I'm not familiar with algorithmic steps in eprom
programming. Could this roughly be likened to a
It depends on the chip :-). If we ignore the really old devices for the
moment, and concentrate on the 2716 and later, there are 2 common algorithms.
The old, slow, one goes as follows :
1) Power up the chip
2) Apply Vpp. This is the programming voltage, and was 25V on old chips,
21 on slightly newer ones, and 12.5V on newer ones still. If you use too
high a voltage, you'll damage the chip. If you use too low, either it
won't program at all (which isn't too bad), or more worryingly, it'll
programm, but will loose data soon afterwards.
3) Select he location to be programemd using the address pins (normal TTL
levels)
4) Apply the data to be programmed into that location to the data pins,
again at TTL levels
5) Give a 'programming pulse' This is a 50ms TTL level pulse applied to a
particualr pin o nthe chip. Some devices have a special PGM input. On
others you apply it to the OE/ pin or something like that. The data sheet
should tell you.
From what I rememebr, this algorithm is the _only_
supported one on 2716
and 2732s, it's OK on 2764s and 27128s, but at least some
manufactuers
don't recoemnd it on larger chips
The other algorithm is called the 'intelegent algorithm' and is faster.
The first 4 steeps are as above, but instead of applying a 50ms
programming pulse, you apply a 1ms one, and then try to read the data
back. If it doesn't read back correctly, apply another 1ms pulse, and so
on, until either it verifies or you've given 50 pulses. In the later
case, the EPROM is bad (!). In the former case, you then apply 5 times as
many programming pulses as you already have given (I believe either as
sepearate 1ms pulese, or as one long pulse).
Some manufactuers suggest doing that verification of the programmed data
with the 5V power pin at some other voltage (often 6V). And that you
should do a final verification of the programmed data both with a high
supply voltage and a low one (4.5v) just to be sure.
IIRC, for some later EPROMs there was a way of reading out a
manufacture/device ID. It involved applying Vpp, and probably pushing the
5V lione up to 6V, and then doing odd things with the other pins (but at
TTL levels). This may be one some programmers refuse to program some
manfuacturer's devices, they don't recognise this ID (even though using
one of the 'geneirc' algorithms would work.
-tony