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

  • KTM-4000/M80 ponderings

    From Louis Ohland@21:1/5 to Louis Ohland on Sun Aug 29 07:40:43 2021
    Oops, 2 rows of KM44C256AJ-8 on the front, 2 rows on the back.

    1 row of KM41C256AJ-7 front, 1 row on the back.

    So, this is equivalent to FOUR 1MB SIMMs?

    On 8/29/2021 07:36, Louis Ohland wrote:
    Burndy RP196U30RB

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Louis Ohland@21:1/5 to Louis Ohland on Sun Aug 29 07:52:01 2021
    Oh, great. Burndy became part of Framatome by 1997.

    On 8/29/2021 07:40, Louis Ohland wrote:
    Oops, 2 rows of KM44C256AJ-8 on the front, 2 rows on the back.

    1 row of KM41C256AJ-7 front, 1 row on the back.

    So, this is equivalent to FOUR 1MB SIMMs?

    On 8/29/2021 07:36, Louis Ohland wrote:
    Burndy RP196U30RB


    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Louis Ohland@21:1/5 to All on Sun Aug 29 07:36:23 2021
    https://www.ardent-tool.com/8580/Planar_Memory.html

    Kingston KTM-4000/M80 4 MB Memory Card (REV.A00)

    https://www.ardent-tool.com/8580/Memory_Kingston_KTM4000_Photo_Front.jpg

    Buffer/Inverter Based Peripheral Driver, 0.2A, BIPolar, PDIP20, PLASTIC https://www.ardent-tool.com/datasheets/AMD_Am2965_2966.pdf

    U3-10,15-22 KM44C256AJ-8 256Kx4 DRAM [2 rows]
    U11-14,23-26 KM41C256J-8 256Kx1 DRAM (parity) [1 row]

    Looks like someone SIMMply took the memory off of two 2MB FPM and tossed
    it on a card.

    The Am2966PC is not used on any other SIMM [72 pin -OR- Model 80
    flavored] that I have seen.

    Notice the pads for PD bits "R1" and R2". Based on that, I'd make a SWAG
    that the IBM cards are internally masked.

    Many cards use a Burndy RP196U30RB-1, 8735 to 9043. The Kingston card
    uses a Panduit 130-096-553, "15 0 3", whatever that date code is. One
    other connector, looks "Molexy" or "AMPy", T74652-2

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Louis Ohland@21:1/5 to Louis Ohland on Sun Aug 29 07:54:16 2021
    I H8te old sites that used Java...

    On 8/29/2021 07:40, Louis Ohland wrote:
    Oops, 2 rows of KM44C256AJ-8 on the front, 2 rows on the back.

    1 row of KM41C256AJ-7 front, 1 row on the back.

    So, this is equivalent to FOUR 1MB SIMMs?

    On 8/29/2021 07:36, Louis Ohland wrote:
    Burndy RP196U30RB


    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tomas Slavotinek@21:1/5 to Louis Ohland on Sun Aug 29 15:57:56 2021
    On 29.8.2021 14:40, Louis Ohland wrote:

    So, this is equivalent to FOUR 1MB SIMMs?

    Essentially yes.

    Not sure why exactly are the extra drivers needed...

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Louis Ohland@21:1/5 to Tomas Slavotinek on Sun Aug 29 10:33:08 2021
    My SWAG is they took care of out of tolerance signals? reduced components?

    But if you look at the Kingston card, there's a lot of SMD caps...

    On 8/29/2021 08:57, Tomas Slavotinek wrote:
    On 29.8.2021 14:40, Louis Ohland wrote:

    So, this is equivalent to FOUR 1MB SIMMs?

    Essentially yes.

    Not sure why exactly are the extra drivers needed...


    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tomas Slavotinek@21:1/5 to Louis Ohland on Sun Aug 29 19:44:23 2021
    On 29.8.2021 17:33, Louis Ohland wrote:
    My SWAG is they took care of out of tolerance signals? reduced components?

    The IBM DRAM chips probably have somewhat different characteristics...
    but yeah, not sure what exactly is going on there.

    But if you look at the Kingston card, there's a lot of SMD caps...

    Bypass caps. Usually you want at least one cap per IC in applications
    like this, and they should be physically close to the individual chips.

    And that's exactly what you see on the Kingston card, and all the IBM
    cards as well for that matter...

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From moussa@21:1/5 to Louis Ohland on Mon Aug 30 09:56:35 2021
    On 29/8/21 8:36 pm, Louis Ohland wrote:
    https://www.ardent-tool.com/8580/Planar_Memory.html

    Kingston KTM-4000/M80 4 MB Memory Card (REV.A00)

    https://www.ardent-tool.com/8580/Memory_Kingston_KTM4000_Photo_Front.jpg

    Buffer/Inverter Based Peripheral Driver, 0.2A, BIPolar, PDIP20, PLASTIC https://www.ardent-tool.com/datasheets/AMD_Am2965_2966.pdf

    U3-10,15-22 KM44C256AJ-8 256Kx4 DRAM [2 rows]
    U11-14,23-26 KM41C256J-8 256Kx1 DRAM (parity) [1 row]

    Looks like someone SIMMply took the memory off of two 2MB FPM and tossed
    it on a card.

    The Am2966PC is not used on any other SIMM [72 pin -OR- Model 80
    flavored] that I have seen.

    Notice the pads for PD bits "R1" and R2". Based on that, I'd make a SWAG
    that the IBM cards are internally masked.

    Many cards use a Burndy RP196U30RB-1, 8735 to 9043. The Kingston card
    uses a Panduit 130-096-553, "15 0 3", whatever that date code is. One
    other connector, looks "Molexy" or "AMPy", T74652-2
    RAM = 80ns
    Parity = 70ns

    !!

    IIRC, there was a time of shortage, and they used whatever available
    ATT, as the hunger for RAM grew thanks to MS.

    also there was that transitional moment from 100ns to 80ns up to 70ns

    --
    Moussa

    "People alike with a similar circumstances, tend to find each
    others across space and time, given enough time in life, no
    matter distance, language, race, colour and believes" (c) MEK
    *** Do Not Copy, Duplicate or Use without my Permission ***

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tomas Slavotinek@21:1/5 to moussa on Mon Aug 30 08:56:35 2021
    On 30.8.2021 3:56, moussa wrote:
    RAM = 80ns
    Parity = 70ns

    Thanks, I'll fix that.

    (Ultimately it doesn't matter, the module access time is dictated by the slowest part of course.)

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