I am building a 20MHz 6502 project and it occured to me - which is better, the EPROM or the EEPROM?

I have heard some EEPROMS can only be written a certain number of times before they fail and sometimes even let the smoke out, but an EPROM takes forever to program and needs a UV programmer. Help?

(If it makes anything easier, I have decided to use SRAM for this project.)

The link to another topic on this same machine: http://forum.6502.org/viewtopic.php?f=2&t=3416

Please do comment about your choice on the poll.

_________________
It may look hard, but it won't take long if you take it one byte at a time.

EPROMs also tend to be slow in comparison to modern EEPROMs. WDCs boards use the SST39SF010A 70ns EEPROMs. The fastest in that series is 45ns (22Mhz). Flash typically has a write endurance of 100,000 cycles. That's a lot reprogramming operations.

You could use another microcontroller to stop the 6502 and copy a ROM image in a RAM chip as we have discusses elsewhere.

_________________
Andrew Jacobs
6502 & PIC Stuff - http://www.obelisk.me.uk/
Cross-Platform 6502/65C02/65816 Macro Assembler - http://www.obelisk.me.uk/dev65/
Open Source Projects - https://github.com/andrew-jacobs

Have a link to the discussion?

_________________
It may look hard, but it won't take long if you take it one byte at a time.

You can look here:
viewtopic.php?f=4&t=3374

My SBC-4 also used an AVR to hold the 6502 in reset while it loaded RAM.

http://sbc.rictor.org/info4.html
Don't let the CPLD scare you away - it could be replaced by a GAL or simple logic. It was mostly for the expansion port buffering.


Daryl

_________________
Please visit my website -> https://sbc.rictor.org/

Both EPROMs and EEPROMs have write-endurance limits. I've never used a parallel EEPROM, but with the EPROMs I've used over the last nearly three decades, I apparently never reached the limit as I developed firmware. My (E)EPROM programmer which goes into an ISA (I think that's what it is) slot in an old PC programs 8Kx8 EPROMs almost as fast as you can get them in and out of the ZIF socket. I would have another ZIF socket on the target board. My UV eraser holds 4 EPROMs or microcontrollers at a time, and takes about 16 minutes for a good solid erasure; so if I kept it going full time, I could take another EPROM out every four minutes, cycling them through, if indeed I could have the next image ready to program into it that soon.

The PIC microcontrollers I've put in many products use a serial programming method. When they were still EPROM-based, I used the above method. Now they're flash-based, so there's no need to remove them from the board to erase them before re-programming.

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

I am using AMD's 27C256-55 EPROM in POC, running at 12.5 MHz. My eraser is a cheap kitchen counter fluorescent fixture with a germicidal lamp in it. I can erase 18 EPROMs at a time with it. I even modernized it a while back with a built-in timer.


Unlike the SST39SF010A series, the AMD EPROM is a DIP, which is more hobby-friendly.

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

I've been using the Atmel 28C256 EEPROM and have programmed them hundreds of times already using a Dataman programmer. I also have byte write mode support to write data to it insitu on my 65C02 board. According to the doc, write cycles are pretty high, ~ 100,000 times.

_________________
Regards, KM
https://github.com/floobydust

The SST39SF0x0 series is (or was, a year or two ago) available in DIP for one, two, and four megabit (128k, 256k, and 512k) parts. Typically 70ns speed, IIRC.

28C256. In-place reprogramming is very nice, though I understand they are not the fastest ROMs around.

_________________
8 bit fun and games: https://www.aslak.net/

I have a decent quantity of Atmel 28C256 EEPROMs rated at 150ns, I've been running consistently at 10MHz without issue. Atmel also make these chips rated at 70ns which I have in PLCC-32 format (AT28HC256) but have not put them into use as of yet. Overall they seem to be under-rated on speed but as always you're better off not pushing them past published specs for reliability.

_________________
Regards, KM
https://github.com/floobydust

Which seems to have better performance speedwise? This is my first ROM, and am curious which is best.

_________________
It may look hard, but it won't take long if you take it one byte at a time.

AMD 55ns EPROMS are available from JAMECO and others. Atmel makes an OTP EPROM that is rated at 45ns. Most EEPROMs "top out" at 70ns.

The particular EPROM I use in POC is AMD's 27C256-55, JAMECO part number 266079. I have run POC as high as 15 MHz with that device.

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

I seem to have lost a draft post: don't forget NVRAM and FRAM as possibilities. No separate programming step, full speed reads and writes. FRAM is touted as having effectively unlimited capacity, usually spec'd as at least 10^14 or 10^15 accesses, with one manufacturer saying they'd never observed a wearout failure. NVRAMs have an internal cell with a life of 5 or 10 years, or connections for an external cell.

You get the best speed from SRAM. If you add a serial flash, you can copy the serial flash data to SRAM at reset, and then execute from there. Of course, you'll need a (small) parallel EPROM for booting, but that code can be relatively static and slow.

About the NVRAM and FRAM... to protect from malicious users (just to be safe) is there usually a write protect feature?

Edit: being the bonehead I can be sometimes, write protect comes in handy, especially when I mean to program $E000 (I/O) and actually type $F000 (ROM)

_________________
It may look hard, but it won't take long if you take it one byte at a time.