SVGA/A news

[Wolfgang Gehl]

While playing with a Solaris setup on a 9595 I ended up creating a DOS
partition and giving SVGA Adapter/A a try.
This mysterious card currently has no OS support beside OS/2 with VGA
resolution and 256 colors and VESA VBE support for DOS.

First I tried the VESA driver from m95svga.exe. It offers VESA VBE 1.2
support for 256 colors for resolutions up to 800*600 pixels customized
for the display's horizontal refresh rate. Since I use a LCD monitor, I
disabled 640*400 and the 16 colors interlaced 1024*768 mode and choose
60Hz for 640*480 and 800*600 mode.

Do you remember trying to play "Simcity 2000" or "The Settlers" with
XGA/2? Never got it running. With SVGA/A and loaded VESA driver these
games are playable without a hitch. It's kind of a crazy to play DOS
SVGA games on a 9595 RAID5 Server, however, it works!

Simcity 2k for DOS contains a VESA driver collection with an IBM
directory. Inside are two VESA drivers. One is for XGA/XGA2 (for Simcity
dysfunctional, LOL) the other one is for the PS/2 25SX planar video. A
look at William’s 25SX site shows that the SVGA/A chipset is sitting on
the 25SX planar. And he mentions that it is common on 40SX and 56/57
SX/SLC planars too.

A search for "IBM 84F7985" took me to a thread on a PS/1 forum with the
title "The mystery of the PS/1 2121’s (S)VGA"
https://ps1stuff.wordpress.com/2016/09/13/the-mistery-of-the-ps1-2121s-svga/
Obviously IBM used this chip in several x86 machines, and for us most
important, inside the micro channel PS/2 ThinkPads too. A ThinkPad
without Windows drivers is hard to imagine. And right, Louis has the 256
colors Windows 3.1 driver, which is in fact the 25SX_AND_700C_AND_720C
display driver:
http://ps-2.kev009.com/ohlandl/9552/9552_700C.html
Since we have the driver, we have the chipset name too: IBMVGA256c

Well, don't expect to much, the Windows 3.1x setup offers 256 colors for
640x480 pixels only. It works in conjunction with the DOS VESA VBE
driver and seems to be an early if not the first Windows VESA VBE driver
at all. I've tested with windows 3.1, 3.11, and 95. With 95 the 16bit
driver slows down the system and there are severe incompatibilities e.g.
with explorer.exe.

Here is the incomplete information I could find so far:


The IBMVGA256c chipset
======================


IBM PS/1 2121 [1]
-------------

Year:            1991 (?)

Video chipset:   IBM 84F7985 (= TI CF62011BPC)
                 IBMVGA256c

RAM DAC:         INMOS IMSG171P-35 [2]
                 35MHz 8 Bit DAC up to 256 colors from a total of
262,144 colors

RAM:             256KB 100ns VRAM


IBM PS/2 8540SX
---------------

Year:            1991

Video chipset:   IBMVGA256c (?)

RAM DAC:         ?

RAM:             ?


IBM PS/2 8557SX [3]
---------------

Year:            1991

Video chipset:   IBM 84F7985
                 IBMVGA256c

unknown:         IBM 85F0120

RAM DAC:         INMOS IMSG171P-50 [2]
                 50MHz 8 Bit DAC up to 256 colors from a total of
262,144 colors

RAM:             512KB 80ns VRAM


IBM PS/2 8525SX [4]
---------------

Year:            1991 (?)

Video chipset:   IBM 84F7985
                 IBMVGA256c

RAM DAC:         INMOS IMSG171P [2]
                 (?)MHz 8 Bit DAC up to 256 colors from a total of
262,144 colors

RAM:             512KB 80ns VRAM


IBM PS/2 Thinkpad 700C/720C [5]
---------------------------

Year:            1992

Video chipset:   IBM 39G8600
                 IBMVGA256c

Clock chip:      ICS ICS2494 [6]
                 135MHz Dual Video/Memory Clock Generator for VGA,
Super VGA, XGA video adapters, 8514A, TMS34010, TMS34020

RAM DAC:         ?MHz on chip?

RAM:             512KB 80ns VRAM


IBM SVGA Adapter/A [7]
------------------

Year:            1992 (together with Server 85)

Video chipset:   IBM 42G3410 (= TI CF63011PCM)
                 IBMVGA256c

unknown:         IBM 85F0120 (= OKI M7U016-008)

Clock chip:      ICS ICS2494 [6]
                 135MHz Dual Video/Memory Clock Generator for VGA,
Super VGA, XGA video adapters, 8514A, TMS34010, TMS34020

RAM DAC:         INMOS IMSG176J-80Z [8]
                 80MHz 8 Bit DAC up to 256 colors from a total of
262,144 colors

RAM:             512KB 80ns VRAM


Driver Support
--------------

DOS:             generic: 700c120-exe [9]
                 SVGA/A:  m95svga.exe [10]

                 VESA/VBE driver*

                 with 256KB VRAM and at least 40MHz RAM DAC
                 800x600 pixels, 16 colors, 60Hz**

                 with 512KB VRAM and at least 30MHz RAM DAC
                 640x480 pixels, 256 colors, 60Hz

                 with 512KB VRAM and at least 40MHz RAM DAC
                 800x600 pixels, 265 colors, 60Hz**

                 with 512KB VRAM and at least 50MHz RAM DAC
                 1024x768 pixels, 16 colors, 43Hz interlaced***

                 with 256KB VRAM and at least 30MHz RAM DAC
                 132-column x 25-row text

                 * RAM DAC requirements are calculated with the formula
                 line frequency in kHz = ([vertical resolution] *
[horizontal refresh rate] Hz + [vertical resolution] * [horizontal
refresh rate] Hz * 0,1)/1000
                 RAM DAC pixel frequency in MHz = ([horizontal
resolution] * [line frequency] kHz + [horizontal resolution] * [line
frequency] kHz *0,2)/1000

                 ** the IBMVGA256c chipset does 800x600 pixels, 56Hz
too but LCDs usually don'tsupport refresh rates below 60Hz and you will
not accept a 56Hz refresh rate on aCRT monitor.

                 *** LCDs usually don't support an interlaced video
resolution. 1024x768 pixels,16 colors, 60Hz requires at least a 70MHz
RAM DAC


OS/2:            from v2.11 up build in under the name "IBM VGA 256c"
                 640x480 pixels, 256 colors with 512KB VRAM and at
least a 30MHz RAM DAC[11]


Windows 3.1x:    generic: 700c120-exe [9]
                 640x480 pixels, 256 colors with 512KB VRAM and at
least a 30MHz RAM DAC together with DOS VESA/VBE software driver


[1]
https://ps1stuff.wordpress.com/2016/09/13/the-mistery-of-the-ps1-2121s-svga/
[2] IMSG171P is compatible to BT476:
https://www.datasheetarchive.com/pdf/download.php?id=527994991358bc5dc0a9e619b66f0a2d71ba0d&type=O&term=IMSG171
    BT47x circuit description:
https://www.datasheetarchive.com/pdf/download.php?id=dd6f6622ae7025772fc684bf7840a63ab87fae&type=M&term=bt476
[3] http://ps-2.kev009.com/ohlandl/8557/8557SX-Planar.html
[4] http://www.walshcomptech.com/ps2/25sx.htm
[5] http://www.vgamuseum.info/images/vlask/ibm/39g8600f.jpg
    http://www.vgamuseum.info/images/vlask/ibm/39g8600fz.jpg
[6]
ftp://bitsavers.informatik.uni-stuttgart.de/pdf/integratedCircuitSystems/1994_Integrated_Circuit_Systems_Data_Book.pdf,
pp. 161-165
[7]
http://www.vgamuseum.info/images/palcal/ibm/44_ibm_fru_71f4877_top_hq.jpg
[8]
https://www.datasheetarchive.com/pdf/download.php?id=30c4f464de799e8bfcadd50feda3f34ef84d46&type=M&term=IMSG17
[9] http://ps-2.kev009.com/ohlandl/9552/r700c120.exe
[10] http://ps-2.kev009.com/ohlandl/video/m95svga.exe
[11] a higher resolution may be possible according to:
http://www.edm2.com/index.php/DDDR/2_-_32-Bit_Super_VGA_Display_Driver

[WBSTClarke]

There's an XGA(-2) VESA VBE DOS TSR in the XGA v2.1.2 Drivers Diskette:
http://ps-2.kev009.com/ohlandl/video/xga212.exe

[Wolfgang Gehl]

Am 23.11.19 um 17:33 schrieb WBSTClarke:

> On Monday, 18 November 2019 22:16:07 UTC, Wolfgang Gehl wrote:

>> [snip]
>> ... too much text ...
>> [snip]

> There's an XGA(-2) VESA VBE DOS TSR in the XGA v2.1.2 Drivers Diskette:
> http://ps-2.kev009.com/ohlandl/video/xga212.exe

Thanks for the link Tim!

But I'm not talking about XGA. I'm talking about SVGA Adapter/A.
http://ps-2.kev009.com/ohlandl/video/SVGA.html
a video adapter with VESA DOS TSR and OS/2 support for 640x480x256 colors.

It took a while to dig through the IBM archives but finally I found the
Win3.1x SVGA drivers for the card in the Warp 3 FixPak XR_W043. I took
what was usable from the ThinkPad Win3.1 driver, rewrote the
oemsetup.inf and here is now what the world is probably not waiting for:

Win3.11 drivers for SVGA Adapter/A:
http://www.franken-online.de/ymmv/src/ibmsvgaa.exe
with 640x480x256 colors and 800x600x256 colors support

Since they all use the same chipset (IBM VGA256c) here are Win3.11
drivers for PS/2 8525SX and Thinkpad 700C/720C:
http://www.franken-online.de/ymmv/src/ibm256c.exe
with 640x480x256 colors support

These drivers should support the PS/2 planar video of the 8540SX and the
8557SX as well as long as 512KB video ram is installed and the VESA DOS
TSR makes the resolutions (640x480 and 800x600) available to the driver.
Other test results are very appreciated since I could test with the SVGA
Adapter/A only. Be careful with original IBM/PS2 tube monitors because a
800x600 resolution may damage them.

Wolfgang

[WBSTClarke]

On Monday, 16 December 2019 20:58:24 UTC, Wolfgang Gehl wrote:
> Am 23.11.19 um 17:33 schrieb WBSTClarke:
> > On Monday, 18 November 2019 22:16:07 UTC, Wolfgang Gehl wrote:
> >> [snip]
> >> ... too much text ...
> >> [snip]
> > There's an XGA(-2) VESA VBE DOS TSR in the XGA v2.1.2 Drivers Diskette:
> > http://ps-2.kev009.com/ohlandl/video/xga212.exe
>
> Thanks for the link Tim!
>
> But I'm not talking about XGA. I'm talking about SVGA Adapter/A.
> http://ps-2.kev009.com/ohlandl/video/SVGA.html
> a video adapter with VESA DOS TSR and OS/2 support for 640x480x256 colors.
>

<snip>

Sorry Wolfgang, I remembered the SVGAVESA.EXE on there, but not that it required an XGA(-2) chipset in VGA mode. Ooopsie!!!

Cheers, Tim

[WBSTClarke]

On Monday, 16 December 2019 21:23:45 UTC, WBSTClarke wrote:
> On Monday, 16 December 2019 20:58:24 UTC, Wolfgang Gehl wrote:
> > Am 23.11.19 um 17:33 schrieb WBSTClarke:
> > > On Monday, 18 November 2019 22:16:07 UTC, Wolfgang Gehl wrote:
> > >> [snip]
> > >> ... too much text ...
> > >> [snip]
> > > There's an XGA(-2) VESA VBE DOS TSR in the XGA v2.1.2 Drivers Diskette:
> > > http://ps-2.kev009.com/ohlandl/video/xga212.exe
> >
> > Thanks for the link Tim!

I'm talking about SVGA Adapter/A.
http://ps-2.kev009.com/ohlandl/video/SVGA.html
a video adapter with VESA DOS TSR and OS/2 support for 640x480x256 colors.

Sorry Wolfgang, I remembered the SVGAVESA.EXE on there, but not that it required an XGA(-2) chipset in VGA mode. Ooopsie!!!

Louis, if you're reading this, the page could use an update. I'm sure there is VGA256C adapter support for OS/2. Especially in the later DDPak (2?) and OS/2 Developer's Toolkits (e.g. v3 and v4.5). I've not tried 800x600 with an SVGA/A as I didn't have a compatible monitor at the time.

Cheers, Tim

[Louis Ohland]

Do tell... I looked, but without sufficient caffeine, it is not apparent.

[WBSTClarke]

Some downloads to add to the page, from Wolfgang's post, and I've seen "IBM VGA256C" as an OS/2 supported driver under at least Warp 4:

[Louis Ohland]

http://ps-2.kev009.com/ohlandl/video/SVGA.html

[dav...@wnmt.com]

Hi Wolfgang,

> IBM PS/2 8540SX
> ---------------
>
> Year:            1991
>
> Video chipset:   IBMVGA256c (?)
>
> RAM DAC:         ?
>
> RAM:             ?

The same planar is shared between the 35SX and 40SX - I have a couple of each of these systems.

[WBSTClarke]

On Tuesday, 17 December 2019 03:16:51 UTC, Louis Ohland wrote:
> http://ps-2.kev009.com/ohlandl/video/SVGA.html

Thank you, Louis and Wolfgang both. This is a significant addition for the DOS gaming users.

FWIW, OS/2 Warp 3 runs pretty well with at least 8MB RAM and a reasonably quick 386-class CPU. I used a Model 80 with a 20/40MHz clock-doubling Cyrix Cx486DRx2 upgrade processor (later upgraded to the 8580-A16 25MHz planar and 25/50MHz) for a number of years. It served my brother and his kids for more years after that.

[Lionel Harris]

This card can be run in a standard slot right? Obviously it wont fit AVE, but if you have a system with AVE, could you put it in a regular slot and deal with the wonkyness of both video outputs? I thought I saw a thread about that, but I can't find it.

[Louis Ohland]

Gender bender, heart in a blender...

The SVGA /A is a Base Video card, see the small tab on the right?

Remember, the SVGA /A was released for the 95A as the lone video card in
a system sans a base video...

Not sure what configuration you envision. Placing a Base Video Extension
[BVE] card in a slot without a video extension [AVE or BVE] means no
other card or planar video can share Base Video. So therefore, you will
have possibly two separate video cards that can operate two monitors at
once.

Now IIRC, one model of the 8580 has TWO video slots, IBM says not to
have two video sources active at one time. I'm confused, but not gender
related...

[Louis Ohland]

8580-Axx J6,7 16-bit AVE slot

http://ps-2.kev009.com/ohlandl/8580/8580_A.html

[Wolfgang Gehl]

Am 17.12.19 um 06:50 schrieb dav...@wnmt.com:

Hi David,

Could you please have a look whether the chipset is 84F7985, which
RAM-DAC is used (INMOS IMSG_anything-xxMHz) and whether there are 2 or 4
ZIP-VRAM chips on the planar?

I have thoroughly tested the device driver with SVGA Adapter/A, but
would feel better if the device driver were checked by another NG
member. At least ibm256c.exe should work.

[Louis Ohland]

http://www.mcamafia.de/pdf/ibm_hitrc10.pdf

Page 97 physical

The video buffers for the base video can be turned off, and the AVE card
can output via the base video DAC to the display.

The video extension can be driven only one direction at a time...

[Wolfgang Gehl]

Am 16.12.19 um 22:33 schrieb WBSTClarke:

Hi Tim,

it seems that IBM never released an 800x600 capable driver for the
VGA256c chip set, not for Win3.1x and not for OS/2 (at least I couldn't
find it). Maybe because there was no compatible IBM tube monitor
available in 1992, or because early implementations of the VGA256c
chipset don't support resolutions above 640x480.

As a lucky side effect of the device driver (wspdsf.drv) it supports -
beside IBM256c - also early ATI, Trident, Tseng, Ahead, Oak, Paradise,
Video7 and even Cirrus Logic chipsets. Since these video cards support
resolutions above 640x480, the driver does it too.

Both driver packages, ibmsvgaa.exe and ibm256c.exe are no offical IBM
releases but more a kind of an arragement of IBM device driver files
that probably never was meant to be put together. For that reason I
would prefer to hear from successful installations before the packages
are placed on the ardent tool.

Wolfgang

[dav...@wnmt.com]

Hi Wolfgang,

> Could you please have a look whether the chipset is 84F7985, which
> RAM-DAC is used (INMOS IMSG_anything-xxMHz) and whether there are 2 or 4
> ZIP-VRAM chips on the planar?
>
> I have thoroughly tested the device driver with SVGA Adapter/A, but
> would feel better if the device driver were checked by another NG
> member. At least ibm256c.exe should work.

Will do.

[dav...@wnmt.com]

I even have a good picture if there is a way to send it to you. TI-branded chip with the 84F7985 FRU, copyright IBM 1990, manufactured the 46th week of 1991. RAMDAC (plastic) INMOS IMSG171P-50, manufactured the 41st week of 1991. Two ZIP-VRAM chips populated, two more pads left vacant - and a 3-pin jumper ('J1') nearby.

[dav...@wnmt.com]

Hi Wolfgang,

> > > I have thoroughly tested the device driver with SVGA Adapter/A, but
> > > would feel better if the device driver were checked by another NG
> > > member. At least ibm256c.exe should work.

This 35SX will be good to test from - I had put WfW 3.11 on it years ago and it has the standard VGA driver - If the floppy doesn't work I can put it on my network.

And it was very easy to put it in my equipment room connected up to the KVM...

[dav...@wnmt.com]

Hi Wolfgang,

Have you been able to run any code for the video BIOS? - INT 10h, Function 4Fh can be used to get SVGA information - I'll look through for what the UNPC code may have too.

[WBSTClarke]

No, it has the BVE extension, for server-class planars, but will fit a non-extended slot Okay for an add-on adapter.

[davidr...@gmail.com]

Going down the rabbit hole on this thought, but didn’t Microsoft’s DDK at one time include the source for the SVGA driver? I seem to recall looking for the XGA sources and got disappointed as I only found the SVGA source...

[WBSTClarke]

On Wednesday, 18 December 2019 15:40:09 UTC, davidr...@gmail.com wrote:
> Going down the rabbit hole on this thought, but didn’t Microsoft’s DDK at one time include the source for the SVGA driver? I seem to recall looking for the XGA sources and got disappointed as I only found the SVGA source...

I don't know about Windows SDKs, but there was SVGA driver source in the later OS/2 SDKs.

[Wolfgang Gehl]

Am 18.12.19 um 01:22 schrieb dav...@wnmt.com:

Hi David,

that is interesting. It is the IBMVGA256c chipset. The good news is that
the RAMDAC has 50 MHz. 39 MHz should be enough for 800x600 with a
horizontal refresh rate of 60 Hz, which is perfect for your LCD. The two
empty ZIP VRAM pads are less good. This means 256 KB of video RAM. It
gets maximum with 800x600 with 16 colors. No 16-color device driver is
available for the chipset.

Maybe you can pull two ZIP VRAM chips from another 40SX and fill the
empty pads?

A picture would be great. Send it to
wolfgang_no_spam(at)maxi-dsl(dot)de. Do not make any changes to the
address. It is valid and will be read

The only tool that can make SVGA visible to the operating system is the
VESA VBE driver (vesa.com). Use the svgaa.exe package and follow the
instructions in the readme.txt file. Without a VESA-VBE driver, the
chipset is recognized as a VGA.

Wolfgang

[Louis Ohland]

Better yet, what are those ZIPs?

[Wolfgang Gehl]

Am 18.12.19 um 21:04 schrieb Louis Ohland:

> Better yet, what are those ZIPs?

Hi Louis,

I guess two TC511665BZ like those on the SVGA Adapter/A but plugged in,
not soldered.

Wolfgang

[Louis Ohland]

Thanks. Sendt a link to Dave...

[dav...@wnmt.com]

Hi Wolfgang,

> The two empty ZIP VRAM pads are less good. This means 256 KB of video
> RAM. It gets maximum with 800x600 with 16 colors. No 16-color device
> driver is available for the chipset. Maybe you can pull two ZIP VRAM
> chips from another 40SX and fill the empty pads?

It's just solder pads, as you will see in the picture - If I put in ZIP sockets (and I did have extra VRAM to fit it - somewhere) on one of my systems it will be some time. I'll try to take a side picture so the we are sure of capacity.

[dav...@wnmt.com]

Hi Wolfgang,

With ECA 085 talking about an "OEM video card" for the 8535/8540 - It appears that IBM may have installed one in certain instances to get around those video limitations.

[dav...@wnmt.com]

I had thought that William Walsh tried to populate the ZIP VRAM sockets of an 8556/8557 without success - Let me dig.

[Louis Ohland]

Ditto. Songs in the tune of Czech...

[dav...@wnmt.com]

I found this from Peter: https://groups.google.com/forum/#!msg/comp.sys.ibm.ps2.hardware/GWBCPVddmpY/EeddPRl5GD4J

[dav...@wnmt.com]

> I found this from Peter: https://groups.google.com/forum/#!msg/comp.sys.ibm.ps2.hardware/GWBCPVddmpY/EeddPRl5GD4J

Peter is wrong about the video chipset being the same as the initial release microchannel PS/2s - but I wandt to find the empirical evidence - in any regard, I'll attempt it when the video RAM chips arrive.

[dav...@wnmt.com]

>> I'll attempt it when the video RAM chips arrive.

They arrived - but the VRAM upgrade to the sockets on 8556/8557 planars is a confirmed FAIL. I changed the jumper position, but IBM's SIT (System Information Tool) still only reports 256Kb VRAM. It won't be a success to solder in sockets or chips to the solderpads of the 8535/8540 planars.

I am going to scope out where the jumper pins connect for more of a grasp of what may be going wrong...

[Louis Ohland]

Dave, ever think of asking Tim to look at the system programs for the
8556? Is this a limitation of the system configuration, the BIOS, or yes
to the above?

[dav...@wnmt.com]

> Dave, ever think of asking Tim to look at the system programs for the
> 8556? Is this a limitation of the system configuration, the BIOS, or yes
> to the above?

The 8535/8540 planar has the same video chipset - no system configuration under ISA.

I just looked at the SVGA/A in more detail (good images at http://www.vgamuseum.info/index.php/companies/item/191-ibm-42g3410-svga)

As Wolfgang's chart shows, 42G3410 on SVGA/A works with 512Kb VRAM - 84F7985 chipset on 8535/8540/8556/8557 doesn't seem to work with adding VRAM. Is that a fault of the 84F7985 chip? Datecode on the 84F7985 here shows '9208' - The SVGA/A 42G3410 at the link has a datecode of '9406'.

I say 84F7985 -> 42G3410 -> 39G8600 (Thinkpad 700/720 video)...

[dav...@wnmt.com]

Now we're at the level of de-soldering PQFP chips - beyond my forte

[dav...@wnmt.com]

On Saturday, December 21, 2019 at 4:07:43 PM UTC-7, dav...@wnmt.com wrote:
> Now we're at the level of de-soldering PQFP chips - beyond my forte

Ok, datecodes on the 39G8600 look to be '9239' close to dates of the 84F7985 but a much bigger package. But two years' difference to the 42G3410 with the same pinout. As soon as I locate an SVGA/A here I'll start making sure something isn't miswired.

[dav...@wnmt.com]

On Saturday, December 21, 2019 at 4:07:43 PM UTC-7, dav...@wnmt.com wrote:

> Now we're at the level of de-soldering PQFP chips - beyond my forte

Oh, here is the 39G8600: http://www.vgamuseum.info/index.php/cpu/item/190-ibm-39g8600

[dav...@wnmt.com]

Datasheet on the TC511665BZ (starting on page A-471): http://bitsavers.trailing-edge.com/components/toshiba/_dataBook/1991_Toshiba_1M_DRAM_1M_VRAM.pdf

[dav...@wnmt.com]

William Walsh reports this on his Model 25SX page:

"U42 - IBM SVGA 84F7985 (same as on the 512K SVGA/A, 40SX and 57sxx planars)"

So find me an SVGA/A with 84F7985 for him to have made that declaration.

There's also a "J16" that I think is for the VRAM expansion just like on the 8535/8540 and 8556/8557 planars with the 84F7985 chip.

I think the 42G0120 chip is for interfacing to the microchannel bus - It is on the SVGA/A and 8556/8557 planars, but not on the 25SX or 8535/8540 planars that are ISA-based.

http://www.walshcomptech.com/ps2/25sx.htm

[dav...@wnmt.com]

> So find me an SVGA/A with 84F7985 for him to have made that declaration.

Especially with 512Kb VRAM...

[Wolfgang Gehl]

Am 21.12.19 um 23:04 schrieb dav...@wnmt.com:

>>> I'll attempt it when the video RAM chips arrive.
> They arrived - but the VRAM upgrade to the sockets on 8556/8557 planars is a confirmed FAIL. I changed the jumper position, but IBM's SIT (System Information Tool) still only reports 256Kb VRAM. It won't be a success to solder in sockets or chips to the solderpads of the 8535/8540 planars.
>
> I am going to scope out where the jumper pins connect for more of a grasp of what may be going wrong...

Hi David,

how much video RAM is reported by vesa.com from savgaa.exe?

Wolfgang

[dav...@wnmt.com]

Hi Wolfgang,

> how much video RAM is reported by vesa.com from savgaa.exe?

Today I am going to make a few diskette images - Ironically from Windows XP or Server 2003 R2 systems.

[IBMMuseum]

So I'm looking for datecode information for the 84F7985 - Anything later than very early in 1992? And also for the 42G3410 - before very early in 1994. If I had a datecode midway between what I have seen, something in early 1993, which chip would it be?

[IBMMuseum]

I booted up an unexpanded VRAM 8557 to the OS/2 2.0 it had, then ran the VESA TSR from a command prompt. It reported 256Kb VRAM correctly. So I upgraded the VRAM and moved the jumper - The same sequence with 512Kb VRAM still reported 256Kb.

I've got to get to somewhere that gives me expected, consistent results...

[IBMMuseum]

And finally located a 9585 'X' with an SVGA/A that I pulled out of the basement...

[IBMMuseum]

On Monday, December 23, 2019 at 3:20:32 PM UTC-7, IBMMuseum wrote:
> And finally located a 9585 'X' with an SVGA/A that I pulled out of the basement...

The 9585 'X' has a bad hard drive - so I have pulled the SVGA/A for being a way to test by a multi-meter. A 9585 'K/N' I have also has the SVGA/A and boots to DOS - It will be my test system.

Now we are cooking...

[IBMMuseum]

Hi Wolfgang,

I've learned why the PS/1 "Pro" 2123 (the limited-production PS/1 in a Model 30-style case) has the same video chip as the Model 25SX: It is the same planar! Well, the HDB-15 on the back panel is populated and the 2x8 video header is removed for the first difference. It also has a SIMM socket populated in the MEM2 position.

http://www.walshcomptech.com/ps2/25sx.htm

I'm prompting whether all the VRAM chips are populated like the 25SX...

[IBMMuseum]

VRAM is 256Kb on the PS/1 Pro 2123, with sockets to expand it...

[Wolfgang Gehl]

Hi David,

so IBM uses a PS/1 mainboard in a PS/2 machine? IBM also has to live on
something, but it sounds . . . disillusioning.

Wolfgang

[Tomas Slavotinek]

On 27.7.2020 14:32, Wolfgang Gehl wrote:
>
> Hi David,
>
> so IBM uses a PS/1 mainboard in a PS/2 machine? IBM also has to live on
> something, but it sounds . . . disillusioning.
>
> Wolfgang
>

It's the other way around - they recycled various PS/2 hardware to build
the Pro.

But yeah, it was pretty much the same thing, just packaged
differently... and well, the support/warranty program was probably
different too.

[IBMMuseum]

>> so IBM uses a PS/1 mainboard in a PS/2 machine? IBM also has
>> to live on something, but it sounds . . . disillusioning.

> It's the other way around - they recycled various PS/2
> hardware to build the Pro. But yeah, it was pretty much
> the same thing, just packaged differently...

Tomas has it right (PS/2 hardware put into a PS/1 system) - At least the "PS/1 Pro" 2123 was limited production. I'm probing if the system BIOS is the same, but that is very likely. Completely different "Starter Diskette", maybe that also has a video driver?

It also means that my IDE BIOS extension hack should work to gain 2Gb HDD partitions using the planar connection...

[IBMMuseum]

>> It's the other way around - they recycled various PS/2
>> hardware to build the Pro. But yeah, it was pretty much
>> the same thing, just packaged differently...
>
> Tomas has it right (PS/2 hardware put into a PS/1 system) - At
> least the "PS/1 Pro" 2123 was limited production. I'm probing

> if the system BIOS is the same, but that is very likely...

I've discovered that the PS/1 Pro 2123 *was* supposed to be a "PS/2 Model 30SX" - It appears that the planned release of the Model 35SX and 40SX likely made IBM redirect it to be a PS/1.

[Louis Ohland]

http://ps-2.kev009.com/pcpartnerinfo/ctstips/937e.htm

On 8/16/2020 22:30, IBMMuseum wrote:
> "PS/1 Pro" 2123

[Louis Ohland]

https://www-01.ibm.com/common/ssi/ShowDoc.wss?docURL=/common/ssi/rep_ca/1/877/ENUSZG92-0161/index.html&request_locale=en

[WBSTClarke]

640x480x16-colour VGA sounds horribly constrained memory-wise, 128KB?

[IBMMuseum]

> 640x480x16-colour VGA sounds horribly constrained memory-wise, 128KB?

With the same planar as the 25SX, but it looks like the VRAM might have been reduced to 256Kb again - with sockets or solderpads to expand it to 512Kb. The irony is that all of the PS/1 "Pro" 2123 submodels are a 386SX-20 - On the 25SX it was only listed as 16MHz, if you got a 20MHz it was just luck, no particular submodel was spec'd at that. The PS/1 Pro adds the SIMM sockets and has more planar memory too (the 'K' 25SX submodel came with 1Mb soldered in place, the 'G' had 4Mb like the PS/1 Pro and the external VGA connection - it looks like the 16MHz planars were put in the 25SX 'G' and the 20MHz planars cycled over to be PS/1 Pro's).

[Louis Ohland]

Dave, I haven't been able to come up with an original announcement
letter for the Pro. It might be it was called something else or the
wording was way different.

On 8/18/2020 15:16, IBMMuseum wrote:
>> 640x480x16-colour VGA sounds horribly constrained memory-wise, 128KB?
>

> With the same planar as the 25SX, but it looks like the VRAM might have been reduced to 256Kb again - with sockets or solderpads to expand it to 512Kb.. The irony is that all of the PS/1 "Pro" 2123 submodels are a 386SX-20 - On the 25SX it was only listed as 16MHz, if you got a 20MHz it was just luck, no particular submodel was spec'd at that. The PS/1 Pro adds the SIMM sockets and has more planar memory too (the 'K' 25SX submodel came with 1Mb soldered in place, the 'G' had 4Mb like the PS/1 Pro and the external VGA connection - it looks like the 16MHz planars were put in the 25SX 'G' and the 20MHz planars cycled over to be PS/1 Pro's).
>

[Daniel Hamilton]

I actually have this system. Sounds like you do also... but if you need photos of it I can take some....

[IBMMuseum]

Hi Daniel,

> I actually have this system. Sounds like you do also... but if you need photos of it I can take some....

No, I don't have a PS/1 Pro 2123 - I have 'K' and 'G' submodels of the 25SX and realized the PS/1 Pro had the same planar (and a Model 30-style case) from pictures on a Facebook post (from "Hattan Man"). Then I saw UnalZ's BIOS notes that referenced a "30SX" with the same submodel coding. It would have been quickly eclipsed by the 35SX and 40SX (the 25SX "Starter Diskette" covers all three PS/2 models).

Does yours have 512Kb of VRAM (all four modules of ZIP chips) or 256Kb with sockets or solderpads? Are you able to get a BIOS date and version? I have an example image of a 25SX ('K' submodel) up at http://ibmmuseum.com/Planars/Test25SX.jpg (yes, I recognize it is dusty and will start cleaning other planars when I take pictures of them).

[Louis Ohland]

JP30-37 that's a jumper layout you don't normally see.

'92, uses a coin cell.

Fixed video oscillators. 25.175, 28.322, 1.5390 [???]

U33-36, U48-49 ???

[Tomas Slavotinek]

I have one of these undercover PS/2s as well. Type 2123-E31.

The PCB itself is exactly the same as in Model 25 SX afaik, just loaded
differently. Can't see the FRU sticker, it's probably under the DASD cage...

Only half of the video RAM is populated by default, U43 and U51 are
empty sockets, and obviously it has the external video connector (J15)
instead of the internal pin header (J17). In addition to U34/U36, it
also has 2 more RAM chips in positions U48/U49. Both SIMM slots MEM2/J4
and MEM3/J6 are available, but neither has a stick in it in my machine.
There are some other minor differences - 90° J16 pin header instead of
the straight one...

The BIOS ROM is dated 17/06/92, P/N 87F4794.

I'm slowly building PS/1 pages for my version of the Ardent Tool, but it
will take some time til it's actually publishable.

[IBMMuseum]

> JP30-37 that's a jumper layout you don't normally see.

I intended to clean the planar better to show traces - that is likely for a 9-pin COM port instead.

> '92, uses a coin cell.

I should check for what LSI chip has the RTC, and at what level...

> Fixed video oscillators. 25.175, 28.322, 1.5390 [???]

IDE and FDC connections are close-by too - I need to compare it to a 35SX/40SX (or just take another picture with my macro photo rig - I have a planar out)

> U33-36, U48-49 ???

'K' submodel with 1Mb planar RAM - 'G' submodels and PS/1 Pro 2123 have 4Mb planar RAM.

[IBMMuseum]

Hi Tomas,

> The BIOS ROM is dated 17/06/92, P/N 87F4794.

My 'G' with the same FRU is BIOS Model F8, Submodel 34, 01/09/1992, Rev. 11 - I have 'K's also with the same FRU I will need to test now. If you are able to test for a BIOS revision it is appreciated.

[IBMMuseum]

My 'G' has RAM populated for 'U11' and 'U13' too...

Despite the revision 11 BIOS, the 25SX 'Starter Diskette' (also shared with the 2.88Mb-capable 35SX and 40SX) shows that the floppy drives ('0' and '1') can be set at 1.44Mb or "Disabled" - With a 2.88Mb drive as "Disabled" it functions and boots from the drive.

[IBMMuseum]

> I need to compare it to a 35SX/40SX (or just take another picture with my macro photo rig - I have a planar out)

Image of a 35SX / 40SX planar now up at http://ibmmuseum.com/Planars/35SX40SX.JPG (cAsE sEnSiTiVe)

It appears I'll just need to wash the boards beforehand - I can also adjust lighting (at least the reflective metal crystal cans have frequency values silkscreened on the planar here) and I'm starting to learn the nuances to macro photography.

[Louis Ohland]

Ah, 41.5390

N82077AA

DS1287 :(

[Tomas Slavotinek]

On 20.8.2020 20:14, IBMMuseum wrote:
> If you are able to test for a BIOS revision it is appreciated.

Will do...

[Tomas Slavotinek]

Finally had time to fire this baby up:

Machine model: 0xF8
Machine submodel: 0x33
BIOS revision: 0x11
BIOS date: 01/09/92

[IBMMuseum]

Hi Tomas,

> Finally had time to fire this baby up:
>
> Machine model: 0xF8
> Machine submodel: 0x33
> BIOS revision: 0x11
> BIOS date: 01/09/92

Exactly the same BIOS version (and planar, with slight changes) as on my 25SX 'L' (I finally looked at the submodel to correct myself from saying 'G').

[richard smice]

Wouldn't it be nice if someone (like the folks that did the sound card)
did a video card for mca I use a ET6000 Tseng video card and mirage software
to run true color in my cross breed mca systems pci/mca system

[WBSTClarke]

There are some interesting video adapters in the German Treasure Trove, more later...