View Full Version : OT: Making backup copies of Eproms

Billy T
06-07-2003, 04:09 PM
Hi Team

I am thinking of buying some obsolete but very specialised second-hand test equipment that has its operational sequences programmed via two Eproms. I can buy additional units for spares but only one still has its Eproms, the others having been lost.

Before I decide to buy, I need to be sure I can make spare copies of the existing Eproms in case of failure. This is outside my area of expertise and I would like to know if Eprom duplication services exist. In my dreams I would be able to plug an Eprom into a duplicating tool and program a blank from the data read off the first.

Can somebody either wake me up (please be gentle) or tell me how I can make copies or get backups made.


Billy 8-{)

[b][pre]BTW- This is not about piracy of games or copyright
infringement in any way. It is a practical problem of
protecting an investment by ensuring the equipment
will remain operational if a program chip fails.

Stumped Badly
06-07-2003, 04:22 PM
Did a google on eprom & selected within NZ & it came up with a whole stack of sites.
One in particular may interest you

06-07-2003, 08:29 PM
The site does not like the auction number????

Stumped Badly
06-07-2003, 08:49 PM
eprom nz
into google & make sure you check the radio button beside pages from New Zealand.
It is the 2nd entry on page 2.
There are plenty of related sites, I just thought the phone number & email addy on the page might be handy as he was asking for info & may be able to help.

06-07-2003, 09:23 PM
Eprom programmers are still available to DIY.

Like this (http://www.dse.co.nz/cgi-bin/dse.storefront/3f07e8a90845e720273fc0a87f9906bc/Product/View/K3605)

You need to check that the EPROMs are still available blank, and that any unit (or service) can actually handle that type of chip. I suspect some earlier ones are like hens teeth now.

Billy T
07-07-2003, 02:01 PM
Hi godfather

Your link didn't work, but is this the unit? (http://www.dse.co.nz/cgi-bin/dse.storefront/3f08d06b08ee7764273fc0a87f990717/Product/View/K3605) And am I correct in believing it will read the code from an existing eprom and copy it to a new chip? I assume the code is first written to a computer file then written back.


Billy 8-{)

07-07-2003, 02:23 PM
Also keep in mind that some eproms were "masked" which didn't allow the program to be extracted from them. Mind you, they are not that common and had to be manufactured like that.
Used to have an e-prom programmer and you can upload/download from PC fine, so can be done no problems.
Must add my voice to GF in that you need to ensure you can get the blank eproms and that the programmer can handle that type.

Graham L
07-07-2003, 03:13 PM
I think GF will concur with my comments. :D That unit will do it ... just check that it will handle the size you need.

The Eproms are available ... and cheaper than they used to be. I still remember that the first one I bought was 2kx8 and cost me $84. Now 256k ones (that's 32kx8) cost much less than that.

DSE might still have them, or Jaycar. South Island Components (http://sicom.co.nz/) do, even 2708! They have up to 4Mbit ... (512kx8). RS will certainly have them at a price ;-). However, there is a source of cheap EPROMs: old PCs. They used to have erasibles for the BIOS. Its only in the last days before they started using flash memory that they used windowless EPROMs as OTP (one time programmable). So early motherboards are a fruitful source of good 25xxx UV erasibles. Just make sure that there's a quartz window under the label on the BIOS chip.

The tricky bit might be erasing them. You need fairly intense UV (which the neighbourhood fusion reactor is supplying "not much of" at the moment. You might think it's worth buying new "blank" ones.,at $12-$24. By the way, the blank state is all "1"s. The propramming clears bits which have to be off. :D

The programmer just uses the chip as a device. It can read any byte, by setting the address lines , and can write by setting the data lines and pulsing the programme pin to the critical voltage (used to be 25V, but now mostly 12.5V). The ones I have salvaged have usually had Vpgm marked on them.

The programmer will copy an existing unit to a file, which can then be used for duplication. The first computer I built ( a Z80) had an EPROM programmer and a hex keypad and 7segment displays. I finished up with a full 2kx8 with a new monitor programme, a basic interpreter, a Lisp interpreter, and keyboard and full s100 video software in it. All hand coded in assembly. :D And entered byte by byte on a keypad. :-(

There are various algorithms used to minimize the time taken --- usually instead of holding the programme voltage high for the specified 50 ms per location, they use a series of shorter pulses, checking after each one to see whether it has "taken". It wasn't too bad with 2k locations ... with 512k production was a bit slow if the full 50 ms was taken for each.

Billy T
07-07-2003, 04:05 PM
Okay guys, thanks for the help so far, I've checked the equipment specs and here's some details:

The existing eproms are Fuji MBM2764-30Z NMOS (28 pin?) and I see that generic 2764 eproms are still reasonably available. I also see that some types are listed as 27C64 but I don't know what the difference is, if any.

There seem to be two different speed ratings too, 200nS and 250nS so I guess the faster speed would be the safest. The equipment is about 6-8 years old but the CPU is a Z80 and the design goes back to the mid 1980's so slower chip speed would probably have been the go.

I am well outside my comfort zone here as I have never played with Eproms before, so I would appreciate some reassurance as to likely interchangeability with the OEM Fuji type. There is considerable investment involved and supply is on a no-return basis, though on-sale is always a possibility to others wanting spare parts.


Billy 8-{)

Graham L
07-07-2003, 04:32 PM
Those Fujitsus will be 300 ns speed ... 27C64 will be perfectly OK and pin compatible. They are CMOS, rather than bipolar. For a while there was a problem with 27xx and 25xx which was to do with power supplies. At least they had fixed NMOS to not need +5, -5 and 12 volt supplies by the time they got to 64k (8kx8). I think some 2716 needed multiple supplies, whereas 2516 was single 5V. At the bigger sizes, the shortage of pins made them fix it. :D

I think the fastest Z80s (made as 40 pin Z80s) were 6 (or 8) MHz. The chip layout is still in use as the core of hundred-lead-latpacks used in disk drives, disk controllers, SCSI controllers ... etc -- it's a magnificent design. Alan Sugar's last project in Amstrad was a brilliant (too late) word processor with a 20MHz Z80 core.

The 2 MHz Z0s I used worked well with 450 ns EPROM, and RAM :D (I had two boards one of 4kx8 and 1 of 8kx8).

You might find a second-hand programmer ... there must be hundreds around. The DSE kit (designed by EA or SC) will have adjustable programming voltage. The older kits will have limited address range (the lastest allow up to 1 Mbit ... though there are now 4 Mbit EPROMS) but will certainly do 64k.

You should be able to find the specifications of the EPROMs through google.

Billy T
07-07-2003, 05:51 PM
Thanks Graham,

I think I can move forward on that now. One last question though, I noticed that the DSE programmer has a 32 pin socket so may I assume it can be configured for 28 pin as well?


Billy 8-{)

Graham L
08-07-2003, 04:05 PM
They changed to a 32 pin socket when they changed the design for 1 Mbit chips. You put the chip in at the appropriate end. :D There will be instructions. You will have seen some RAM (and cache) sockets which had "spare" pins if you used the lower capacity chips.