cpc keyboard fault testing

[arnoldemu]

Here comes the info, MacDeath style 

!!!!!As always be careful, if you're not sure, don't do it, because there is a possibility you could do something wrong!!!!!!

for keyboards where some keys do not work.

problem for cpc+, if f1 key doesn't work, you can't test the keyboard this way, because you can't go into basic
with system cart.


cpc/cpc+ (if f1 works):

switch computer on and press every key
see which keys do not give a response
(turn off caps lock, press shift and a letter to see if it works.
press ctrl and a letter to see a different char.)

next consult the keyboard matrix diagram.
(keyboard document on www.cpctech.org.uk)

If all keys are horizontal, then 1 bit is bad out of the data.
if all keys are vertical, then 1 line is bad out of the data.

taking apart you can then check if the problem is with the keyboard or not.

if you have a spare keyboard with the same connectors try this. This may help to eliminate keyboard quickly.

for plus:

remove cart, remove all screws, push triangle tabs,
turn back over.
lift top off.
insert cart into left side (pcb that stands up) carefully
remove keyboard from pcb, lift away.
connect power and monitor.
switch on using switch on top piece.

two connectors (looking as if you were typing on keyboard).
top is LINE. All 10 are used.

bottom is BIT. 1 on either side is not used, so we have NOT USED, 8 x USED, NOT USED.

Left to right is line 9 down to 0. So testing joystick, down to cursors.


Testing keyboard via the connectors:

You can connect a piece of single core wire between the top and bottom or left and right side connectors.

You can fix the the single core wire between the metal part and the bracket.
then just touch the bottom one and see a key press appear.

For each line you do, you must then do all of the 8 BITS. (e.g. for plus, for each of the top you choose, you must do each of the 8 bottom ones with it).

If all keys produce a result then keyboard is not functioning.

If one or more keys does not produce a result (remember shift and ctrl must be pressed with another key).

Then if you find a fault, try connections using a multi-meter and continuity tester.

- if a row is bad, you need to check connections to:

CPC: 74LS145 (ic101 in cpc6128)
2,11,12,13,14,15,16,17,18,19 (y10-y1)

and from this 12,13,14,15 to 17,16,15,14 of 8255

in CPC+:
ky1-ky9 80-71 on IC101 on AMS40489

- if a bit is bad, you need to check connections to:

AY-3-8912, (ic102 in cpc6128, ic114 in CPC+) port A.
7,8,9,10.11.12.13.14. (bit 7..bit 0)
same for cpc and plus

for cpc6128:

turn over keyboard. remove screws.
turn back again and lift up keyboard.
detach keyboard.
lift it and position it upside down but above base.
connect power
switch on.
connectors are to the left.
leftmost is line, rightmost is bit.
Testing is the same as for plus.

cpc664:

Removing keyboard from pcb is very difficult. I would not advise.

for cpc464:

!more difficult!, power switch connects to cassette which then connects to pcb.
Also watch out, different keyboard designs, the follow is for membrane keyboards.

turn over
remove all screws
turn back.
lift off keyboard. detach keyboard and cassette.
then remove keyboard so you have a hole.
connect cassette back again and put case back on without keyboard.
access to connectors will be difficult, not sure if it's even possible. <- to be checked.

[Gryzor]

Niiice! Shall I copy it to the wiki? That would be its natural home...

[arnoldemu]

Niiice! Shall I copy it to the wiki? That would be its natural home...

not yet.
I've got 2 amstrads with keyboards that don't work.

a 464 (costdown) - this one seems to have a single bit not working on almost all lines
a 464+ (this one has various keys not working, mostly column, however this computer *was* stored with something leaning on the keyboard - so it's my fault really).
But once I've fixed or got furthur with these I'll do a proper write up for the wiki.

[Gryzor]

Cool Please don't forget!!

[arnoldemu]

I took apart my CPC464 costdown.

I found that if I turned over the pcb, then I could easily see the connection between the right keyboard connector and the AY
chip. I checked all the contacts for connection and all checked out ok.

For the costdown it appears AY pins 17,18,19,20,21,22,23,24 are connected to the right keyboard connector.
The left keyboard connector is connected to the 74LS145 to the other side.

So this seems to indicate that the membrane is broken, but really I can't tell even after tracing the expected line.
(from the F4,F5,F6 back to the connector).

I guess that if the AY was faulty, then I would expect some other strange problems with it.. so the next thing may be to write a program to check it's OK. Writing some data to it's ports, writing to it's registers and reading them back just to eliminate this.

So assuming it's the membrane, is there anyway I can check for continuity working just from the ends?
Is there any way I can simulate a key press and identify where the broken part is?

Looking at the membrane, I assume the two contact sides are inside, and that the key actually presses the two contacts together. So if I split the membrane (which would be a crazy idea, especially to get it back together again, how can I find the problem?

Anyone got any ideas or can help?

[arnoldemu]

A suggestion from Devilmarkus is to carefully split the membrane into two. Clean both inner surfaces and then stick back together again. So I will probably attempt this first with my 464+ keyboard (which really seems the most broken - see if it sorts out *some* keys, and then do the same on the costdown 464 keyboard.

[arnoldemu]

A suggestion from Devilmarkus is to carefully split the membrane into two. Clean both inner surfaces and then stick back together again. So I will probably attempt this first with my 464+ keyboard (which really seems the most broken - see if it sorts out *some* keys, and then do the same on the costdown 464 keyboard.

I did this with my 464+ which has the worst keyboard.
I carefully split the two membranes and could do this with little force.
I cleaned the contacts just with a bit of spit and kitchen towel and carefully put them back together ensuring the dots lined up.
It didn't make a difference, in fact it seemed worse. I'll give it another go with some other cleaner and perhaps try the multimeter with it.
I tried to take a picture of each half, but they didn't come out well

[Bryce]

If you have a multimeter, you should be able to test the continuity from each pad back to the flat cable contact, but don't just poke the sharp point of the probe against the foil, otherwise you could damage the foil further. Use some sort of self-made rounded surface on the end of the probe. Sometimes the foil gets tiny barely visible tears. As soon as you know where the tear is it can be easily repaired.

Bryce.

[redbox]

I recently took my 6128 Plus keyboard apart to clean it all and had some problems putting it back together.


At first I thought I'd knackered the membrane because I was getting multiple key inputs when pressing one key - for example, if I pressed Q or S then W was showing as well.  I followed the keyboard membrane input tracks that arnoldemu was talking about and seriously thought I had scratched one off or something and this is why I was having problems.


However, I took it all apart again to double check and found that some of the springs weren't seated exactly in the middle of the plastic arm of the key holder underneath.  This was the case for all of the keys I was having problems with.  I could find the keys that were doing this by booting into BASIC and putting the minutest of pressure on each key - the ones where the spring wasn't seated would start printing characters on the screen and the ones that were okay would do nothing.


Worth checking before you panic and think you've wrecked your membrane 

[Vorian]

I need advice for troubleshooting keyboard issues on a plus.

I have no input from line 0 keys ( F9,F6,F3,F.,enter, right, up, down ). All other keys work.
Using another keyboard I get input from line 0 keys.
Would I be correct in saying I have a membrane issue?
I have looked over the membrane and nothing jumps out at me to indicate an issue.
I'm thinking new membrane?