Yes, you are correct. I'm a bad counter. I knew a 2732 was a 24 but suffer from brain rot. Still it is likely a standard Intel part. At most it may have a pin swap on something like a select pin. Measuring the pins, will tell which pins are tied and which are active. Dwight ________________________________ From: cctalk <cctalk-bounces at classiccmp.org> on behalf of Paul Berger via cctalk <cctalk at classiccmp.org> Sent: Monday, March 21, 2022 5:19 PM To: dwight via cctalk <cctalk at classiccmp.org> Subject: Re: ID UV erasable PROMS used on an IBM PC board? They won't be 2732 as the EPROM in the picture has 28 pins, but 2732 is a 24 pin package 2764, 27128, 27256 and 27512 are all in 28 pin packages. The 8519602 does look like an IBM house number but unfortunately it is not listed in the cross reference I have . I would agree it should be fairly easy to figure out what they are by looking at the circuitry surrounding them. The unpopulated U3 position beside the EPROM looks like it might be for another EPROM and it looks to have pins 26,27,and 28 all tied together on everything larger than a 2764, pin 26 is A13 so my guess would be 2764. If you read a 2764 as a 27128 the upper and lower halves of the image will be identical because the A13 pin is not connected on a 2764 so you would effectively read the EPROM twice. Pin 26 tied high would eliminate anything larger than a 2764. Paul. On 2022-03-21 20:49, dwight via cctalk wrote:

On Mon, Mar 21, 2022 at 8:25 PM Chuck Guzis via cctalk <cctalk at classiccmp.org> wrote: Some device programmers can read the manufacturer and device ID codes check for a Intel standard part. Intel 2764 - 89h / 02h Intel 2764A - 89h / 08h Intel 27C64 - 89h / 07h Intel 27128 - 89h / 83h Intel 27128A - 89h / 89h Intel 27C128 - 89h / FCh Intel 27256 - 89h / 04h Intel 27C256 - 89h / 8Ch Intel 27512 - 89h / 0Dh Intel 27C512 - 89h / FDh

The only new information I can gather from those picture is it would appear that pins 1 and 28? of the EPROMs is tied together and since 28 is Vcc, connected to +5V this would eliminate 27512 as pin 1 on a 27512 is A15, but on 2764, 27128, and 27256 it is Vpp and would normally be tied to +5V for read.? As I mentioned previously the empty position has pin 26, 27, and 28 tied together, if the positions that are populated with sockets are the same that would eliminate 27128 and 27256 as well, and I would stand by my previous suggestion that it is 2764.? I would suggest that you check to see if 26 and 27 are connected to pin 28 on those sockets to verify. Paul. On 2022-03-22 07:03, D. Resor via cctalk wrote:

On Tue, Mar 22, 2022 at 11:56 AM Chris Elmquist <chrise at pobox.com> wrote: http://www.bitsavers.org/components/intel/_dataBooks/1993_Intel_Memory_Prod… Page 5-78 (Page 1201 of the PDF) Table 1. Mode Selection Intelligent Identifier - Manufacturer 89H Intelligent Identifier - Device 8DH That is for the A27C256 "Automotive" rated version. My BP Microsystems BP-1610 device programmer software doesn't have an entry for that one either.

On Tue, Mar 22, 2022 at 12:25 PM Chris Elmquist <chrise at pobox.com> wrote: (Flogging a dead horse here slightly) Just took a look at a previous version of that Intel databook: http://www.bitsavers.org/components/intel/_dataBooks/1991_Intel_Memory_Prod… Page 5-56 (Page 325 of the PDF) 27C256 Table 1. Mode Selection Intelligent Identifier - Manufacturer 89H Intelligent Identifier - Device 8DH NOTES: 4. Programming equipment may also refer to this device as the 27C256A. Older devices may have device ID = 8CH So apparently the Device ID change is not specific to the A27C256 "Automotive" rated version, just that in the 1993 version of the databook where I first looked the only EPROMs listed are the "Automotive" ones.

On Wed, 23 Mar 2022, Chris Elmquist wrote: I didn't even know that primitive EPROMs have device IDs... Without looking for a data book, how is the ID mechanism implemented? Christian

On Wednesday (03/23/2022 at 04:18PM +0100), Christian Corti via cctalk wrote: Intel called it "Intelligent Identifier(tm) Mode". From page 5-43 of my 1991 Intel Memory Products book, for the 27256, You put Vh (+12V) on address A9 and then, with A0=0, you will read out the manufacturer ID with A0=1, you will read out the device ID Chris -- Chris Elmquist

There are a few other problems. The power supply does not power up. Three of it's four voltages are missing (+5vdc, +8.5vdc, +13vdc). With no connections to the power supply board, the +5vdc measures 4.92vdc while the other two voltages are still zero volts. The second issue I found upon removing the driver and processor board is moisture had entered the typewriter's electronics and the connector between the two boards has corrosion on many of the pins and on the PWB near the connector. I do not have access to a microscope. I know it has been explained to me elsewhere that UV exposure from florescent lamps can erase the proms. The way I understand is the glass of the florescent lamp filters out any UV radiation (as does the piece of glass mounted in front of quartz halogen lighting). In order that a UV lamp can operate correctly the tube must be made of quartz, not glass. Am I wrong here? Don Resor From: dwight <dkelvey at hotmail.com> Sent: Monday, March 21, 2022 4:50 PM To: 'General Discussion: On-Topic and Off-Topic Posts' <cctalk at classiccmp.org>; D. Resor <organlists1 at sonic.net> Subject: Re: ID UV erasable PROMS used on an IBM PC board? The dies look to have consistent wire bonding. That would mean they are one of the standard EPROMs made by Intel, just preprogrammed by Intel before shipping. The numbers wouldn't be intel numbers they would be IBMs inventory numbers. My guess is that they are 2732s. You might use a microscope and look at the edges of the dies. They often have the die type in the metal layers around the edge some place. Avoid using florescent ring lights as a large amount of UVC leaks from these. White LED are or incandescent lights. Also power it up and note which pins look to have signals. if any of the lines have what looks like a constant voltage measure it to the nearest .01 volts if you can. That will help determine if it is driven by a signal or a hard tied wire. Lower left and upper right should be ground and power pins. Dwight _____ From: cctalk <cctalk-bounces at classiccmp.org> on behalf of D. Resor via cctalk <cctalk at classiccmp.org> Sent: Sunday, March 20, 2022 6:53 PM To: 'General Discussion: On-Topic and Off-Topic Posts' <cctalk at classiccmp.org> Subject: ID UV erasable PROMS used on an IBM PC board? I cannot find a datasheet by any of the numbers silkscreened on these ICs. Could these be proprietary IBM P/N numbers? https://www.dropbox.com/s/f6rvemx9ldbbv5x/EPROMS1.jpg?dl=0 No need for a Dropbox account, close the login pop up and you can view the image. Thanks Don Resor

It is true that the glass blocks most of the UVC from florescent lamps. The key word here is "most". It is not a 100% block. When looking at aged data in EPROMs one should error on the side of caution. As an example, I have a pole lamp that I use a standard florescent bulb, with the typical spiral of 3 turns. I used it for about 4 hours a day for about a month when I bumped into the lamp shade. It crumbled into pieces. At the base where the lamp shade was protected from direct light of the lamp, the plastic seemed to be of full strength, even though the temperature was higher at the base. I don't know the intensity of the various levels of UV at different frequencies leaking from the lamp. I do know that it takes a specific high frequency of UV in the UVC range as a minimum to erase EPROMs. This plastic may be sensitive to UV closer to the visible spectrum. I'm only saying that one should use a lamp that has zero UVC as apposed to a lamp that has a filter reduced level of UVC. Years ago, I won a bet by allowing an EPROM to sit exposed for 3 months in an industrial setting, using florescent tube lighting. It did not erase any of the data in that condition. I suspect it had some effect but unlikely much more than that lost by cosmic rays over the same period of time. Holding an aged EPROM about 2 inches from a florescent lamp is still not something that I'd bet on. Dwight ________________________________ From: D. Resor <organlists1 at sonic.net> Sent: Tuesday, March 22, 2022 2:40 AM To: 'dwight' <dkelvey at hotmail.com>; 'General Discussion: On-Topic and Off-Topic Posts' <cctalk at classiccmp.org> Subject: RE: ID UV erasable PROMS used on an IBM PC board? There are a few other problems. The power supply does not power up. Three of it?s four voltages are missing (+5vdc, +8.5vdc, +13vdc). With no connections to the power supply board, the +5vdc measures 4.92vdc while the other two voltages are still zero volts. The second issue I found upon removing the driver and processor board is moisture had entered the typewriter?s electronics and the connector between the two boards has corrosion on many of the pins and on the PWB near the connector. I do not have access to a microscope. I know it has been explained to me elsewhere that UV exposure from florescent lamps can erase the proms. The way I understand is the glass of the florescent lamp filters out any UV radiation (as does the piece of glass mounted in front of quartz halogen lighting). In order that a UV lamp can operate correctly the tube must be made of quartz, not glass. Am I wrong here? Don Resor From: dwight <dkelvey at hotmail.com> Sent: Monday, March 21, 2022 4:50 PM To: 'General Discussion: On-Topic and Off-Topic Posts' <cctalk at classiccmp.org>; D. Resor <organlists1 at sonic.net> Subject: Re: ID UV erasable PROMS used on an IBM PC board? The dies look to have consistent wire bonding. That would mean they are one of the standard EPROMs made by Intel, just preprogrammed by Intel before shipping. The numbers wouldn't be intel numbers they would be IBMs inventory numbers. My guess is that they are 2732s. You might use a microscope and look at the edges of the dies. They often have the die type in the metal layers around the edge some place. Avoid using florescent ring lights as a large amount of UVC leaks from these. White LED are or incandescent lights. Also power it up and note which pins look to have signals. if any of the lines have what looks like a constant voltage measure it to the nearest .01 volts if you can. That will help determine if it is driven by a signal or a hard tied wire. Lower left and upper right should be ground and power pins. Dwight ________________________________ From: cctalk <cctalk-bounces at classiccmp.org> on behalf of D. Resor via cctalk <cctalk at classiccmp.org> Sent: Sunday, March 20, 2022 6:53 PM To: 'General Discussion: On-Topic and Off-Topic Posts' <cctalk at classiccmp.org> Subject: ID UV erasable PROMS used on an IBM PC board? I cannot find a datasheet by any of the numbers silkscreened on these ICs. Could these be proprietary IBM P/N numbers? https://www.dropbox.com/s/f6rvemx9ldbbv5x/EPROMS1.jpg?dl=0 No need for a Dropbox account, close the login pop up and you can view the image. Thanks Don Resor

Years ago when I was at Intel, they would take EPROMs that had only a few defects and attempt to program them with specific applications. They would then usually paint the window black and put on that particular product number. When we needed EPROMs in the lab, we would take a pile of these that were no longer used and scrape the paint off the window. Some would take the desired program and some wouldn't. I suspect that they can fuse identify specific bad EPROMs to use for such harvesting. The fused value may be laser blown as well. My guess is that that is the reason for the different ID number. If it will take pre-programming with a particular data, it is worth harvesting. This way they can find value in an otherwise unsellable product. I doubt the harvesting methods have changed much over the years. Dwight ________________________________ From: cctalk <cctalk-bounces at classiccmp.org> on behalf of Paul Berger via cctalk <cctalk at classiccmp.org> Sent: Tuesday, March 22, 2022 1:14 PM To: D. Resor via cctalk <cctalk at classiccmp.org> Subject: Re: ID UV erasable PROMS used on an IBM PC board? I can tell you what some of the chips are on the power supply board, but the list I have is mostly TTL parts 1582684 LM324 1582606 74LS74 F/F D-TYPE DUAL +E/T 2392122 7417 BUF/DVR HEX O/C (TPD0 30NS) 8272147 74LS05 INV HEX O/C Aside from the house numbering of parts which is not unique to IBM, IBM had their own foundry for many years so you will encounter custom parts for which there is no industry equivalent, this is also not unique to IBM, I have some HP computers that many of the parts are house numbered, and HP also manufactured their own chips so there are some parts that are unobtanium as well. Paul. On 2022-03-22 06:40, D. Resor via cctalk wrote: