Topic: Bizarre C128D keyboard problem

[orb85750]

Suddenly, a group of keys (letters) has stopped working on my C128D keyboard.  However, they sort of work with the shift key on.  Actually, they double print when I have the shift key on.  Anyone have a clue what has happened to my beloved C128D tonight?  I probably won't sleep all night.

[Hydrophilic]

This sounds like one of the lines of CIA#1 or perhaps the VIC-IIe have gone bad.  You didn't say which group of keys, so take a look at a keyboard map (here's one) and see if your problem keys all lie in one row or all in one column.  If so, it's a good bet that line is some how damaged... I suppose there are multiple ways it could happen, but it seems the most common cause is static discharge by touching one of the joystick ports.
 
If that is the case, then hopefully it is the CIA that is bad, because these are fairly easy to obatain.
 
Also, because pressing SHIFT seems to make the keys work, it sounds like the line is not completely dead, just 'weak'.  In that case cleaning the keyboard may help... but it may not... Even if it does help, it may be just a matter of time before 'weak' line dies completely.
 
You probably already know this, but joystick/mouse/etc can cause problems with the keyboard, so unplug any devices and see if that helps.  Anyway, good luck!

[Antti]

I had a problem with one of my joysticks which caused similar problems. Didn't try with shift key, though. So unplug joysticks and see if the keys start working again.

[orb85750]

Wow, thanks for the advice.  I was using a CD32 competition pro pad in the port.  I just unplugged it and everything works fine again (somehow it did work fine for a while the pad plugged in yesterday).  Well, I won't use that pad again -- hopefully no trouble with other joysticks. 

[crrn]

Hello
I have a simmilar problem like described on first post. I figured out that keys that dont work are the ones from first row (F1, F3, F5,F7, Ins Del etc)
I also tried to change a CIA chip which I took from old disk drive. My C128D boots nice but still dont work and effect is still the same - the same  keys not working.
Is it CIA or VIC then?

Also, I dont know if it is connected to the same couse but th1 1571 in my C128D doenst work. When typing DIRECTORY directive only result is a LED flashing but the header is not mooving and no noice is made by disc drive.
What might be the problem?

[Hydrophilic]

Those keys you mention are all on row 0 of CIA, so it sounds like the problem is CIA (it seems the the VIC is okay)
 
The only other electronic component that I think that could cause the problem is the EMI arrestor (one for each line) between the CIA and keyboard connector.  These little things look like a diode but they have 3 leads instead of 2.  You should see a whole bunch of them (22) near the keyboard connector.
 
It could be the EMI arrestor (or whatever its called) is bad, although I've never heard of that happen.  Row 0 of the keyboard maps to port A bit 0 (CIA pin 2) and keyboard pin 13.  You could use a continuity tester (with C128 power off) to check the line is not broken.  If you don't get continuity, then the EMI arrestor is bad or there is break in the line on the cirucuit.  However both seem unlikely.
 
It seems more likely there is something wrong with the keyboard itself.  Maybe the wiring is bad... or maybe a small piece of metal or some liquid got inside the keyboard.
 
Anyway be sure the problem is not caused by something plugged into a joystick port, like a mouse with the right-button stuck could cause the problems you mention.
---------
So you put the CIA into the 1571 and now the drive head does not move?  That IS strange!  The CIA is only used for fast-serial communication.  The stepper head / motor / track sensor are all handled by one of the two VIAs.  I don't why you would get those results unless maybe you swapped a CIA for a VIA !!  But if you did that, it seems like the C128 would not work... really strange!

[crrn]

I'll check keyboard tommorow morning. open it and see if its OK. But I didnt mention that in the C64 mode keyboard works fine.

Regarding the 1571. I took CIA (MOS 6516) from other old/broker 1751 i have. The one in C128D has all the original chips/parts but the head is not mooving while the DIRECTORY directive is typed. So first question could be how to determine if the electronics or moving header parts are broken?

Regards

[Hydrophilic]

Keyboard works fine in C64 mode... that is strange... the VIC keyboard lines are not used in C64 so that makes me think it could be the VIC, but the VIC lines don't have anything to do with F1, F3, F5, etc... maybe if the VIC lines were 'stuck' it could cause problems with function keys, but then it should cause problems for ALL the keys... really weird!
 
As for the 1571, bunch of things could be wrong.  The motors run off of +12V so if that is bad (but +5V is good) then the chips would work but not the motors.  Of course the stepper motor could be burnt out (or the transistors that drive it).  There should be a trouble-shooting chart somewhere on the web for that...

[crrn]

OK, So I used the compressed air to clean the VIC areas on the mainboard and also to clean the keyboard.
Now it seams that the kesy work, but from time to time pressing return causes to print different chars (mostly char '8') but when i press it harder then it works fine for some time. the same is fo F1-F7 keys. So I can assume it could be a keyboard problem...

but the 1571 is still not working. and i actually start to think there must be something more complicated broken with it. I was thinking of using the electronics from my spare 1571 disk drive but the board is a bit different - especially that is has an different power supply adapters (well its taken from standalone 1571 drive).

[Hydrophilic]

The RETURN key is on the same row as the F-Keys of the keyboard, but the number 8 is not on the same row or same column of the main keyboard.  However, the number 8 of the extended keyboard (VIC line) is in the same column as RETURN.
 
So here is a theory.  I noticed when testing MP128 that the extended keys work and trigger the same effect as keys of the main keyboard (for example pressing 1 and 3 on the numeric keypad will active the response as SPACE and STOP of the main keyboard).  Which is because the IRQ that I wrote doesn't touch the extended keyboard lines of the VIC; that is, it leaves them in the same state that the KERNAL left them in...
 
So it seems that the KERNAL leaves the VIC keys active at the end of the IRQ routine, and when the next IRQ occurs it starts over by testing the main (CIA) keyboard rows.  If you notice in the keyboard layout that the RETURN / F-Keys are in row 0 of the CIA.
 
So anyway, it may be that the VIC keyboard line 0 is partialy stuck, which is to say, it is slow to rise to +5V from the previous IRQ of 0V when the new IRQ starts to scan row 0 of the CIA.  If this were true, it would interfere with all the keys of row 0 for the CIA, as you reported.  But after a few microseconds the VIC lines finally reach a good value (close to +5V) and the other rows of the CIA will read ok.
 
So that's just a crazy theory I had.  If it is correct, you may be able to fix it by connecting a small resistor (480~1000 ohm) between the K0 line (such as pin 26 of the VIC) and any +5V (such as pin 48 of the VIC).  This should help pull-up the 'slow' K0 line of the VIC.  That's up to you... I would try it before doing a major cleaning of the keyboard because it has a LOT of little screws and you have to unsolder all those locking keys.
 
---------
If you have a voltmeter, you could try testing the +12V on the disk drive when it is on but not doing anything.  If it is good when doing 'nothing' then you should try issuing a disk command and continue to watch the voltage (you might need an assistant for this).  The voltage may drop a little bit (say to 11.5V), but if it drops a lot (9 volts or less), then it is likely the stepper motor or drive circuits have issues.  If the voltage stays the same, then it tells you nothing (a burnt out circuit / motor may not use any power).
 
As for swapping power supplies, I imagine both supply +5V and +12V with adequate amps, although I only have external 1571 units so I can't say for sure.  I guess the only problems you would have would be the physical mounting and perhaps the connectors.  If the connectors are the same it should be easy to test without mounting.  Of course if the connectors are different you would need to check the voltages and splice the cables before you could test...
 
Anyway, good luck!

[JamalBG]

Hi! I'm new on this forum, and I'm not fluent English language speaker, so please excuse my bad spelling. I had bought Crrn's C128D and tried to fix keyboard problem, but soldering 1k resistor between 26 and 48 pin of VIC-II shown no results, problem still exists and i don't have other 48 pin VIC-II chip to check if this problem is caused by short circuit somewhere on mainboard or by fault VIC. Dismantling keyboard and checking with ohmmeter every wire between keyboard and connector shown that all are okay and theres no short circuits between them. What should i do next? Did anyone know where i can buy MOS 8566? I checked eBay but with no results.

P.S.
Pressing F1 results in printing 2, F3 - 4, F5 - 7, and pressing F7 results in inserting a tab, which (according to table on this page http://sites.google.com/site/h2obsession/CBM/C128/keyboard-scan), are in one row with the number 8 on extended keyboard.

P.S.2 So the pattern goes like this: Pressing a key from black row (attachment) results in printing a character from red row in the same exact column. Every other key works fine, also, keyboard works just fine in C64 mode.

[Hydrophilic]

Sorry to hear you bought crn's problem!  The shared line of the VIC-II and the CIA is definately the problem.
 
Since you tried the crazy resistor trick and it did not work, then maybe the VIC-II is okay but the CIA is bad.  The CIA's will often get blown out by electrostatic discharge when people try to connect a joystick or mouse.  And as you may know, the joystick and keyboard are both connected to the CIA.
 
The good news is the CIA is very easy to replace (not soldered) and is very common (for Commodore chips).  In fact, there are two CIA chips in the C128.  You should check to be sure they are both 6526 CIAs before you try to swap them.  I've seen some that have 6526 in one socket and 8526 in the other... (I don't know if it would be safe to swap them).  But if they are both the same, swap them and see what happens.
 
If they are not the same, you can also find a CIA inside a C1571 if you have one... I don't remember if they are soldered or socketed... and you can also buy them on eBay pretty cheap.  Anyway, good luck!
 
P.S.  You're English is pretty good, I had no problem reading your post.

[JamalBG]

Thank you for fast reply. I've just replaced the CIA chip (the one sitting close to the keyboard port) with the chip from working 1571 disk drive. Both are marked as MOS 6526A-1. After replacement, nothing has changed, problem still exists.

The other CIA chip on my C128 mainboard is sitting close to the user port, but it is marked as 6526A, and according to this http://www.commodore.ca/manuals/funet/cbm/documents/repair/troubleshooting-c128.txt its responsible for user port and serial.

Is there a way to check if those EMI's are fault? Because i dint have any other ideas what could caused such problem.

After small research i have found something interesting. Keyboard is pretty simple device, works when circuit is made when pressing a key. But there's  As i remember, original commodore 64 keyboard used 20 pin interface. Commodore 128D have interface with 25 pins, I'm assuming that those additional pins are used for extended keyboard. Pressing F1 results in making circuit between pin 13 and pin 8, and pressing 2 on extended, makes circuit betwen 21 and also with pin 8. So i checked with ohmmeter keyboard port with commodore turned off and pin 13 with pin 21 and shows that they are circuited. So I'm pretty that between CIA and keyboard port is short circuit between 13 and 21.

P.S.
Ok, now im really confused. After fully disasembling of commodore, i took mainboard and flipped it over to check for short circuits. And when it was flipped over, i decided to test CIA for the last time, but without keyboard plugged in. Knowing  that circuit between 11 and 13 should result in reacting as it was return key pressed. I turned C128 on (without keyboard) and made circuit between pin 11 and 13 (where keyboard port is soldered in) using small resistor and what happened? Everything was okay! Every other reaction was like it should be. CIA, VIC, and even EMI's are working fine, so it must be a problem with keyboard port, or keyboard itself.

P.S.2
After double checking of keyboard, i found the problem. There is short circuit in cable. I will solder new cable and post results. Thank you very much

Now i will try to repair disk drive, but i am afraid that previous owner could plug some connectors from drive to a drive pcb not in the correct way. Does anyone have photos, diagrams or schematics of how should be connectors plugged in internal drive, and in the external 1571?