ROM identification and duplication?

[James Meyer]

On Sat, 27 Dec 2003 10:04:53 +0100, Dieter Wiedmann
<Dieter.Wiedmann@t-online.de> wroth:

James Meyer schrieb:

I've erased many OTP devices. I put them close to the electron beam
dump of a 3 MEV linear electron accelerator. The X-rays generated there would
be fatal in just a short time. The facility would cost about 10 megabucks to
duplicate.

And afterwards you will nevermore reach a sufficing data storage time.
Done that with a dentists X-ray, first bit faults occured after less
than a year.


Regards,
Dieter

Dentist's X-ray machines are just barely able to erase an OTP part. Are
you sure you damaged the parts that way? Or did you only marginally erase them.
If all the faults were zeros that magically became ones or vice versa, then
marginally erased is my guess.

Jim

 

[DaveC]

How can I tell what configuration this 40-pin DIP ROM is? Specs of printer
say 4 MB ROM, but that could be 256k x 16 or 512k x 8, right?

Which means the correct EPROM is either 27C0400 (configurable as 256k x 16 or
128k x 8??) or 27C4096 (512k x 8).

How to tell?

Thanks,
--
DaveC
me@privacy.net
This is an invalid return address
Please reply in the news group

 

[Bruce Esquibel]

DaveC (me@privacy.net) wrote:
: How can I tell what configuration this 40-pin DIP ROM is? Specs of printer
: say 4 MB ROM, but that could be 256k x 16 or 512k x 8, right?

: Which means the correct EPROM is either 27C0400 (configurable as 256k x 16 or
: 128k x 8??) or 27C4096 (512k x 8).

: How to tell?

Trial and error.

What is a 27C0400 ? Can't seem to find that anywhere.

Anyway, I suppose the big question is what exactly are you going to copy the
rom with. Some of the programmers like the Dataman stuff have a signature
read, just stick the rom in and it'll tell you what it is, provided it has a
signature to read, most modern day eproms do.

If it's some home brew setup, you'll just have to do more work narrowing the
device down. At least figure out the grounds and power supply connections.
Like with the datasheet for the 27C4100 (256kx16) and the 27C4096 (512kx8),
the 4100 has vcc on pin 21, the 4096 has it on pin 40. Both use 11 and 30 as
grounds.

The 4096 seems to be a easy device to spot, all the address lines are on one
side of the chip (21 to 29 and 31 to 39). On the 4100 they are on both sides
(1 to 9 and 32 to 40). Generally the address lines are all bundled together
on the foil traces into another device. Doesn't have to be perfect, "close
enough for government work" is ok.

Once you can narrow down the likely suspects just try them, provided the vcc
and grounds match, it's unlikely you are going to hurt anything if you
picked the wrong one.

I'm hoping you are at least dealing with something in a socket, if the rom
is soldered directly into the board, maybe figuring out the eprom is going
to be the least of your worries.

-bruce
bje@ripco.com

 

[Dieter Wiedmann]

James Meyer schrieb:

Dentist's X-ray machines are just barely able to erase an OTP part. Are
you sure you damaged the parts that way? Or did you only marginally erase them.
If all the faults were zeros that magically became ones or vice versa, then
marginally erased is my guess.

Surely not marginally earased, the bits went to logic '1'.


Regards,
Dieter

 

[Lewin A.R.W. Edwards]

Surely not marginally earased, the bits went to logic '1'.

But under this digital behavior lie analog characteristics. We just
defined some threshold point and said "less than threshold = 0, more
than threshold = 1". A digital input is after all nothing more than a
high-gain voltage amplifier. The essence of a reliable and predictable
digital circuit is keeping all input states well away from the
threshold point where analog behaviors are observable.

Using the dentist's X-ray may have set each cell close enough to the
"Erased" state that it read all erased under some set of conditions.
But later on, with some different Vcc, after time, who knows?