Question: IBM 6580 Displaywriter S/36 -15pin Keyboard

[MauriceH]

At the moment working on a 6580 S/36 system with dual 8" FFD-drives
Problem.....No Keyboard, damnn,We looked at that basement around but no keyboard to be found.

Must be a 15pin D-connector.
If any one has something useful laying around, let me know.

Try to hook up a XT keyboard by using +5V, Data, Clock, Reset, Gnd.
Think I need also a Kybd Acknowledge signal, as the system now hangs on Error 02.
First error code 01, Keyboard cable test , it now passes ,but error 02 is Physical keyboard test, it fails.

 

[Chuck(G)]

I believe that the keyboard protocol on the Displaywriter is very different from that of the XT. (note pins for B1-1 and B1-14, which the XT keyboard does not provide).
Too bad that most of the Displaywriter keyboards have been appropriated by the keyboard vampires and hacked to provide USB output, discarding the original circuit board.
Maybe someone has or wants to do a project that duplicates the original protocol.

 

[MauriceH]

Chuck, I noticed indeed those "Vampiers"Thought these boards were that obscure, seems though they manage to get it going.
That way I thought hook it up the other way,

Maybe I can draw some extra wire from a standard XT from that CPU. But can't find a proper manual at the moment.
Indeed strobe also. Think this writer needs to know when a key is pressed.

 

[Chuck(G)]

Indeed this will be a challenge. You might inquire over at the Deskauthority forum to see if anyone has one of the original PCBs from the keyboard available. At least that would give you a start with reverse-engineering.

 

[MauriceH]

So at the moment I found out the org keyboard has an 8049 cpu, the IBM XT has an 8048.
I manage to extract a wire Data Strobe from the XT keyboard and so far a bit luck I can manage to step through the Error codes and so manage
to get the DisplayWriter to get the IBM Start screen . Top.

But have no luck Booting with the disks I have, Couple of Org IBM disks and a copy set TextPack 4 1to3 disks.
Have the code 02 so that tells a bad disk. Just only reads track 0.
Only with 1 disk TextPack4/1 I get a very long error code starting with 9xx ending with CCCCCCC
Yeb cleaned the heads.
So some progress.

 

[Swamp_Donkey]

What is a displaywriter system/36 unit and how does it differ from a regular displaywriter?

 

[Swamp_Donkey]

You will have a hard time getting a non-displaywriter keyboard controller to work with the Textpack system because the keyboard contains the key map for the Textpack software, some diagnostics and also the system sound. Each keyboard has a number identifier associated with it for what language layout it has, and this is identified by the Displaywriter BIOS during POST and used by the Textpack software. The error you are getting is because the BIOS has not identified a keyboard number and Textpack has asked for what number keyboard is attached and the system replied null. I've emulated the situation you are facing by unplugging the keyboard on my unit and stepping through the POST diagnostics with the white button on the back of the unit and I got code: "*900* 00D0 CCCC C9E2 E2 003C 0044 1000 20 DADCFA CCCC" using a known good diskette of Textpack 3.
DOS and CP/M for Displaywriter do not use this Keyboard information however and you will be able to boot those programs if you have them.

 

[Al Kossow]

photos of the keyboard pcb are here http://bitsavers.org/pdf/ibm/6580_Displaywriter/photos/keyboard
a firmware dump http://bitsavers.org/pdf/ibm/6580_Displaywriter/firmware/4333923_KB_8048.BIN

there is a simulation of the keyboard in MAME

 

[Al Kossow]

the displaywriter has nothing to do with system/36

 

[Chuck(G)]

Looks like the 74165 furnishes the ID of the keyboard to the 8049. How the 8049 inserts the ID into a message isn't clear yet.

 

[Al Kossow]

Since I did my last work on the keyboard, I bit the bullet and spent way too much for a real keyboard
I need to verify it is in fact a 2K part instead of the 1K that I had assumed.
The mystery IBM part in the upper right corner of the schematic is probably the capacitive sensor

 

[Swamp_Donkey]

Yeah that's the 8 channel capacitive matrix sensor.

 

[Al Kossow]

I just dumped another microcontroller with a completely different part number, stamped 8048 underneath and
it matches what I dumped a long time ago, so the reverse-engineered schematic claiming its a 8049 is wrong

 

[Swamp_Donkey]

I have one with the same part number but what looks like an updated chip with updated IBM code. Plastic package instead of ceramic, stamped:
i 8493174 PI M P
L7200630 8719
c IBM 1984
ⓜ © INTEL '82

Significant, Al? Could this be the 8049? This info is probably of little use but I am curious on your thoughts.

 

[Al Kossow]

It's possible, the tell is the 1982 intel copyright
I don't know many people that would risk damaging a $1000+ keyboard to find out if the rom is any different, though.
I haven't actually opened up my working keyboard to see whats inside.

 

[MauriceH]

the displaywriter has nothing to do with system/36

Yeh thanks I made a mistake about that, was reading stuff about these old comps

 

[MauriceH]

A model F keyboard is on its way with 15pin connector (think of a Terminal IBM), and a Keyboard CPU Displaywriter board taken out of an bigfoot, as we know why.
That contains the org 8049 cpu.
So hope to construct that way a working keyboard.

As the Software, MSDOS Boots, CE12A1 diagnostics boots, and PDD12A1 also.
All Textpacks 1 and 4 fails with the long error code. Pack For I have org IBM 8"Disks.
I'll wait till I get the keyboard working and manage to do a system check with the system setup disks.
As I now Boot that disk at top row it mentioned 128k, have 4 MEM block installed on that MEM card.Strap is in the middle
So should have more mem available I think.

As for the 8049 I will dump the software, when it arrive.
HAve to see How I can readout that software from 8049

I'll be back

 

[MauriceH]

It is 192kB Present as MSDos found out after Boot.
Can't find any document about these mem block, or how many each block should be.

Do I need a system setup or check to store that way the amount of mem installed?
As I have a second mem-card I populated those empty sockets as well and took out that middle strap, as
I found on picture on the internet, with full populated the straps are all out.
But no difference at that mem count.

 

[MauriceH]

These are some fail-codes I got.Loading the TextPacks

 

[MauriceH]

As I Already had a 8"FFD Box of FFD's with that software I found in that box some documents also.
These are NOT from this system! So serial number does NOT match my system!
But it listed a 192k Memboard and an additional 256k? Or a second mem card?

SO if any one have the sn: 7176, I have the paperwork what was original installed.
Nice that org service card with all kinds of service notes when that machine was in service.