Casio CZ-1000 Repair

I have never owned a CZ series synth, only little casio samplers (SK-1 and 5). The CZ’s are intriguing, unique and make cool sounds (see video at bottom for inspiration). A little different architecture and technique for sound production. Be sure to use 8 stage envelopes and cross fade voices! How does it work? Readme

Why Fix a Casio? I wrote and recorded this song using ONLY the CZ-1000 and my custom patches. This is the synth I repaired. Read on!

Get the Manual!

The service manual is fantastic, and has so much info you hardly need the datasheets for the special chips in use (main cpu, sub cpu, ram, rom). It has pinouts, data flows, timing charts, waveforms (analog and digital), truth tables, and descriptions of what pins do. Good functional information on how key scanning matrix (for buttons and keys) and LED drivers work, too.

There are many other synths (Roland Jx-8p and Ensoniq Mirage are good contemporary examples) that have this same type of architecture. If working on 80’s synths, this is a great guide, along with the CZ-1 manual.

The system uses a few different voltages, control by some really critical trimmers and a couple of power transistors.

I found the schematics and others docs as a bunch of separate files so I combined them into this single pdf:

Service Manual and Schematics

Power Up

The AC/DC converter for this unit is the AD-5. It puts out 9v DC at up to 1 amp. Center negative- the tip of the plug is negative. I don’t have a power supply for it so I am adapting an old one, The socket is unusually deep (9mm), but I found a barrel plug for it and an old power supply that I can cut up and put the new plug on, ensuring the polarity is correct, of course.

I could buy a new one from Amazon or ebay for about $10, but my efforts keep the old parts out of a landfill.

I changed the plug on my power supply and fired it up. The synth plays and the controls work, but it sounds pretty bad (video below). Takes a few minutes to warm up, too (volume is low and unsteady at first). The oddest thing so far is the pitch wheel- go to the high extreme and the synth squeals, go low and it dies out. In the middle it works.

UPDATE: The pitch wheel problem is due to how the synth stores some things in RAM and then they are lost if you remove the batteries or unplug for more than about 10 minutes. The fix is to press the ‘P’ button (the reset).

The Usual Suspects

RAM uPD449 (x2) – user internal memory (no traditional battery backup but settings stay in memory if plugged in or batteries are installed)

ROM MB83256PN26 – synth memory for things like presets.

74LS373 provides addresses for RAM and ROM access.

CPU uPD7811 – main microcontroller for the synth.

uPD933 (Music LSI) – gets 8-bit signal from the CPU and creates a 12-bit digital music output. The schematic says the clock signal for the uPD933 is 8.96MHz, but testing it (after calibrating my scope with a DSS) shows it is half that. Synth sounds fine in regards pitch.  I think 4.48 is the correct value since the pin description table says “about 4.5MHz” and my scope reads the S&H frequency at it correct value, so I know it is pretty accurate.

DAC

Several 4049’s (3) for interfacing to the DAC from the uPD933. These invert the input signal. A low signal about around 2.5V. High is 5v. I tested all of the inverter gates with a voltmeter and they all work. Just press a key to send in a signal.

EXK-S19Z2052R – this is be a ‘ladder network’ DAC that runs the inputs through different resistors to create a combined analog signal for how to represent a 12-bit input.

Test points are shown in the manual. Compare test point output on oscilloscope to expected waveforms in the manual for the CPU and the DAC. The DAC inputs look totally digital- curves made of little dashed lines.

VR2 trimmer adjusts the DAC output voltage.  It can’t do much since it is in series with a 330k resistor, and it is only 10k. VR2 is on the edge to the left (when looking at the bottom).

Expander Circuit

This is made of a voltage divider network and a 4051. The expander circuit helps restore the voltage signal to its proper shape and smooth it out. The input signal looks like it is clipped, but it is really more like the top of the wave is pushed down into the rest of it. The manual calls it “contracted,” which explains why this component is the “expander”. I feel like it is because the high order bits are maxed out and the lower order bits still retain their shape, they just start from a lower voltage.

The 4051 (8 channel analog switch) appears to be working. You can also listen to the audio at the inputs to the 4051.

VR4 is the most important trimmer and has a calibration procedure. Expected to be crucial for reducing noise and making chords play better. VR4 is by the P switch. VR4 controls the voltage drop across a voltage divider network in the expander circuit. The network controls input signal voltage to the different channels in the 4051. The output varies from 1/16 of the input to output equalling the input (no change), depending on the channel. Uses inverse powers of two- 1/16, 1/8, 1/4, 1/2, 1. The uPD933 selects the channels.

VR1 is by the corner below VR2. Has to do with power down connector. Setpoint adjustment, I think.

LPF chip – parts list says LPF-V20. I think that is in the Cz1. Schematic says LPFv17. The 17 or 20 refers to filter cutoff point in KHz. The LPF removes frequencies above 17KHz.

Transistors are 2SC17400, unless noted otherwise.

CZ 101/1000 Block Diagram

Automatic Power Off Circuit

Keep in mind that this circuit works whether on battery power or plugged in. When troubleshooting, you should turn this off so that the synth will remain properly powered. Otherwise, you will get confused and freak out thinking your synth keeps powering off all by itself, not that I know anyone that happened to. 😉

Audio Output

A dual op amp (4558) is used in the audio amp circuit. One half provides the 1/4 out directly, but for the headphone output the 4558 feeds into a Sanyo LA4138. Similar to a 4137 (same pinout) but requires more voltage and puts out more power (louder).

Tapping pin 2 or 6 (electrically equivalent) of the 4558 will let you listen to the audio through an amp or look at it on an o-scope before it hits the op amps. This is useful if trying to identify if the op amp is adding noise to the output.

Calibration

5 trimmers. 1 is on the power and output board M452-MA2M and the other 4 are on the big board that generates the sounds, M4512-MA1M.

+VBR adjustment – With the power off BUT PLUGGED IN OR ON BATTERY, adjust variable resistor VR1 on pcb M452-MA2M so that the emitter of T11 is 4.3v +-5%. Mine was reading 3.9v. The adjustment was simple- just need the digital multimeter to get it accurate. This part of circuit is powered when off, since the unit needs power to keep user RAM stored.

pcb M4152-MA2M

Schematic view of pcb M4152-MA2M

IC’s indicate whether receiving VBR or VDD. It does not look like VDD can be adjusted. My VDD is perfectly steady at 5v.

VBB Adjustment – with no sound/no keys pressed, adjust VR4 on pcb M4512-MA1M so that the potential difference (voltage) between pins 5 and 7 of the op amp is at a minimum. The op amp in question is the unlabeled one in diagram below. It is only one nearby receiving power to pin 5 and outputting on pin 7. The op amp is also in the DAC circuitry and I know from other reading that VR4 adjustments affect the DAC. My VR4, however, is doing NOTHING.

The op amp is a quad op amp- NJM2058D. In stock at Mouser!

op amp used in 4 places of sound generating circuit

The audio output goes from T2 emitter to the volume slider and then to the audio output section (with its amplifiers) on pcb MA2M.

What about VR2? It is not discussed in the manual. 😦  It controls the voltage level going into pin 6 of the 2058, which is the first step of entering the expander circuit.

“Once VR4 gets in the right range, VR2 starts being helpful. You have to adjust between a hiss and a crackle, but it’s not too bad.”- found on reddit

Adjusting VR2 should change this level, affecting the output at pin 7, too. Expect 0.2v peak-to-peak based on waveform diagrams in the manual.

Power Down Detector – Apply 6.3v (vice 9v) to the power jack and adjust VR1 on pcb M4152-MA1M so that the pilot lamp shuts off. The pilot lamp is a red LED under the circuit boards in that area. Don’t confuse with the red LED under the ‘P’ switch that flashes when you press the switch. Now where is my 6.3v power supply??

Location of VR1

Audio Signal Tracing

For a few glorious seconds, the synth played PERFECTLY and I have no idea why. I hate intermittent faults; at least this one is bad most of the time which should make it easier to find. This gave me hope.

My audio is not (so) bad when playing one note. Adding notes makes it rapidly go bad. I have been back-tracing across the both pcb’s to try to find what is causing this. The output of the pd933 to the resistor network acting as a DAC seems as expected. You can listen to the audio at the various stages in the DAC (the resistor network). You are effectively listening to the decimated version of the signal- it goes from normal to noise. It can be pretty crazy. Imagine the mods that you can install in this synth!

What does an op amp do? It amplifies the differential signal between the inputs. The output does NOT have to be greater in magnitude than a given input, but for a gain > 1, will be greater than the difference signal.

NOTE: The CZ-1 Service Manual has some better explanations of the block functions. Even though some different components are used, the concepts are the same and it is well worth the read.

My testing so far has not found any problems on the output of the DAC. The components all look like they are doing what they should with the signals given. Time to look on the other side of the DAC. There are 3 4049 hex inverters. I need to look at their input and output signals. I think they are being used to create square wave pulses of different frequencies. Maybe the decimation I hear in the DAC is NOT right- could be a bad inverter? They are also being used in ways I don’t understand yet. Some of the gates are in parallel and their outputs are joined. Is that doubling the voltage?

I tested all the 4049’s and they are good. It looks like everything is doing what it is supposed to do, so I went back to look more at VR4 which does nothing.

VR4 calibration steps does nothing to the sound. The voltage at pin 5 is not steady, and clearly looks like a capacitor is being charged up. VR4 has little to no effect on the voltage at pin 5, so I can’t calibrate it. Why? I checked the trimmer output versus each other connection. The trimmer is 500 ohm and it read properly across that part. However, the readings between the other pins are bad. One reads 36k, and the other reads 3.5M!! This is the source of the noise and why VR4 does nothing- it is simply broke.

This video lets you hear the problem I was having shows a quick hack to verify it and identify what part needs to be replaced. This is the “before” video (but it also shows the basic fix).

The Fixx

If I can find a part that fits well-enough, I will install it; otherwise, I will add a trimmer in the right spot in the circuit so it can perform the same function. It matters how it would get tapped in, since in the wrong place it would provide a resistance of 500 – 1000 ohm vice 0 – 500 ohm. I think I already got it right, though, since the synth sounds nice.

So here’s what I ended up doing. I put in a trimmer and tried it out. No change to waveform- it just affect the DC offset of the waveform by about 10mV. No audio change. What is the fix/hack? A 10 ohm resistor soldered in to jumper around the bad part of the trimmer. This fixes the noise problem and the warmup issues where synth was real quite for a minute after startup.

Picture below shows circuit board layout and also the schematic for where the 10 ohm resistor (essentially a jumper) was installed.

Sample and Hold

I checked the frequency of the input and it is accurate (as good as I can tell with my scope)- 40KHz.

LPF

I can detect no change in the signal passing through the LPF. Then again, I can’t hear anything at 17KHZ anymore. It is not causing the trouble, though, which is some modulation/beating between notes in a chord.

Various Fixes (or not)

I found four traces that were cut- all on purpose. They were jumpered over but two jumpers have broke/detached. I have repaired them but no change in the sound.

The 26 pin ribbon cable on MA1M has several wires NOT connected to the board. I think I fixed them but it is ugly. There is a lot of stress on the joints due to bend of the ribbon cable, which is why they failed. These are connected to the cartridge socket (which I already broke a pin off of- oops). Cartridge won’t work anyway.

Everything has tested great. All the buttons and keys work. I cleaned the keys well and also re-glued one edge of the control panel label sheet down better.

LCD Contrast

I found this one the hard way. I adjusted it, thinking it might affect playing. It did not. Later after doing some soldering on a broken jumper, I noticed the screen had gone out. Arg- already broke it! Then I thought, maybe one of the trimmers was LCD contrast? CZ1 has one for that. Turns out that is all it was.

LCD contrast trimmer – VR3

MIDI

It works over MIDI. Since it is upside down when working on it, I plugged in my Microfreak so I can play it as needed during testing.

It supports sysex, too, and works with MIDI-OX so you can get all kinds of patches for it.

• https://coffeeshopped.com/synth/casio/cz-1000 – general good info
• https://www.kasploosh.com/cz/13549-how_to_transfer_sysex/ – sysex how to
• http://czounds.com/ – patches
• http://machines.hyperreal.org/manufacturers/Casio/CZ/ – patches

I am able to send sysex patches to the CZ1000 using MIDI-OX. Plug in both MIDI cables and makes sure memory protect is OFF.

Note that only one patch can be sent at a time and it is sent to the edit buffer. This means you will have to WRITE the patch to INTERNAL memory. Of course, if you don’t have batteries or power plugged in, you can forget about it being saved… but read on

Battery

User memory lasts 10 minutes once power is turned off or batteries are removed. There is no internal battery just for the RAM. The assumption was you would keep ‘D’ cells in it all the time, which also powered the RAM. That explains why some perform a CZ101/1000 battery mod.

Memory Cartridge

Crazy expensive and not worth it. Plus, the plug has one pin broken on this one and they are all needed. Without replacing the male pin header the cartridge plugs in to, this will NOT work with a cartridge. May need ribbon cable work, too. Instead, buy this MIDI gadget and store 2000 patches! Thank you, Lestyn Lewis.

https://lewismidi.com/czpl88mini/

References

• http://zine.r-massive.com/casio-cz-101-keybutton-repair-guide/
• https://www.synthtopia.com/content/2009/11/15/the-casio-cz-1000-synthesizer/