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

  • Double vision [ATI 8514 Ultra variants]

    From Louis Ohland@21:1/5 to All on Sun Jan 30 08:50:26 2022
    Were two PCB layouts? Earlier one had a jumper "JU8 CLK" near the center
    osc. The later card does not have this osc and no JU8.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tomas Slavotinek@21:1/5 to Louis Ohland on Mon Jan 31 22:10:33 2022
    On 30.01.2022 15:50, Louis Ohland wrote:
    Were two PCB layouts? Earlier one had a jumper "JU8 CLK" near the center
    osc. The later card does not have this osc and no JU8.

    Hmm, the box shows even an earlier version with through-hole PALs and
    what not:
    http://www.vgamuseum.info/images/pirx/ati/010_ati_03g_8514ultra.jpg

    A pre-production sample? No clue...

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Louis Ohland@21:1/5 to Tomas Slavotinek on Mon Jan 31 15:16:57 2022
    While transporting a fully-armed attack Brittany to a live-fire range,
    there were at least two Dot Clock Generators, ICS1394 [35 MHz] and
    ICS1494 [85 MHz]

    I'm sure there were more PCB versions...

    On 1/31/2022 15:10, Tomas Slavotinek wrote:
    On 30.01.2022 15:50, Louis Ohland wrote:
    Were two PCB layouts? Earlier one had a jumper "JU8 CLK" near the
    center osc. The later card does not have this osc and no JU8.

    Hmm, the box shows even an earlier version with through-hole PALs and
    what not:
    http://www.vgamuseum.info/images/pirx/ati/010_ati_03g_8514ultra.jpg

    A pre-production sample? No clue...

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Louis Ohland@21:1/5 to Tomas Slavotinek on Mon Jan 31 15:22:00 2022
    Note the blue ZIPP sockets, U10,12,14,16

    The video controller is fabbed by Toshiba.

    On 1/31/2022 15:10, Tomas Slavotinek wrote:
    On 30.01.2022 15:50, Louis Ohland wrote:
    Were two PCB layouts? Earlier one had a jumper "JU8 CLK" near the
    center osc. The later card does not have this osc and no JU8.

    Hmm, the box shows even an earlier version with through-hole PALs and
    what not:
    http://www.vgamuseum.info/images/pirx/ati/010_ati_03g_8514ultra.jpg

    A pre-production sample? No clue...

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tomas Slavotinek@21:1/5 to Louis Ohland on Mon Jan 31 22:59:49 2022
    On 31.01.2022 22:22, Louis Ohland wrote:
    Note the blue ZIPP sockets, U10,12,14,16

    The video controller is fabbed by Toshiba.

    On 1/31/2022 15:10, Tomas Slavotinek wrote:
    On 30.01.2022 15:50, Louis Ohland wrote:
    Were two PCB layouts? Earlier one had a jumper "JU8 CLK" near the
    center osc. The later card does not have this osc and no JU8.

    Hmm, the box shows even an earlier version with through-hole PALs and
    what not:
    http://www.vgamuseum.info/images/pirx/ati/010_ati_03g_8514ultra.jpg

    A pre-production sample? No clue...


    Probably a gate array.

    One half of the memory is socketed even on the later boards it seems.
    (this explains the memory size jumper)

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tomas Slavotinek@21:1/5 to Tomas Slavotinek on Mon Jan 31 23:04:49 2022
    On 31.01.2022 22:59, Tomas Slavotinek wrote:
    On 31.01.2022 22:22, Louis Ohland wrote:
    Note the blue ZIPP sockets, U10,12,14,16

    The video controller is fabbed by Toshiba.

    On 1/31/2022 15:10, Tomas Slavotinek wrote:
    On 30.01.2022 15:50, Louis Ohland wrote:
    Were two PCB layouts? Earlier one had a jumper "JU8 CLK" near the
    center osc. The later card does not have this osc and no JU8.

    Hmm, the box shows even an earlier version with through-hole PALs and
    what not:
    http://www.vgamuseum.info/images/pirx/ati/010_ati_03g_8514ultra.jpg

    A pre-production sample? No clue...


    Probably a gate array.

    One half of the memory is socketed even on the later boards it seems.
    (this explains the memory size jumper)


    Actually, the "9151" board seems to have all DRAMs soldered in. Yet the
    jumper header is still populated. Hmm...

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Louis Ohland@21:1/5 to Tomas Slavotinek on Mon Jan 31 16:10:31 2022
    You could get 512K with 4 sockets, or a 1MB with all positions filled
    with soldered ZIPPs.

    The Model 90 could come 8 sockets, 4/4, and 8 soldered.

    ATI chose to interleave socket/soldered...

    On 1/31/2022 16:04, Tomas Slavotinek wrote:
    Actually, the "9151" board seems to have all DRAMs soldered in. Yet the jumper header is still populated. Hmm...

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tomas Slavotinek@21:1/5 to Louis Ohland on Mon Jan 31 23:49:24 2022
    On 31.01.2022 23:10, Louis Ohland wrote:
    You could get 512K with 4 sockets, or a 1MB with all positions filled
    with soldered ZIPPs.

    The Model 90 could come 8 sockets, 4/4, and 8 soldered.

    ATI chose to interleave socket/soldered...

    On 1/31/2022 16:04, Tomas Slavotinek wrote:
    Actually, the "9151" board seems to have all DRAMs soldered in. Yet
    the jumper header is still populated. Hmm...

    Yep, I understand that, but why populate the 512K/1M jumper if all chips
    are soldered in on the "9151" sample? Welp, it may come handy if there's
    a defect in the top 512K...

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