Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.sys.ibm.ps2.hardware > #71514 > unrolled thread

Incredible expanding ROM [increasing complex BIOS density]

Back to article view | Back to comp.sys.ibm.ps2.hardware

Contents

  Incredible expanding ROM [increasing complex BIOS density] Louis Ohland <ohland@charter.net> - 2026-08-06 22:48 -0500
    Re: Incredible expanding ROM [increasing complex BIOS density] "Josh Rodd" <joshua.at.rodd.dot.us@rodd.us> - 2026-08-08 14:46 +0000

#71514 — Incredible expanding ROM [increasing complex BIOS density]

This kerfuffle started by Joshua coincided with my evaluating a 8570-161 
that Elizabeth Holes turned me on to.

Turns out that it is one of TWO systems that use EPROMs that are 
compatible with Mask ROMs. 8570-161 [short sysboard] and the 8573-P70 
[late sysboard].

I then listed all the micro channel system ROMs and foundt that almost 
all MCA PS/2s use JEDEC except those two systems.

But I noticed a strange disturbance in the force. The T0- complex uses 
an AM27C020, which is a 256Kx8 ROM. All later T1 /T2 complexi [except T1 
upgrade DX2-66] use the AM27C010 [128Kx8].

According to AMD, the AM27C0x0 series is upward and downward pin 
compatible in density. As noted already, the T0- used a HIGHER density 
than the T0+ -AND- all EPROM-based Type 1. The datasheets say the 
AM27C0x0 series are compatible up to 8Mbit.

My SWAG is that IBM decided to split the POST from a single image into 
two, Stage 1 and Stage 2. MAJ Tom identified ROM in the T0-, not just 
IML... With the split of stages, IBM could use the lower density 27C010 
for stage 1 and put stage 2 on the HD.

With the incredible super powers granted to me by the white alkaloid 
poison known as caffeine, the walls flexed.

How about dropping in a higher density complex EPROM? Combine Stage 1 
and Stage 2 T1 complex BIOS into a single image and not need to do the 
convoluted bank loading / switching needed by IML. In effect, we have 
Flash ROM convenience where we can use IDE or RAID if we wandt.

If Joshua whips the refdisk into shape and slims it down, perhaps the 
lean 'n mean refdisk can be placed into the larger EPROM [say a 
AM27C080]. Instead of triggering the floppy, there could be a hook in 
the complex BIOS like "F1".

Although I am using the complex BIOS and the AM27C0x0, I am not sure if 
the other systems have EPROMs that are pin compatible with lower density 
EPROMs.

[toc] | [next] | [standalone]

#71532

My decompilation of the Type 2 Model 80 shows a certain delineation
between the E000 and F000 parts of the ROM, as if they eventually
planned to do this. I haven't yet touched the Type 3 or 4 machines, but
I'll eventually get around to a source level reconstruction of all the
ROMs. It's surprisingly easy to do, partly because the ROM isn't very
different from the XT 286 ROM (the most recent AT BIOS for which we have
source). The only real changes are addition of ABIOS, integration of
VGA, removal of any disk drivers (this is a Type 2 thing; the Type 1
could drive MFM/A cards, I think?), and setup code for a 386.

The 50/60 is even more worthy of analysis since it should be even closer
to the XT 286 code.

If we are able to get a good source reconstruction of the 90, 95, 77,
etc. ROMs, we could get a lot closer to fixing a lot of the more
irritating things about these machines such as the many quirks with IML.

Louis Ohland <ohland@charter.net> wrote:
> This kerfuffle started by Joshua coincided with my evaluating a 8570-161 
> that Elizabeth Holes turned me on to.
> 
> Turns out that it is one of TWO systems that use EPROMs that are 
> compatible with Mask ROMs. 8570-161 [short sysboard] and the 8573-P70 
> [late sysboard].
> 
> I then listed all the micro channel system ROMs and foundt that almost 
> all MCA PS/2s use JEDEC except those two systems.
> 
> But I noticed a strange disturbance in the force. The T0- complex uses 
> an AM27C020, which is a 256Kx8 ROM. All later T1 /T2 complexi [except T1 
> upgrade DX2-66] use the AM27C010 [128Kx8].
> 
> According to AMD, the AM27C0x0 series is upward and downward pin 
> compatible in density. As noted already, the T0- used a HIGHER density 
> than the T0+ -AND- all EPROM-based Type 1. The datasheets say the 
> AM27C0x0 series are compatible up to 8Mbit.
> 
> My SWAG is that IBM decided to split the POST from a single image into 
> two, Stage 1 and Stage 2. MAJ Tom identified ROM in the T0-, not just 
> IML... With the split of stages, IBM could use the lower density 27C010 
> for stage 1 and put stage 2 on the HD.
> 
> With the incredible super powers granted to me by the white alkaloid 
> poison known as caffeine, the walls flexed.
> 
> How about dropping in a higher density complex EPROM? Combine Stage 1 
> and Stage 2 T1 complex BIOS into a single image and not need to do the 
> convoluted bank loading / switching needed by IML. In effect, we have 
> Flash ROM convenience where we can use IDE or RAID if we wandt.
> 
> If Joshua whips the refdisk into shape and slims it down, perhaps the 
> lean 'n mean refdisk can be placed into the larger EPROM [say a 
> AM27C080]. Instead of triggering the floppy, there could be a hook in 
> the complex BIOS like "F1".
> 
> Although I am using the complex BIOS and the AM27C0x0, I am not sure if 
> the other systems have EPROMs that are pin compatible with lower density 
> EPROMs.
> 

[toc] | [prev] | [standalone]

Back to top | Article view | comp.sys.ibm.ps2.hardware

csiph-web