IBM 5170 AT 286 new bios

[Enrico Lazzerini]

I managed to find a bios for an IBM 5170 AT 286 Type1 Mainboard. It is
here: 30aprile1989 (list here:
http://ibm-pc.org/firmware/ibm/5170/5170.htm)
In practice I have manually added the number of cylinders, heads, and sect
/ trk that allow the machine to recognise a disk IBM 250MB. At this point
regularly boots DOS.
But I wish to program these data permanently in the bios. Does anyone know
what program to use and give me the procedure? I tried with BIOSUTL, but
the new EPROM do not boot the machine.
With BIOSUTL i made what follow: read actual BIOS, you can add new disk
geometry parameters at free 47 position, then you have to recalculate bios
checsum, then BIOSUTIL devides BIOS into EVEN and ODD file so i can finally
program them into two 27256 150nS eprom.
Thank you
Enrico

--- news://freenews.netfront.net/ - complaints: ne...@netfront.net ---

[gm1...@aol.com]

Enrico. You will need an EPROM programmer and that will have the correct software to burn the two EPROMs. The nasty bit is you need to take the BIOS and split it into two files, one all the even bytes and the other all the odd. Then you burn the two EPROMs. Just nmake sure you get them in the right sockets !

Cheers, Andy

[Johny B Good]

On Tue, 22 Jul 2014 05:42:33 -0700 (PDT), gm1...@aol.com wrote:

>Enrico. You will need an EPROM programmer and that will have the correct software to burn the two EPROMs. The nasty bit is you need to take the BIOS and split it into two files, one all the even bytes and the other all the odd. Then you burn the two EPROMs. Just nmake sure you get them in the right sockets !
>
>Cheers, Andy

That bit about the IBM AT using a PAIR of roms for the bios to gain
16 bit access intrigued me. I can't recall seeing anything shoddier
than the IBM PS/2 machines (which by rights should have been thrown
through the windows of IBM's offices by way of 'recycling the tat'.

Apart from the original IBM PC MoBos (2nd hand) I used to build my
very first PCs, the only other PCs have been built as clones using
MoBos made by third party manufacturers. My first upgrade using a
brand new MoBo was an 80286 33MHz clocked AT with its own built in
cmos setup menu (which set the standard in regard of my expectations).

I must admit this is so long ago that I can't really recall but I
think this used only a single eprom for its bios (at a 33MHz clock
speed, I suspect there was an option to shadow copy the 8 bit rom into
16 bit dram to compensate for the loss of speed due to throttling
access down from 16 bit to the 8 bit data bus of the rom).

Almost certainly, the quaint idea of using a pair of eproms to
improve access speed was a dead duck with the advent of the first
decent CPU to come out of Intel's fabrication plant, the famous 80386
and it's cut down cousin, the 80386SX.

I did google for the IBM 5170 AT to see whether or not the BIOS was
programmed into 1 or 2 chips but found no mention of such a
technicality. However, I landed up on this related web page:

<http://books.google.co.uk/books?id=uqInkwnBdacC&lpg=PA6&pg=PA55&redir_esc=y#v=onepage&q&f=true>

Which revealed a rather startling ad on page 157 by Microsoft with
the headline:

"I Fell 20,000 feet and lived" a reference to their flight simulator
and a chance to blow their own trumpet regarding their OS warez.

What really caught my attention, however was the penultimate
paragraph:

"The next time I flew, they had repaired the Sears Tower. My next
stop? The World Trade Centre.".

I wonder if this ad was the inspiration for Al Quaeda's 11th of
September attacks? It's a chilling thought, isn't it?
--
J B Good

[gm1...@aol.com]

I worked on the first PC's at IBM Greenock as well as checking and advising on the spec for some of the PS/2's that were for EMEA only. I am still using a model 80 at home and is the longest running PC I have had. The AT used two eproms as at that time there were no 16 bit Eproms available. Internally we had the AT board running up to about 12 Mhz. A few pull up resistors were required but that was all. On the later AT's we had to disable the speed check or it would not boot but I cant remember what had to be done. I have had a look through my old note books but cant find it.

Cheers, Andy

[fritz chwolka]

Am 19.07.2014 23:08, schrieb Enrico Lazzerini:
> I managed to find a bios for an IBM 5170 AT 286 Type1 Mainboard. It is
> here: 30aprile1989 (list here:
> http://ibm-pc.org/firmware/ibm/5170/5170.htm)
> In practice I have manually added the number of cylinders, heads,

>.....
>...

As my answers in cctalk seemd to be lost I post it here too.

It's not the best group for dos-problems but one of the best of the best.
:-)



Hi..

the biosutil maybe one way. The version I use told me:

Biosutil V1.1
InfoMatrix Bios utilities for the AT BIOS.
Brad Gibson Copyright (C) 1990 by Secret Software

and I believe there's no other version.


*----------------------------------------*

Other idea

It maybee better to add a enhanced bios to your PC.

There is the xtide bios at

https://code.google.com/p/xtideuniversalbios/

The xtide bios works on most xt/at.

You can easily ad an enhanced bios if you have an free rom place in a
network card like a ISA 3com etherlink.

As Im from germany and my english isn't as good as it should be here is
an explanation from

You should read it.

http://flint.cs.yale.edu/feng/research/BIOS/BIOS-report.pdf

The IBM BIOS gains much of its versatility by being an extendable BIOS.
That is, the full extent of the BIOS is not cast forever in the silicon
of the single PROM chip holding the firmware. The IBM BIOS can accept
additional code as its own into one integrated whole. Hence additional
PROM chips containing BIOS routines can be added to the PC.
The BIOS will incorporate these new routines.
The key for making BIOS extendable is a Firmware routine that enables
the BIOS to look for add-in code. During the boot up, BIOS code reads
through the address range that is set aside for firmware looking for
codes stored on add-in boards. If a valid section of code is found, the
instructions are added to the BIOS repertory. For instance a new
interrupt routine can be added or the functions of existing routines can
be changed.

During POST after interrupt vectors have been loaded into RAM, the
resident BIOS code instructs the computer to check its ROM memory for
the occurrence of the special preamble bytes, that mark the beginning
of add-in BIOS routines. The BIOS searches for these preamble bytes in
the absolute address range 0C8000 - 0F4000.

Here the XTIDE BIOS is found and added so you can use it.


Greetings

fritz

[glen herrmannsfeldt]

Johny B Good <johnny...@invalid.ntlworld.com> wrote:

(snip on even/odd EPROMs)

> That bit about the IBM AT using a PAIR of roms for the bios to gain
> 16 bit access intrigued me. I can't recall seeing anything shoddier
> than the IBM PS/2 machines (which by rights should have been thrown
> through the windows of IBM's offices by way of 'recycling the tat'.

All the AT and AT clones that I know of used two EPROMs. I think
it was the 386 and 486 days where I saw one, but note that by
then the usual OS did minimal BIOS access. Well, I was running
OS/2 on an AT clone, then upgraded to a 486 machine, again running
OS/2. I am not sure by now how much MS-DOS in the later days used
the BIOS, other than for booting.

> Apart from the original IBM PC MoBos (2nd hand) I used to build my
> very first PCs, the only other PCs have been built as clones using
> MoBos made by third party manufacturers. My first upgrade using a
> brand new MoBo was an 80286 33MHz clocked AT with its own built in
> cmos setup menu (which set the standard in regard of my expectations).

I only remember the 80286 up to 12 or so MHz. The 486 I remember at
33MHz and up. At some point, I had something with a really fast AMD
chip that fit into the Intel 486 socket.

> I must admit this is so long ago that I can't really recall but I
> think this used only a single eprom for its bios (at a 33MHz clock
> speed, I suspect there was an option to shadow copy the 8 bit rom into
> 16 bit dram to compensate for the loss of speed due to throttling
> access down from 16 bit to the 8 bit data bus of the rom).

I think some did that, but with Windows or OS/2, once booted there
is very little in calls to the BIOS.

> Almost certainly, the quaint idea of using a pair of eproms to
> improve access speed was a dead duck with the advent of the first
> decent CPU to come out of Intel's fabrication plant, the famous 80386
> and it's cut down cousin, the 80386SX.

I don't remember any 32 bit systems with four EPROMs.

-- glen

[Johny B Good]

On Wed, 23 Jul 2014 01:29:53 +0000 (UTC), glen herrmannsfeldt
<g...@ugcs.caltech.edu> wrote:

>Johny B Good <johnny...@invalid.ntlworld.com> wrote:
>
>(snip on even/odd EPROMs)
>
>> That bit about the IBM AT using a PAIR of roms for the bios to gain
>> 16 bit access intrigued me. I can't recall seeing anything shoddier
>> than the IBM PS/2 machines (which by rights should have been thrown
>> through the windows of IBM's offices by way of 'recycling the tat'.
>
>All the AT and AT clones that I know of used two EPROMs. I think
>it was the 386 and 486 days where I saw one, but note that by
>then the usual OS did minimal BIOS access. Well, I was running
>OS/2 on an AT clone, then upgraded to a 486 machine, again running
>OS/2. I am not sure by now how much MS-DOS in the later days used
>the BIOS, other than for booting.

Definitely by the time the 486 MoBos appeared, it was a common
feature in the bios to choose to shadow the roms into ram (16 bit wide
dimms had become standard by then so the rom region of the 1st MB
memory map had to disable that region of ram addresses to make way for
the roms - it was a no brainer to shadow copy the rom contents into
that otherwise mapped out ram space and have the rom code stored in 32
bit wide ram).

As you say, windows 95 and up made no use at all of the rom code once
booted but I think msdos still used the bios rom routines. The
shadowing into ram space considerably accelrated the performance
compared to running directly from the 8 bit wide roms (and they were
single 8 bit wide roms - I programmed enough replacement eproms to
know this for a fact).

>
>> Apart from the original IBM PC MoBos (2nd hand) I used to build my
>> very first PCs, the only other PCs have been built as clones using
>> MoBos made by third party manufacturers. My first upgrade using a
>> brand new MoBo was an 80286 33MHz clocked AT with its own built in
>> cmos setup menu (which set the standard in regard of my expectations).
>
>I only remember the 80286 up to 12 or so MHz. The 486 I remember at
>33MHz and up. At some point, I had something with a really fast AMD
>chip that fit into the Intel 486 socket.
>
>> I must admit this is so long ago that I can't really recall but I
>> think this used only a single eprom for its bios (at a 33MHz clock
>> speed, I suspect there was an option to shadow copy the 8 bit rom into
>> 16 bit dram to compensate for the loss of speed due to throttling
>> access down from 16 bit to the 8 bit data bus of the rom).
>
>I think some did that, but with Windows or OS/2, once booted there
>is very little in calls to the BIOS.

Almost certainly no calls whatsoever except probably only in 'safe
mode'.

>
>> Almost certainly, the quaint idea of using a pair of eproms to
>> improve access speed was a dead duck with the advent of the first
>> decent CPU to come out of Intel's fabrication plant, the famous 80386
>> and it's cut down cousin, the 80386SX.
>
>I don't remember any 32 bit systems with four EPROMs.

There weren't any. Being able to shadow the rom code into 16 or 32
bit wide ram did away with the need for 16 or 32 bit wide rom chips.
Shadowing would only cost a second or so, if that, of additional
POSTing time so was a no-brainer feature. Using a single 8 bit rom
(flashable or eprom) kept costs down without performance penalty when
shadowing was enabled.

I think it not being a default setting was more to help diagnose ram
issues if that region of the ram was faulty - the user, after hitting
trouble by selecting the shadow option only had to clear the cmos
settings back to default to get a working system so they could play
around with the ram (swapping dimms out or between slots or whatever).

I simply can't remember when this shadowing option first appeared.
Most likely as a result of the use of 72 pin DIMMs (16 bit?) which
started off at 1MB a dimm which meant the 1st MB of ram had 384MB
mapped out going unused. Shadowing just made better use of the
existing dram that would otherwise be mapped out.

386 and 486 MoBos would have been using 72 pin dimms in pairs so the
shadowing would have conferred a speed advantage without the need to
double or quadruple up the roms to make a 32 bitwide data bus. A
single large 8 bit eprom would (and did) suffice quite nicely for the
job, that much I _do_ know.

I could well imagine the use of a pair of bios rom chips (odd and
even addressed bytes) in the early 286 systems using memory made up in
two banks of discrete memory chips but I can't specifically recall the
use of a pair of rom chips (but then, they only came to my attention
when I wanted to reprogram the bios code in 486 and later system
boards).
--
J B Good

[Steven Hirsch]

On 07/22/2014 09:29 PM, glen herrmannsfeldt wrote:

> I don't remember any 32 bit systems with four EPROMs.

It may simply be early onset of senility, but I could swear I had a Mylex 386
motherboard with four EPROMs. Would have been ~1990.