Identify mystery EPROMs

[Andrew Lynch]

Hi, I have some old EPROMs and need some help identifying them. They look to be AMD 2732s but when I put them in the programmer, I get an error message regarding Vpp.

I have some actual AM2732 EPROMs with earlier date codes and they program without issue. So even though they have an AMD logo, I don't think they are AM2732s.  These appear to be 1984 or 1985 vintage.

Thanks in advance

 

Sent from Yahoo Mail on Android

[horus falcon]

Hi, Andrew,

Well, they are the correct pin count for a 2732 EPROM.  Have you tried wiping the tops with IPA to see of they are fakes?  Where/how did you acquire these, and are you using a TL866 programmer? (Some of these will occasionally misidentify chips).

Later On,
David

[Alan Cox]

On Sun, 10 Mar 2024 at 21:18, horus falcon <horus...@gmail.com> wrote:
>
> Hi, Andrew,
>
> Well, they are the correct pin count for a 2732 EPROM. Have you tried wiping the tops with IPA to see of they are fakes? Where/how did you acquire these, and are you using a TL866 programmer? (Some of these will occasionally misidentify chips).

866II also can't program some 2732s because they need a pretty high
voltage on the Vpp and also because there is a bug in the 866II which
handles CS wrongly in some cases.

Don't recognize the parts. They don't look like any AMD2732 or 27C32
EPROM which makes me think they are probably genuine because I can't
see why a faker would make up a set of unique EPROM labels when they
could copy the known ones 8)

Alan

[horus falcon]

Ack!   I sent too quickly.  You might also check that the programmer meets these requirements (from the AMD Datasheet published by Jameco):

"The programming mode is entered when a voltage greater than 12.0, but less than 13.3 V is applied to the Vpp pin (OE-bar/Vpp for 32k)."   (Pin 20 of the DIP package).

Here's a link to the datasheet, in case you might need it.  It explores requirements for programming and reading these devices.
https://www.jameco.com/Jameco/Products/ProdDS/40133.pdf

Later On,
David

[David T]

They look fake. Some windows are different sizes and yet the data written on the package is the same.

--
You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/CAK9X0%2BsA45SXEt2ohd4jKhfhj15aUu%2BNNcUaGqpNBaUQrm67jQ%40mail.gmail.com.

[horus falcon]

Excellent points, all.  It's still worth doing the check if these were acquired from a Chinese source (e.g., alibaba) who does recovery/recycling of chips and other components.
I had forgotten about the bug for CS, but, yeah, the TL866 series are kind of limited.

Later On,
David

[andre...@gmail.com]

Is the part number not visible on the die?

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/CAOE-EuKZ%2BAdUc3sqGL1QkHNfR0T5O7NUF1vFALrm%2BREYkhe0RQ%40mail.gmail.com.

[horus falcon]

Yeah, it's a sad fact of life in the 21st century that fake electronics abound, especially in high-demand items.  Since EPROMS and their like are usually critical components, I try to source them from reputable vendors and observe their specs when programming.  I retired recently, and no longer have access to the Data I/O programmer at work, so will likely be buying some kind of programmer for use at home.  The TL866 series is the least expensive option I've found so far, but research is the key here.

Later On,
David

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/CAOE-EuKZ%2BAdUc3sqGL1QkHNfR0T5O7NUF1vFALrm%2BREYkhe0RQ%40mail.gmail.com.

[horus falcon]

I could see part numbers, and what look like date codes pinning these down to 1984 if they are genuine.  Note that the format of some of the labels present differently than others.  This may just be a difference from lot to lot, or it may be indication that some of these are fakes and some are genuine.  Man!  A Retro Chip Tester Pro would come in handy about now...

Later On,
David

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/070401da7332%24ad6f1bf0%24084d53d0%24%40gmail.com.

[Andrew Lynch]

Hi

Yes, I did clean the tops with GooGone to dissolve the sticker residue and the labels are on there for good.

I acquired these EPROMs several years ago as part of a large batch of components.  I think it was from a small-time reseller who retired.

I am using a TL866II T48 which is the upgraded model and I know it handles 25Vpp 2732 EPROM because I successfully programmed a small batch of various kinds.  However, you are right about the TL866II+ because it cannot handle the really old 2732s with 25V or 21V Vpps.  It is why I upgraded to the T48 programmer.

Unfortunately, the 2732s are from before EPROMs used Device IDs.  If I could read a Device ID, this would be trivial to identify since you can backtrack from a Device ID to a manufacturer and model.  I think the first EPROMs to support Device ID were the 2764s.

Given the apparent date codes and that they threw a Vpp error trying to program as an AM2732, I am going to try as an AM2732B since they are about the right vintage for 12.5V Vpp EPROMs.

Thanks, Andrew Lynch

--
You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.
To view this discussion on the web visit

https://groups.google.com/d/msgid/retro-comp/60d551bd-02f5-47ba-829a-a6b4821f0af6%40gmail.com

.

[Andrew Lynch]

Hi

No, the closest to a part number is "2732-25" label on the top of the chip.  I am inferring AMD as a manufacturer because of the AMD logo on the botttom.  They are missing the usual label like "AM2732" or equivalent.

Thanks, Andrew Lynch

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/070401da7332%24ad6f1bf0%24084d53d0%24%40gmail.com.

[Andrew Lynch]

Hi

I really don't think they are fakes because I've had them for at least 10 years.  I think they are legitimate and act/read like a normal 2732 EPROM.  I suspect they are OEM specialty parts probably for a manufacturer.  Possibly automotive or some other equipment.  

I just need to find out what they are closest too so I can reprogram them.  At least in the T48, there are only 3 options for AMD 2732 series NMOS EPROMs: 2732, 2732A, and 2732B.

Chip manufacturers used to make common parts with in-house part numbers for large manufacturers.  You'd see it even in common logic TTL 74xx type chips that would make it into the grey market.

Thanks, Andrew Lynch

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/ebec2db4-30da-4a01-a4b8-7bed008c0553%40gmail.com.

[andre...@gmail.com]

On the AMD EPROMs I have seen, AMD put their logo, a block text  “AMD”, the part number, and a copyright date in a text area on the die.  Sometimes that is the only surefire way to know what programming voltage to use.

[Andrew Lynch]

Hi

Yes, these EPROMs are "special" which is what makes them a mystery.  On virtually all of the other EPROMs I've dealt with, identifying them is trivial and practically instant.  Just read the label.  But not these for some unknown reason.

I have a quantity of the "usual" AMD EPROMs and they are exactly as you describe.  Super easy to identify even without Device ID.

Thanks, Andrew Lynch

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/073201da7336%24756ff270%24604fd750%24%40gmail.com.

[Andrew Lynch]

Hi

Good news!  I pulled out a sample of the EPROMs and they program successfully as AM2732As.  I tried programming them as AM2732Bs and failed.  Trying as AM2732s failed too so it only left the AM2732As and thankfully, it worked.  

They verified successfully and seem to work fine.  I've separated these into their own tube and labeled accordingly so I think the mystery is solved.  

Thanks everyone for helping out!

[Douglas Miller]

I would think they are 2732, as the label on top says. They are probably made for an OEM that requested a special part number, like the "733WO1589". Just a guess, though. Strange, if that is the case, that there is also the "2732-25" - although that label looks as if it was added later/separately.

[andre...@gmail.com]

Maybe it is because I have spent a lifetime buying the cheapest EPROMs I can find, but I have come across a reasonable number where the marking on the case does not match the die.  Sometimes ( as you mention) pre-programmed OEM specials that were re-marked (or supplemental marked) and resold, sometimes factory-fresh mismark screw-ups, sometimes just fakes.

If I get a newly acquired EPROM and it doesn’t program, the first thing I do is look through the window to read the device number right off the die.   The die never lies.

 

 

From: retro...@googlegroups.com <retro...@googlegroups.com> On Behalf Of Douglas Miller
Sent: Sunday, March 10, 2024 3:37 PM
To: retro-comp <retro...@googlegroups.com>
Subject: Re: [retro-comp] Identify mystery EPROMs

 

I would think they are 2732, as the label on top says. They are probably made for an OEM that requested a special part number, like the "733WO1589". Just a guess, though. Strange, if that is the case, that there is also the "2732-25" - although that label looks as if it was added later/separately.

--

You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/905f8793-5554-402d-8352-1c66c9ce45e0n%40googlegroups.com.

[Amardeep Chana]

Hi Andrew,

I just went through that with some 2716s I ordered from Ali Express (yeah...I know).  The little post below the die in the window resembles parts made by SGS/ST.  You might try programming them as an ST M2732 (or more likely M2732A due to the die size).  I used IPA and some of their original labels started showing through under the ceramic sputter paint the remarker used to hide the original labels.

Best of luck!

Regards,

Amardeep

[Andrew Lynch]

Hi

Thanks!  Fake parts are the bane of my existence.  They are ruining this hobby because the fakes are appearing everywhere.  No vendor is safe from high quality fakery.

Going through the whole fake parts ordeal trying to order a legitimate V9958.  The market is flooded with fakes now.  What a pain to deal with.

https://hackaday.io/project/186450/log/216999-lots-and-lots-and-lots-of-fake-v9958

Thanks, Andrew Lynch

--

You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/363a7f1b-6235-4d90-ae56-6b31c2da9697n%40googlegroups.com.

[Amardeep Chana]

Oh, I had missed where you found it worked with the A suffix...makes sense.  2732 had a much larger visible die area than 2732A like these.  Also, just for posterity, the 733W..... number is Xerox part numbering when the build authorize vendors for their BOMs.

[Andrew Lynch]

Hi

Interesting about the 733W part number.  I suspected something like that.  Maybe a contract number or similar but a vendor authorization makes sense too.  Thank you!

Andrew Lynch

--

You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.
To view this discussion on the web visit

https://groups.google.com/d/msgid/retro-comp/02158454-a20c-450c-a26f-5200a43476adn%40googlegroups.com

.

[horus falcon]

AMD sometimes private labels parts, and a lot of these are labeled for military use.  That may explain the lack of "AM" in the part number?

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/073201da7336%24756ff270%24604fd750%24%40gmail.com.

[Andrew Lynch]

Hi

OK, that last sentence really got my attention.  You can read text on the actual die itself?  Are you using a microscope?  I can't see anything.  My vision is not that great, but I see nothing at all.  Text on the die has to be really itty-bitty tiny to fit on that speck of silicon.

Thanks, Andrew Lynch

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/08fb01da7352%241334c970%24399e5c50%24%40gmail.com.

[Wayne Hortensius]

On Tue, 12 Mar 2024 01:01:39 +0000 (UTC)
"'Andrew Lynch' via retro-comp" <retro...@googlegroups.com> wrote:

> HiOK, that last sentence really got my attention.  You can read text

> on the actual die itself?  Are you using a microscope?  I can't see
> anything.  My vision is not that great, but I see nothing at all.
> Text on the die has to be really itty-bitty tiny to fit on that speck
> of silicon. Thanks, Andrew Lynch

I just had to try this too. :) Using a plain vanilla digital camera in
close-up mode, I took pictures of about a dozen or so different EPROMs I
had here. I couldn't find any text in most of them; some I'm pretty
sure didn't the die have any text visible at the magnification I could
manage, in some the focus or lighting didn't work out, but I did get a
couple of very unambiguous "yep, there it is" images and I've included
the best one here.

Regards,
Wayne

[Andrew Lynch]

Oh Wow.  My mind is blown.  I never knew you could do that.

Thanks!

[Bill Shen]

Wow, that's a very clear example.  I've looked at my collection of EPROM and most of them are too fine to read clearly.  In most case the trademark symbols are easier to identify so at least that can be matched with the labeling on the package.
  Bill

[Tom Storey]

If you're lucky enough you might even come across some "chip art". Google it. 🙂

--
You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/1348460277.2369406.1710207059972%40mail.yahoo.com.

[Michael Wheepy Crousaz]

Hi everyone,

IDK it this may be usefull, but searching other infos about 2732 EPROMs, I found out those types may have 3 versions (among various brands and other parameters)

2732, 2732A, 2732B ... And the differences between those 3 is the programming voltage !

Chip - Vpp
2732 - 25v
2732a - 21v
2732b - 12.5v

Knowing yours show 2732-25, I suspect this 25 it the writting voltage.

Source : https://www.thirdgen.org/forums/diy-prom/37598-eprom-differences.html

Cheers,

Michael

[T Gerbic]

Not sure why it seems so mysterious. They just look like preprogrammed original version AMD 2732s that have a speed rating of 250ns. I believe I have some of these marked 2732-25 in my EPROM drawer with preprogrammed numbers on them. Once erased, they are just original version AMD 2732s. I can program them on my Unisite, B&C or Xeltek 610P. These all have selectable programming voltage, either manually set of read from the actual part number.

In the old days many companies could not afford expensive mass programmers so just had them delivered preprogrammed.  If the code was considered ready for production and they needed thousands, many companies skipped the EPROMs and just had them mask programmed into permanent ROMs to greatly cut cost.

As Michael posted above, different versions of the same EPROM are often available with different programming voltages. Generally a number like 25 indicates speed, and this would be 250ns. Usually EPROMs and older RAM/ROM chips would be marked 1 for 100ns, 2 for 200ns, 25 for 250ns and 35 for 350ns access time.

This has been my experience over the decades.