There's the power supply on the
right, what looks like the keyboard logic decoder board in the middle,
and what I assumed was the guts of the Kurzweil sound on the left motherboard,
with the four items on the top, from the left: the two control wheels,
the volume slider, the data slider, and the display circuitry. All this
seemed ok - just dirty. I sprayed and wiped and cleaned it best I could.
Since most of the keys worked, I suspected the problem was somewhere
underneath the keypad. How complex would it turn out to be? Let's see...
I took the keyboard out and turned
it upside down, at right angles to the case. Hmmm. More dirt. I cleaned
it as carefully as I could, with alcohol and Q-Tips.
I unscrewed the two left hand circuit boards and folded them up along
Immediately visible was a jumper
someone had installed previously. Hmmm. So there was a history of a
bad connection here. If I was more confident and knowledgable the solution
was here, staring me in the face. But I missed it this time. The complexity
of the circuit board traces boggled me. If I'd paid any more than $50
for this thing I'd have been chicken to mess with it - but I figured
I had little to lose by trying.
I did notice that one of the keys had a plastic lip broken where a metal
tab engages it, giving the small metal plate connected to the return
spring something to hold to. I managed to shoot that spring across the
room several times, even going so far as to buy some replacements in
case I couldn't find it. A tiny spring in a crowded room is hard to
find, but each time Bonnie managed to find it. I filed a new lip in
the key, and it seems to be working OK.
I turned my attention to the keypads. I took off the rubber contact
things, and cleaned the circuit board contacts underneath. Also cleaned
the contact rings attached to the rubber things. These looked very familiar.
I think the CZX-3 has them. I know I've fooled with these things before.
While messing around here, I noticed that I could trigger notes by touching
a rubber mounted ring to a circuit contact, unless it was one of the
keys that didn't work. However, while the keyboard was upside down,
with the black keys being depressed by the weight of the assembly, an
Eb which was dead began triggering. Bad Connection, it was screaming!
Again, here was the answer shouting at me, but instead I stared back
at it for quite a while, trying to grok the logic of the circuitry.
Those little traces still boggled my brain.
I put it back together, on the off chance that cleaning had helped some
of the keys. Instead, more of them didn't work!
Disassemble, and look closely. Hmmm. The little rubber thingies insert
into holes in the metal support structure, which is where the keys themselves
encounter them, pressing them down so their little round contacts touch
the matching contacts on the circuit board. And these rubber things
are very flexible, and it's easy for them to not quite clear the holes
when you're replacing the boards:
Also, one end of the string of rubber contacts doesn't have a connecting
tab, so it hangs loose:
Have to be careful to get that one in the proper hole, too! Rubber thing
not in hole correctly = no note!
So I took the circuit board things off, checked very carefully to replace
them with the rubber things in the right places, and then put it all
back together. Hmm. The NEW bad notes were gone, but the old ones remained.
At this point I felt the need for help. I joined the Musicplayer.com
Keyboard forum, and posted a help notice. Someone there referred me
to the Kurzweil K1000 forum on Yahoo, and I joined that as well. I had
to leave it alone then for awhile.
When next I had time to fiddle around, I brought my ancient Radio Shack
Micronta voltage/resistance meter to check the contacts. Underneath
the keypad circuit board were diodes, two for each note. I checked these,
and they all seemed ok. However, I noticed that if I checked the continuity
between any of the diodes and one the jumpers nearby it would trigger
a note! Cool! Now I'd have a signal to trace.
About this time I got a message on one of the forums asking "which keys
are bad?" and of course, just listing them, finding some order to them
would be an important step in finding out why they were bad. The fact
that they were all in the lower half of the keyboard, and in fact all
on the leftmost of three circuit boards under the keypad, pointed to
a possible bad connection between boards. I investigated the jumper
Now, here there are two plugs, concentric, so to speak, with the outside
one being bigger. The contacts on that one are labeled S1, F1, ...F4,
S4. The smaller one is labeled 0, 1,...7, 8. Ah, that seemed so much
easier to keep track of! Also, that one had direct connections to the
diodes that were associated with the notes that didn't work. So I started
with that one.
There's a connector that communicates the signals from the keypad to
what I assumed was the keyboard decoder board. I figured that, once
I figured out the pin layout I could check the continuity of the whole
keypad by checking to see if there was a break between any point on
it and this connector. So that was my next goal.
Radio Shack continuity testor to the rescue! Here's how it comes out:
Now it was just a matter of checking
each diode pair and then numbering the associated notes, 0 ->7, which
I did, upside down, on the note ends.
Oddly enough, when I also numbered the diode pairs, the last pair was,
I think, a number 3, but the top note, a high G, was a number 4. This
couldn't be right! I told Bonnie about this, and she immediately said,
"Well, all you've done is number the white notes!" Of course! The black
keys were hidden, supporting the keypad upside down. Sheepishly, I agreed.
This would not be the last time I did something stupid... !
So I got some alcohol and cleaned the numbers off the key ends and then
renumbered them, remembering to skip a number for each of the black
keys. Now the pattern was apparent - all the bad keys were 3s! And all
on the first (furtherest from the big connector) circuit board. My first
thought -- look at those concentric group jumpers again..
I looked underneath those connectors,
and, sure enough, there was a crappy solder joint - and it was the fourth
one down. That would be the number 3 trace.
I got out the soldering iron and touched up that solder joint. In the
little time I had left that day, I checked the keys, and they seemed
to work. All should be right, right?
Wrong! The next day I put it all back together and tried it out -- and
the same 5 keys were dead! What? Maybe I was mistaken the day before
when it seemed to work? Time to tear things apart again.
This time I looked closely at that group jumper. Hmmm. What had I missed?
Well, as I turned the board over several times, checking the numbers,
I realized that I had been counting from the wrong end -- what I thought
was the number 3 trace was actually the number 4 trace! So my soldering,
though though it probably hadn't hurt, couldn't possibly have fixed
the wrong notes. The problem was a bad connection, right? Maybe just
my jiggling and screwing around with things had changed conditions enough
so that the notes worked briefly, enough for me to think I'd fixed it.
Whatever the facts, they didn't work now.
So my attention went back to that jumper that I'd seen the first day,
soldered onto the middle circuit board:
Hmmm. Looking down past the jumper
there was a spot of corrosion I'd missed before. Looked like someone
had spilled beer or something, years ago, which had slowly eaten into
the circuit board:
How had I missed this? You can
even see the mark I made next to the diode pair which corresponded to
a bad key! I checked the traces, between boards and between the diodes
and the off-keypad connector. Hmmm. None of them seemed disconnected...
To check the continuity, I'd been using my ohm/volt meter with the minimum
ohm setting - and now that I thought about it, the resistance seemed
pretty consistent from one numbered diode pair to another - about 3
to 4 ohms. I decided to check ALL the diode pairs, with their similarly
numbered neighbors as well as with the output pins on the big connector
--check all the 1s, all the 2s, etc. Ah HAH! All the readings were in
the 3 to 4 ohm range, EXCEPT for the diodes corresponding to the bad
keys!! Those were in the 12 to 13 ohm range- and just so happened to
come through those corroded traces. So, all I had to do was solder another
jumper to jump over the corroded traces.
Get out the soldering iron again! As long as I was at it, there was
another trace right next to the 3 trace that was bad looking. It's reading
was OK, but why not make a pre-emptive strike and jump that one as well?
So I put another jumper from a pair of 4 diodes across the bad patch,
to the next pair of number 4s. This was tricky, since both jumpers started
and ended at neighboring diode pairs, and I didn't want to screw anything
up. I carefully put masks of duct tape over contacts I didn't want to
disturb, changing them around as I soldered one wire and then the other.
Whew! That might do it, I thought.
Wrong! When I put things back together (not entirely - I was getting
good at just putting the keypad together and setting it inside the case
without screwing anything else down) I was overjoyed to find that all
the keys worked - more or less. Sometimes more - all the 4s and 3s were
now tied together! Playing any 3 sounded both that three and the neighboring
4, and vice versa. An interesting effect, but not what I had in mind...
Back under the keypad. Looking closely, I saw that as I'd soldered one
of the jumpers, being careful not to get any solder bridges between
contacts, I'd missed seeing that the insulation had melted enough for
a wire to penetrate from below, shorting it to the next diode pair -
just what I'd been trying to prevent. A little repair work and it was
This Story is from another member
with permission of course.