1. Forum
  2. Usenet
  3. COMP.SYS.IBM.PS2.HARDWARE

  • I see... dead people. Er, Patterns

    From Tomas Slavotinek@21:1/5 to Louis Ohland on Mon Aug 30 20:58:14 2021
    I'm positive that these 4 chips are all different in some way and the
    patterns are very much intentional - layout reasons. Different P/Ns for
    what's exactly the same part are not all that uncommon, but *not* on the
    same board.

    (see "[Update] Model 80 Memory Cards" -> "Topology Thoughts")

    But eh, at this point I'd say it's time to probulate rather than
    speculate...

    On 30.8.2021 20:38, Louis Ohland wrote:
    https://www.ardent-tool.com/8580/Memory_IBM_2MB_1_Photo_Front.jpg

    Now this seems to suggest a pattern. If the P/Ns were just due to
    accounting alone, why the Tic-Tac-Toe? Note the wider spacing down the middle, resulting in a 3 wide by 4 high grouping.

    Low density sugar cubes?

      96-pin connector [top]

    90X7010 90X7012 90X7010    90X7012 90X7010 90X7012
       X       0       X          0       X       0
    90X7012 90X7010 90X7012    90X7010 90X7012 90X7010
       0       X       0          X       0       X
    90X7011 90X7013 90X7011    90X7013 90X7011 90X7013
       A       B       A          B       A       B
    90X7013 90X7011 90X7013    90X7011 90X7013 90X7011
       B       A       B          A       B       A

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tomas Slavotinek@21:1/5 to Tomas Slavotinek on Mon Aug 30 21:00:04 2021
    Originally I wanted to color the backgrounds of the chips differently,
    to make the patterns easier to spot, but it looked gimmicky. Eh, I'm
    gonna try again...

    On 30.8.2021 20:58, Tomas Slavotinek wrote:
    I'm positive that these 4 chips are all different in some way and the patterns are very much intentional - layout reasons. Different P/Ns for what's exactly the same part are not all that uncommon, but *not* on the
    same board.

    (see "[Update] Model 80 Memory Cards" -> "Topology Thoughts")

    But eh, at this point I'd say it's time to probulate rather than
    speculate...

    On 30.8.2021 20:38, Louis Ohland wrote:
    https://www.ardent-tool.com/8580/Memory_IBM_2MB_1_Photo_Front.jpg

    Now this seems to suggest a pattern. If the P/Ns were just due to
    accounting alone, why the Tic-Tac-Toe? Note the wider spacing down the
    middle, resulting in a 3 wide by 4 high grouping.

    Low density sugar cubes?

      96-pin connector [top]

    90X7010 90X7012 90X7010    90X7012 90X7010 90X7012
       X       0       X          0       X       0
    90X7012 90X7010 90X7012    90X7010 90X7012 90X7010
       0       X       0          X       0       X
    90X7011 90X7013 90X7011    90X7013 90X7011 90X7013
       A       B       A          B       A       B
    90X7013 90X7011 90X7013    90X7011 90X7013 90X7011
       B       A       B          A       B       A


    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Louis Ohland@21:1/5 to All on Mon Aug 30 13:38:56 2021
    https://www.ardent-tool.com/8580/Memory_IBM_2MB_1_Photo_Front.jpg

    Now this seems to suggest a pattern. If the P/Ns were just due to
    accounting alone, why the Tic-Tac-Toe? Note the wider spacing down the
    middle, resulting in a 3 wide by 4 high grouping.

    Low density sugar cubes?

    96-pin connector [top]

    90X7010 90X7012 90X7010 90X7012 90X7010 90X7012
    X 0 X 0 X 0
    90X7012 90X7010 90X7012 90X7010 90X7012 90X7010
    0 X 0 X 0 X
    90X7011 90X7013 90X7011 90X7013 90X7011 90X7013
    A B A B A B
    90X7013 90X7011 90X7013 90X7011 90X7013 90X7011
    B A B A B A

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Louis Ohland@21:1/5 to Tomas Slavotinek on Mon Aug 30 14:11:19 2021
    Or the he-man method of using a legend? So old school, but it might work...

    On 8/30/2021 14:00, Tomas Slavotinek wrote:
    Originally I wanted to color the backgrounds of the chips differently,
    to make the patterns easier to spot, but it looked gimmicky. Eh, I'm
    gonna try again...

    On 30.8.2021 20:58, Tomas Slavotinek wrote:
    I'm positive that these 4 chips are all different in some way and the
    patterns are very much intentional - layout reasons. Different P/Ns for
    what's exactly the same part are not all that uncommon, but *not* on the
    same board.

    (see "[Update] Model 80 Memory Cards" -> "Topology Thoughts")

    But eh, at this point I'd say it's time to probulate rather than
    speculate...

    On 30.8.2021 20:38, Louis Ohland wrote:
    https://www.ardent-tool.com/8580/Memory_IBM_2MB_1_Photo_Front.jpg

    Now this seems to suggest a pattern. If the P/Ns were just due to
    accounting alone, why the Tic-Tac-Toe? Note the wider spacing down the
    middle, resulting in a 3 wide by 4 high grouping.

    Low density sugar cubes?

      96-pin connector [top]

    90X7010 90X7012 90X7010    90X7012 90X7010 90X7012
       X       0       X          0       X       0
    90X7012 90X7010 90X7012    90X7010 90X7012 90X7010
       0       X       0          X       0       X
    90X7011 90X7013 90X7011    90X7013 90X7011 90X7013
       A       B       A          B       A       B
    90X7013 90X7011 90X7013    90X7011 90X7013 90X7011
       B       A       B          A       B       A



    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tomas Slavotinek@21:1/5 to Louis Ohland on Mon Aug 30 21:54:25 2021
    On 30.8.2021 21:11, Louis Ohland wrote:
    Or the he-man method of using a legend? So old school, but it might work...

    On 8/30/2021 14:00, Tomas Slavotinek wrote:
    Originally I wanted to color the backgrounds of the chips differently,
    to make the patterns easier to spot, but it looked gimmicky. Eh, I'm
    gonna try again...

    https://ardent-tool.com/8580/Planar_Memory.html#Cards

    It's like a fookin xmas tree :-D

    What do you think?

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Louis Ohland@21:1/5 to Tomas Slavotinek on Mon Aug 30 15:06:16 2021
    Eh, should do.

    On 8/30/2021 14:54, Tomas Slavotinek wrote:
    On 30.8.2021 21:11, Louis Ohland wrote:
    Or the he-man method of using a legend? So old school, but it might work... >>
    On 8/30/2021 14:00, Tomas Slavotinek wrote:
    Originally I wanted to color the backgrounds of the chips differently,
    to make the patterns easier to spot, but it looked gimmicky. Eh, I'm
    gonna try again...

    https://ardent-tool.com/8580/Planar_Memory.html#Cards

    It's like a fookin xmas tree :-D

    What do you think?


    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tomas Slavotinek@21:1/5 to Louis Ohland on Mon Aug 30 22:40:36 2021
    OK, I'll keep it then (maybe I'll adjust the colors a lil later on...).

    On 30.8.2021 22:06, Louis Ohland wrote:
    Eh, should do.

    On 8/30/2021 14:54, Tomas Slavotinek wrote:
    On 30.8.2021 21:11, Louis Ohland wrote:
    Or the he-man method of using a legend? So old school, but it might
    work...

    On 8/30/2021 14:00, Tomas Slavotinek wrote:
    Originally I wanted to color the backgrounds of the chips differently, >>>> to make the patterns easier to spot, but it looked gimmicky. Eh, I'm
    gonna try again...

    https://ardent-tool.com/8580/Planar_Memory.html#Cards

    It's like a fookin xmas tree :-D

    What do you think?


    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Louis Ohland@21:1/5 to Tomas Slavotinek on Mon Aug 30 16:14:49 2021
    Huh. Maybe there is a difference, pinout-wise. Look at the patterns,
    remember the CAS, RAS, and addressing thing? Maybe IBM designed these
    cards to reduce the connector count.

    I dunno.

    On 8/30/2021 15:40, Tomas Slavotinek wrote:
    OK, I'll keep it then (maybe I'll adjust the colors a lil later on...).

    On 30.8.2021 22:06, Louis Ohland wrote:
    Eh, should do.

    On 8/30/2021 14:54, Tomas Slavotinek wrote:
    On 30.8.2021 21:11, Louis Ohland wrote:
    Or the he-man method of using a legend? So old school, but it might
    work...

    On 8/30/2021 14:00, Tomas Slavotinek wrote:
    Originally I wanted to color the backgrounds of the chips differently, >>>>> to make the patterns easier to spot, but it looked gimmicky. Eh, I'm >>>>> gonna try again...

    https://ardent-tool.com/8580/Planar_Memory.html#Cards

    It's like a fookin xmas tree :-D

    What do you think?


    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Louis Ohland@21:1/5 to Louis Ohland on Mon Aug 30 16:15:51 2021
    Oops, conductor / trace count...

    On 8/30/2021 16:14, Louis Ohland wrote:
    Huh. Maybe there is a difference, pinout-wise. Look at the patterns,
    remember the CAS, RAS, and addressing thing? Maybe IBM designed these
    cards to reduce the connector count.

    I dunno.

    On 8/30/2021 15:40, Tomas Slavotinek wrote:
    OK, I'll keep it then (maybe I'll adjust the colors a lil later on...).

    On 30.8.2021 22:06, Louis Ohland wrote:
    Eh, should do.

    On 8/30/2021 14:54, Tomas Slavotinek wrote:
    On 30.8.2021 21:11, Louis Ohland wrote:
    Or the he-man method of using a legend? So old school, but it might
    work...

    On 8/30/2021 14:00, Tomas Slavotinek wrote:
    Originally I wanted to color the backgrounds of the chips
    differently,
    to make the patterns easier to spot, but it looked gimmicky. Eh, I'm >>>>>> gonna try again...

    https://ardent-tool.com/8580/Planar_Memory.html#Cards

    It's like a fookin xmas tree :-D

    What do you think?



    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tomas Slavotinek@21:1/5 to Louis Ohland on Mon Aug 30 23:42:34 2021
    The chips were probably designed with applications like this in mind,
    possibly down to the pin assignments and recommended layout patterns. We
    need more info about these lil cubes, and find out where else were they
    used, if anywhere. (Perhaps the early 30-pin SIMMs?)

    Ultimately we may have to wait for Ken Shirriff's analysis (if he
    decides to check it out).

    On 30.8.2021 23:15, Louis Ohland wrote:
    Oops, conductor / trace count...

    On 8/30/2021 16:14, Louis Ohland wrote:
    Huh. Maybe there is a difference, pinout-wise. Look at the patterns,
    remember the CAS, RAS, and addressing thing? Maybe IBM designed these
    cards to reduce the connector count.

    I dunno.


    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tomas Slavotinek@21:1/5 to Tomas Slavotinek on Mon Aug 30 23:53:34 2021
    UMMR has some tasty DRAM cubes here:

    http://ummr.altervista.org/ibmtechviews.htm

    http://ummr.altervista.org/scansione0018d.jpg http://ummr.altervista.org/scansione0023.jpg http://ummr.altervista.org/scansione0078.jpg http://ummr.altervista.org/scansIBM30PIN.jpg http://ummr.altervista.org/PS280memories.jpg

    Yawn... Well, have fun. MAJ Tom out!

    On 30.8.2021 23:42, Tomas Slavotinek wrote:
    The chips were probably designed with applications like this in mind, possibly down to the pin assignments and recommended layout patterns. We
    need more info about these lil cubes, and find out where else were they
    used, if anywhere. (Perhaps the early 30-pin SIMMs?)

    Ultimately we may have to wait for Ken Shirriff's analysis (if he
    decides to check it out).

    On 30.8.2021 23:15, Louis Ohland wrote:
    Oops, conductor / trace count...

    On 8/30/2021 16:14, Louis Ohland wrote:
    Huh. Maybe there is a difference, pinout-wise. Look at the patterns,
    remember the CAS, RAS, and addressing thing? Maybe IBM designed these
    cards to reduce the connector count.

    I dunno.



    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tomas Slavotinek@21:1/5 to All on Tue Aug 31 13:37:52 2021
    I've added a few more memory card types, P/Ns, and photos:

    https://ardent-tool.com/8580/Planar_Memory.html

    This page in particular had some "new" card types:

    https://mastodonpc.tripod.com/personal/ps280mem.html

    Unfortunately the photos are too crappy to identify the chips.

    But the 90X8670 layout looks very familiar to me, I may actually have
    that card somewhere... I'll check.

    Also, see the 72X7475 board and compare it to the 15F6773 one... maybe
    some of the P/Ns on the same board are interchangeable after all? Or
    maybe we are dealing with chips that are partially defective or not up
    to spec for all use cases? See these two cards:

    https://ardent-tool.com/8580/Memory_IBM_1MB_3_Photo_Front_alt_1.jpg https://ardent-tool.com/8580/Memory_IBM_1MB_3_Photo_Front_alt_2.jpg

    What's up with the green/red/blue stripes? Additional QC?

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)