LCD-286 Keyboard Repair

[Ellron_Cupboard]

Hello All,

I've been working on repairing a MicroExpress LCD-286 i got a while ago. It came to me very dead. At any rate i've got the hardware mostly sorted out and finished the keyboard tonight. It took a lot of searching on the web to get it repaired so i thought i would summarize it all for anyone else who may have a similar issue later on down the road. anyways here goes:

So the keyboard is a Parex Modern-1 (A60-0014-00). It used what i can only assume was an off brand D8049H micro-controller. labeled (" IC C36081E ; P3A7R6E2x1" date code: 8725). At any rate, the keyboard is selectable AT/XT variant. In AT mode, the keyboard did not function at all. But in the XT mode it sort of did. the LCD-286 is an AT machine so this board needed repaired. Basics were checked first, like continuity across the AT/XT selector switch,etc. all hardware seemed fine. After a few days of tinkering i started to think that maybe a kid had disassembled the keyboard and moved AT/XT switch while the board was powered up possibly goofing something up or the chip had just aged.

i looked around for some options and decided getting a few D8749H IC's and a Willems programmer was the way to go. Fast forward a week later. i'm reading off the original IC's mask rom successfully out. Thinking i'm in good shape and maybe the original chip was just bad.

Spent quite a bit of time tinkering with the Willems programmer trying to get it to burn the D8749H chip with no success. Turns out you need to solder a 5NF cap and 1Meg Ohm resistor in series between pin 20 and pin 25 of the 40pin MCS-48 adapter board zif socket ( so between VSS and the PROG pins ). along with setting VPP on the programmer to 21V via J6 and J7 and Vcc to 5V6 Via J5 (memory on this one). OK so D8749H has the 8049H mask rom on it. installed a dual wipe socket on the keyboard, fired it up and nada.. in-fact its even worse then the original chip.

Time to rethink what i'm up to..... then it occurred to me, i've got a Dolch 486DX2-33 luggable with a cherry mx blue keyboard. identical layouts , maybe... just maybe.

I've attached a few pictures of the keyboard and microcontroller. The Willems MCS-48 mod needed for programming the 8749H microcontroller and the non working ROM, the decompiled assembly code from the 8049H (in case anyone ever wants to tinker with it).

I'll leave the rest for tomorrow so i can get some sleep.

 

[Ellron_Cupboard]

So my Dolch 486dx2-33 has a similar keyboard. I opened it up and found that the boards looked identical same model number but what i can assume is the 7th revision of the board. A60-0014-07. I desoldered the 8049H chip and read out the rom, burned it onto a D8749H and put it in the lcd-286 keyboard; note, i also added a zener diode between pins 21 and 25. powered up the computer and entered the keyboard diagnostics on the bios. to my surprise the keyboard worked like 95%. a few keys were actually swapped though. like the control key and the c key were backwards. plus i got a keyboard error message on startup which was kind of iffy.

i kind of took a step back from it and decompiled the A60-0014-07 rom code into assembly and ran a diff between the two keyboard roms. up front it looked pretty promising but they are actually quite different. although thats probably relative because binary can be decompiled different ways or written different ways. anyways, i thought about this for quite a while and decided it probably wasn't worth the mess digging out my old books to sort through this line by line and isolate the issue. maybe for someone who writes assembly everyday.

well that kind of took that option out of the mix.

i figured the best way to get this working was to modify the PCB and attempt to bring up the A60-0014-00 board to the A60-0014-07 spec. mapping out the differences it ended up being about 6 keys that needed traces cut and jumpers soldered in. the traces on this board are pretty big so if i ever need to get it back to original ... a little scraping , soldering and solder mask would do the job. Ended up going this route after thinking about it for a while and to my surprise everything worked out fine, no keyboard error and every key worked as it should.

so at this point I'm done with the hardware modifications and am moving on to cleaning and retro-brighting.

i've attached a few images of the A60-0014-07 keyboard and the respective ROM and AST files.

hopefully someone down the road finds this information useful, i think the ROM files hold some real value potentially.

I wasn't able to find hardly any information on the Parex keyboard or the various versions aside from this link:

https://deskthority.net/wiki/Parex_Modern-I

most of the stuff on the web is about how people desire the keyboards for the cherry mx blue keys,etc. but nothing too technical.
i had considered using a kitten paw microcontroller or a teensy but honestly i wanted to stick with tried and true stuff while preserving ROM files and the knowledge.

it would be cool if someone gave that rom file from the A60-0014-00 a look over... maybe it can be saved. i do believe theres an emulator for the MCS-48 microcontroller family within MAME.

anyways, cheers.

 

[musovern]

Hi there,
Nice write up on your keyboard.
I've got a LCD-88 portable 8088XT clone that I now have working and running but the keyboard is not working it looks very similar to yours.
I've unsoldered my P3A7R6E2X chip and tried to read it with my needham emp20 programmer but seems like it dead.
I am going to try burning your ROM files to some 8749H chips I have and I'll test in on a AT machine first with the switch in that settings.
What is the zener diode did you use between pins 21 and 25.

Cheers

Vernon

 

[Ellron_Cupboard]

Hi Vernon,
Im very late to the game but, I went and looked at the keyboard and it appears*(would need to desolder the diode to be 100% ). looks like an 1N4148. Must've had a typo when I wrote this up. Hope you got it working. I kind of vaguely recall reading the white papers on the microcontroller for a keyboard application and think that's where I came up with the diode. Memory is quite rusty at this point though. But the keyboard has worked flawlessly since I did the repair.