co.uk/en/products/d ... 2/18159991

[msknight]

Minimum order quantity of 312, totalling over £8,000. Ouch.

[msknight]

Or this? Seems more affordable and still 100 years on the datasheet...

https://octopart.com/datasheet/x28hc256 ... l-16754218

[1024MAK]

Using an AT28C256 EEPROM in a ZX81

Pin 1 is A14. On a ZX81 board, this is connected to +5V.
Pin 26 is A13. On a ZX81 board, this is connected to +5V.
Hence the programmed code must be in the top quarter of the memory in the chip (0x6000 to 0x7FFF).

Pin 20 - as per fitting an EPROM.
Pin 23 - as per fitting an EPROM. See this post.

Pin 27 is /WE. On a ZX81 board, this is connected to +5V. This is okay, as it prevents accidental writes to the chip.

The connections for the X28HC256 are the same as the AT28C256.

Using an AT28C64B EEPROM in a ZX81

As above except:
Pin 1 is an output. Best practice would be to isolate this pin from the board (bend it out).

Pin 26 is a NC (no connect), so being connected to +5V on the board does not matter.

Mark

[Moggy]

Using an AT28C64B EEPROM in a ZX81

Pin 1 is an output. Best practice would be to isolate this pin from the board (bend it out).

Pin 26 is a NC (no connect), so being connected to +5V on the board does not matter.

Mark

The data sheet shows pin 1 as no connect so need to bend out or am I reading it wrong?

https://docs.rs-online.com/d264/A700000006402645.pdf