Easy to make EPROM copier

[Studio263]

I made this EPROM programmer as an easy to use tool for copying the software used in B&O TV and hi-fi products made in the late 1980s and early 1990s. This is contained on 28 pin 27-512 type EPROMs, most of which are now unavailable with the correct software on. The plan was to read off the original software and to write it onto new EPROMs as a means of storage, this being a better long term proposition that computer-based techniques which would rely on the long term availability of hardware which is sold with a large degree of built-in obsolescence – not a good idea. The EPROMs on the other hand should last at least 20 years and can be re-copied at any time using only the programmer itself, which is completely self-contained. Timings and parameters are optimised for AT27C512R EPROMs, available from RS components under the part number 696-3017. These are write-once devices, chosen for their natural UV protection and low cost (about £2 each).

It is based on a single PIC16F877A microcontroller which generates the 16 bit addresses to page through the EPROM, processes each byte of data and generates the required 100uS programming pulse timed by its own quartz crystal reference. The PIC also controls an on-board linear power supply which generates 5V, 6.5V and 12.5V supplies as required to perform the reading and writing operations. The process of operation is this:

1) A 16 bit address is generated (the sequence runs from 0 to xFFFF).
2) The byte addressed is read from the source EPROM (U5 in the diagram) nd stored in the memory of the PIC
3) The byte is then written to the same location in the destination EPROM (U4 in the diagram)
4) The source EPROM is then switched to “read” mode and the stored byte is read back into the PIC
5) The two stored bytes are compared, if they match the address is incremented and the next location is addressed. The process continues until all locations have been copied, at the end LED D2 lights.
6) If the two stored bytes don’t match the writing process is repeated. This re-try routine can be repeated up to 10 times per location, if the two bytes still don’t match the copying process is stopped and LED D3 lights.

At the end, if D2 is lit then the destination EPROM contains an accurate and verified copy of the source EPROM.

The odd resistor values in the diagram are there because the regulators need them to get the correct output voltages. Since RS sell resistors in lots of 25 it was logical to use the rest up in parts of the circuit were the exact value isn’t critical. Every effort has been made to protect the source EPROM; this is why the HEF40373B latch is use. In theory the EPROM’s on-board tri-state outputs could have been used but the latch provides an extra barrier in case the programming voltages get onto the data bus. The resistors at the other pins are there for the same reason, they allow the internal protection diodes to clamp the pins should an excess of voltage appear.

D1 follows the LSB of the upper byte of the address bus, meaning that it flashes slowly during copying. The copying process has been conservatively timed at each step to ensure accuracy, it takes about two minutes to complete. The circuit runs from a single 15V supply, copying starts as soon as this is applied.

I have a B&O MX4000 TV running in my workshop which contains an EPROM copied using this circuit, it functions identically to the way it does with the original chip.

PM me if you would like to build one of these and need the software.

[Dave Moll]

In step 4, do you mean that the "destination" EPROM is switched to read mode?

[Studio263]

Yes, that's right.

[SiriusHardware]

Since this project is primarily for rescuing / refreshing eproms used in antique equipment, I would suggest incorporating Tony Duell's idea from elsewhere in these forums. Read each byte from the source eprom many times and if there are any bits in the byte which sometimes read 0 and sometimes read 1, assume a value of 0 for those bits.

This will help to get more reliable readings from old source eproms which may be hovering on the verge of 'bit-rot'

Also changing the supply voltage too can help in getting out the last bit of stored data (I think it is a bit lower to suck out that last smidgen).

[Mikey405]

Well done Tim. I'd like to join the others in congratulating you on a pretty clever bit of kit. I'm quite sure that it will be very useful for safeguarding the software on older colour sets of a certain age. An excellent effort. A+

Kind regards.

From Mike.