If you have questions about Model M internals, ask now. Project progress here.

[pex]

The point of my Solutions page isn't to provide a bunch of useless trivia and progress updates and fun stuff, so I'll be providing that here to supplement my Solutions page.  You should probably read the page, though, to be familiar with what I'm doing.  http://solutions.sveit.com/modelm.html

Right now, I'm continuing to type in array numbers (i.e.: there are essentially 6 straight rows of membrane contact points on my Model M which I am just using reference [position] numbers for until I can figure out the keys they relate to.  Not all contacts are used!!!) into Excel.  I will then put them into a relational database so I can verify I didn't mess up in documenting them or translating them from my written sheet.  I can also then assign the key names to my array numbers, and start building graphical representations or program simulations.

From there, we can predict ghost/mask issues. I bought 50 IN914/IN4148 diodes today.  I think what I can do when I finish my documentation of the membrane is just get a component board and make my own small version of a keyboard just to see if the diode idea works.  Then I can figure out how to alter my Model M.

I've been sending out inquiries to different organizations about the PCB of the keyboard controller.  I don't know if it's really important for me to know anything about it, but I'd like a tech reference on the IC that controls all of this (part of learning about keyboards and their capacities.)

So far, using the data on the IC, and although it says Lexmark, net searching brought me to Zilog, an IC manufacturer.  They have no record of this IC based on the part number I gave them (emblazoned on the IC.)

I asked Lexmark to tell me about the IC that SAYS THEIR NAME ON IT.  They said 'um talk to IBM, it's THEIR keyboard'.

I asked Unicomp since they inherited all this.  They told me it was a proprietary masked [sic] ROM (apparently it's called mask ROM in the field) that has been discontinued (they use a different one in the Unicomp version of the Model M, apparently) and asked me to "get to the point".

I still have to contact IBM, but I'll bet they'll be a noshow/standard-response.  I mean, jesus christ, you have your copyright and your patent, and there are just not THAT many things that little IC could be doing!  I just want to know what you label the leads!

I may finish the paper-to-Excel translation tonight.  That would be grand.  More to follow (including messages from companies.)


Anything you want to know about my Model M...any pics you want to see, the way things work...ask and I'll see if I can accommodate since it is all in pieces right now.

[pex]

Correspondences.

Zilog customer support website (zilog.com/whateveritis)

     Support               Service Requests      

   
Incident ID
   81337
Description
   Integrated Circuit identification - 1428578
Type
   Misc. Inquiry
Product
   3 - Tech Documentation
Priority
   3 - Medium

Detail Log:

**** Entered By: OCPLead @ 2008-01-29 00:36 ****
I have a p/n 1428578, 40-pin DIP on a keyboard controller and the sparse information I have found points to ZILOG as the manufacturer of this integrated circuit.  Could you tell me if you have ever manufactured such a part, and if so, please provide a data sheet and technical reference for this part.


**** Entered By: nreyes @ 2008-01-29 16:15 ****
Dear John,

Thanks for contacting ZiLOG!
Please help advise what would be the equivalent zilog part number of your P/N 1428578 so as we could help you.

Appreciate your prompt response.

Best Regards,
Zservice


**** Entered By: OCPLead @ 2008-01-30 09:19 ****
This part was made somewhere between 1994 and 1998, and I wasn't sure if it was a discontinued or one-off part not listed on your normal parts page.  Because I do not actually know much about the part other than its relation to all other components, deciding on an 'equivalent' zilog part number, when I thought that actually was the zilog part number, will be troublesome.  If zilog made no part with that part number between 1994 to 1998, I may be searching in the wrong place.

Lexmark (through their stupid printer-only website)

Dear John,

Here is your Service Request # 1-791277321

Please accept my sincere apologies for the delay in our response to your e-mail.  We make every effort to respond to each e-mail within 1 business day, but in this case we have failed to do so.  I am truly sorry for any inconvenience this may have caused you, and I appreciate your patience in this matter.

I kindly request you to contact IBM technical support as the keyboard was under IBM's name. IBM Corp. number 1800-772-2227.

If you have any more questions or concerns, please contact me at your convenience and I will be happy to assist you. (If I am not available, another representative will reply to you as soon as possible.) To respond, please select "Reply" in your e-mail software, and be sure that the past e-mail is included in this reply.

[AOL Users: In order to include the previous e-mail, you must highlight it with your mouse when you are replying.]

If you need to reply, please be sure to include in your message all information from prior e-mail messages & replies. If your e-mail client automatically deletes prior e-mail thread information, it will cause a delay while we look up your support history. If this is the case you may want to save the old e-mails as attachments and attach them to the current e-mail.

Sincerely,
Fahad
Lexmark eSupport Team
http://support.lexmark.com
[THREAD ID:1-D33THL]

Rate Your Support Experience! Your feedback is extremely valuable to us. Please click the link below to participate in the Lexmark Customer Support Survey.

https://surveys.lexmark.com/survey/s?s=9536


**********Original Message**********
This inquiry has nothing to do with printer Lexmark X5470, because your support system is unfortunately too limited to deal with this request.  In previous years (a decade or two ago), Lexmark, in partnership, license, or buyout of IBM, created Model M keyboards or parts for keyboards under the IBM name.

I need to get a datasheet and technical reference for an integrated circuit found on the keyboard controller of a keyboard I have.  It is labeled as follows:



1428578

(copyright)LEXMARK-94

9802     NT



It is a 40 pin DIP.



My keyboards has the following identifying features:

p/n: 42H1292

s/n: 2169838

Date: 18-03-98 (uk date order of manufacture)

EC: D41496

Made at...IBM United Kingdom Limited, Greenock, Scotland, UK



If you need any more identifying information to narrow down and identify the IC, let me know...I have covered every internal and external inch of the keyboard for information that may be applicable.



Thanks for taking on this unusual request.

Unicomp:

The controller you are asking about is a masked ROM part which was proprietary to Lexmark and Unicomp. We do not have a technical datasheet for this part. This part is no longer manufactured and has been replaced by a newer masked ROM part. What specifically do you need to know?
   
  Keyboard support

   
 
Jim Owens wrote:
  Neil, Anything we can tell this fellow?

hotmail.com wrote:   From: <@hotmail.com>
To: ,

Subject: 40-pin DIP p/n1428578 on '98 UK 42H2192 keyboard controller
Date: Tue, 29 Jan 2008 04:18:55 -0500

In previous years (a decade or two ago), Lexmark, in partnership, license,
or buyout of IBM, created Model M keyboards or parts for keyboards under the
IBM name.
I need to get a datasheet and technical reference for an integrated circuit
found on the keyboard controller of a keyboard I have. It is labeled as
follows:

1428578
(copyright)LEXMARK-94
9802 NT

It is a 40 pin DIP.

My keyboards has the following identifying features:
p/n: 42H1292
s/n: 2169838
Date: 18-03-98 (uk date order of manufacture)
EC: D41496
Made at...IBM United Kingdom Limited, Greenock, Scotland, UK

If you need any more identifying information to narrow down and identify the
IC, let me know...I have covered every internal and external inch of the
keyboard for information that may be applicable.

Thanks for taking on this unusual request.




Jim Owens
Unicomp, Inc.

[pex]

Excel transfer is done.  Making an Access database now (probably for future export via CSV and rebuild into SQL when convenient or needed).

Building database...am already finding some consistency issues.  Rechecking my paper-to-excel conversion.

[pex]

So...I'm retarded.

Not sure why I didn't do this before, but I got some 17x11 sheets of paper and I'm going to make a diagram of the keys and the contacts to ease troubleshooting (would have eased figuring out the matrix too! dang!)  Maybe I'll get some pix up of this today or tomorrow (and finish up my DB.)

[pex]

I'm still around, in case anyone wonders.  Haven't accomplished much lately.

I'm looking around for a piece of glass I can put over a lamp to do my tracings since I found my traces of the key paths weren't accurate (probably 95% accurate, but this new paper system will speed verification immensely.)

The quest for n-key rollover is still on!

[iMav]

Glad to see you are still around.  

[pex]

So, I finally got a piece of glass...probably 17x24, nice size, a bit flimsy however but will be plenty sufficient.  Someone was discarding it (is seemed to be from a picture or canvas frame?) so I took it home and cleaned it up...



THEN I REMEMBER I HAVE ABOUT 17 STORM WINDOWS LYING AROUND THAT WOULD HAVE DONE JUST SWELL.

Jawsome.  Time for me to get back to work!

[outofideas]

I was troubleshooting a Model M a while ago and found a site that had the matrix documented.  It's since disapeared, but archive.org has a copy here.  The matrix is near the bottom of the page.

I also ran across this page which details a method to put a Model M back together with screws after cutting off the original plastic rivet thingies.

The Model M I was troubleshooting had broken traces, verified when I took it apart.  I did fix the traces with conductive paint.  But I haven't tried putting it back together again yet.  For one thing I don't have the screws I need just lying around.  

At this rate I'll probably never get around to it.  Especially since I've been eyeing one of those Filco Keyboards.  The lack of n-key rollover is about the only thing that I find lacking in the Model M.  Occasionally having to remap keys to be able to press 3-4 buttons at a time is annoying on an otherwise perfect keyboard.

Also I've been using software(compiz) that's easier to work with, with the otherwise hyper-annoying "windows" keys.  So that'a another reason to try a Filco or something similar.  I could get a 104-key from Unicomp, but that wouldn't solve the rollover issue.

Good luck with your project.

[iMav]

Welcome to geekhack outofideas.  I believe we had a thread around here somewhere regarding Model M repair and had linked to Sandy's page.  But it has long since been buried.  Thanks for mentioning it...definitely good info.

[pex]

I did a recheck on the matrix because my first ran had some data integrity issues.  I should be able to move onto testing-engineering now, yay.

Interesting is this information:
http://web.archive.org/web/20070713030833/http://pupp.edgeemu.com/kbhack.htm

Which defines the matrix of the follow type of Model M:

IBM
Part no. 1391401
S/N No 2051477
Date 12 mar 87
Plt No F2 Model M Matrix: 16x8
Connector: ps2

I have 24 board connects on my test Model M, but it is a 42H1292, made 1998-03-18, and the controller board is set up completely different.  I now have a 1391401, so I will be able to compare my results to what is on that page, but I do not think the matrices are the same for the 42H1292 I have and the 1391401 this other page presents.

Todo from here is just do some quick wire bridging to make sure I can just go with the controller, and then I'll probably get a perfboard from radioshack and set up a diodes test with select matrix areas.

If the diodes test goes well, then comes the structural engineering issue.

[pex]

Tried to do some testing with QWAS.  First confirmed that QWAS was a bad combo by using a bunch of aluminum foil to link together the 4 terminals appropriate for QWAS, testing each separately.

Then I foiled all of them together to make my computer beep because it can't complete the combo.  I also tried lesser combos to make sure combos worked.  They did.

Enter: diodes.  1N914-type (item 276-1620 pkg of 50) They are supposed to be the ones recommended on this page (1N4148):
http://www.dribin.org/dave/keyboard/one_html/

Ideated how the matrix would work...verified I could still get connections behind the diode.  That worked.  However, no combination of single-keypress dioding allowed me to capture ANY keypresses.  That doesn't make sense (I should at least be able to capture a keypress from a diode in the right direction!)

This says to me that maybe the diodes are not the appropriate type?

[pex]

Email to Dave Dribin to find out if he has any field reports of the information he provides on his page being successful upon realworld application:

Date: Tue, 05 Aug 2008 07:01:52 -0400
From: Solutions Master <@gmail.com>
To:  @dribin.org
Subject: Keyboard matrix theory and application to real hardware: diodes

Dave,

I'm trying to do some enginnering to prove whether or not it really is
as simple as dropping diodes on an existing keyboard matrix to alleviate
ghosting/blocking.  My searches indicate that information on this topic,
and the successful application of the information, is not very available.

What I need to know from you is whether you have received any field
reports of diode drop-ins as being successful when applied to existing
keyboard layouts.  Preliminary testing on one keyboard indicates that
either that specific keyboard model, or the diodes suggested on your
website, is not appropriate.  Further testing will be done to make sure
this does not become an erroneous conclusion.

If you are interested, you can keep up with the project:

http://geekhack.org/showthread.php?t=289
http://solutions.sveit.com/modelm.html

- pex

[lowpoly]

While I'm really interested in what you're doing here I have a hard time following you. Maybe it's because I'm not a native speaker, I don't know.

[iMav]

Thanks for the update pex!

[pex]

Message from Dribin

I have not received any information on if it is possible to add this  to existing matrices.  I don't see why it wouldn't work, assuming you  have enough access to the wires/contacts.  I imagine it would depend  on the particular keyboard.  FWIW, it has been used in some new  projects and it seems to work for them:

 
 

-Dave

Interesting linkbacking those projects are doing.  

I'll get some perfboard and components over the next few days and try to figure out what I'm doing wrong and they're doing right.

[pex]

I got a perfboard.  Because I need a new solder tip and have nearly no good experience soldering, I thought I'd make a set of aluminum-cellophane contacts.  (Single layer of the normal alum foil on the contacts of the board with cellophane tape for structure.)

That doesn't seem to work.  I can get some voltages over it, but if I try to perform a key press, it fails.  I can still form a keypress by completing the circuit with the points on the board.

While aluminum is conductive (down the line from silver and copper but still good), the aluminum foil can't handle it.  It didn't burn up or anything, and continuity tests pass.

I will try taping the wires I'm poking into the perfboard for better results.  Unfortunately they are round and harder to secure.

EDIT: now i'm having trouble with the wires.  jesus christ.

i think i will have to solder
sux0r

[pex]

No good update info right now, but I did want to revive this thread because I have more freetime now and I know where both my components and my data are.  I also have the tools to do solderwork.

Maybe I can get the controller soldered for some manipulation this week, and do an honest diode test.

[iMav]

Glad to see this project is still alive.

Keep us updated!

[xsphat]

What are you doing to this Model M? I don't know what you goals are.

[bhtooefr]

N-key rollover on a Model M, IIRC...

[pex]

Snap-in dioding is not working.

I tried expecting the electrons originating from 'rows' or 'columns', and switches the directions of the diodes after the switches.  4 different combinations resulted, therefore, two of which allowed me to 'type' single characters but not reach "QWAS" without my computer beeping, and the other two preventing any recognition of keys.

I presume that pointed in the direction allowing no flow after the switch broke the circuit.  Otherwise, the diode was not effective as stopping backflow for whatever reason.

Examine the following link for a pictures walkthrough.
My electronic playground (199 KB)
Faithful reproduction of Dave Dribin (220 KB)

I'm hoping someone with slightly more electronics knowledge can see enlightenment here. Otherwise, MYTH BUSTED!  I realize my setup is a mess, but if someone thinks I need to take multimeter readings, I can do that after the arduous task of trying to make a cleaner playground.

[pex]

What are you doing to this Model M? I don't know what you goals are.

N-key rollover on a Model M, IIRC...

I have two goals: implementing n-key rollover on a keyboard that has no such feature, and to fix my particular keyboard.

This is a Model M I picked up off a pile of trash.  It doesn't work completely correctly, but the issues it has are consistent and quantifiable.  They should not affect my QWAS test.

[wellington1869]

Here's a question about model m internals - is it relatively easy to replace or swap the spring in a buckling spring stem? (For instance if you wanted to increase/decrease the resistance)

[pex]

Here's a question about model m internals - is it relatively easy to replace or swap the spring in a buckling spring stem? (For instance if you wanted to increase/decrease the resistance)

This question is currently being discussed at: http://geekhack.org/showthread.php?t=1491