I bought an Omnichord OM-84 August 30th, for my son. My hobby is repairing synths and this one needs some repair work. It does not work when on battery. When plugged in, you can hear little squeaks but no notes.
I did quite a bit of work on this unit as described below, then set it aside for a month (while fixing a Moog Prodigy!!). I plugged it in and voila! It plays. It’s perfect. I don’t know why. Maybe because it is 10 degrees cooler in the house? Temperature affects caps and resistors. Maybe I will find out next Summer!
A visual examine found corrosion on some wires on the main circuit board (the one with all the IC’s). The corrosion comes from two leaking electrolytic capacitors. These are attached to the main audio amplifier (a TA7222AP). I know these have to be replaced and am hoping that will be the end of the story. It is missing one button which I have already found at omnichord-heaven.
Useful Information on Disassembly
There are two circuit boards, joined with several pin connectors, so that the boards can be easily separated. The boards are screwed in to the bottom half of the case.
The speaker has a 2 pin molex connector to the circuit board for easy unplugging. So does the battery power. When boards are installed, the speaker molex is the closest one to the speaker.
There are two other molex connectors going from the main board to the membrane switches (under the chord keys). They unplug from the membrane board.
The ribbon cable from the main board to the strumplate can be pulled out. Be careful- there is no latch to hold it tight; it just slides in and out.
I replaced the two capacitors. Odd thing is they tested OK. Still, the stains and corrosion were worrisome. One of the leads was also loose at the pcb solder joint. Once I replaced them, the omnichord plays a single note (in different octaves). The power supply keeps cutting in and out, making troubleshooting impossible.
Since I am getting different octaves, that means part of the voltage divider circuit ((4530 and 4040) is working. Since I only have a single note, the timer (IC 2) is suspect. Its output drives the frequency divider circuit. BUT it is controlled by the m50740-701sp micro-controller, which gets signals from the chord buttons.
Now I have found that the power input plug is not good- power cuts in and out when it is moved. I had to add a jumper from the loose power plug pin (the ground in this case) to another ground point (I chose the ground where the battery plugs in). Now power is stable.
By the way, polarity on an Omnichord wall wart is REVERSED from typical power supplies. I cut my power chord and flipped the leads so the polarity is correct.
There is no rhythm section, chords, or any response from the keyboard. Volume knobs and sustain work, though.
I DO GET a fast (140bpm) tick when in chord mode. It responds to volume control and the trim pots for tuning (which control the master clock circuit). In bass line mode, there is no bass but I can hear a tempo click that speeds up/down in conjunction with the tempo knob.
It PLAYS, but
One day I plugged in the omnichord and it played perfectly for just a few minutes.This only happened once. Usually it just plays a single pitch and only when strummed. Turning on/off/on rapidly will also cause it to NOT play. If you wait about 15 seconds and before turning back on, it will play.
What to Check Next
Look at m50740-701sp. This has been suspect in other units. I don’t see the correct voltage between pins 52 (voltage in) and 26 (ground). Should be around 4.7v. I see 0.8v. I double-checked and that is the voltage- which is completely wrong. All other IC’s are getting 4.9v and the 7905 (5 volt voltage regulator) is putting out a solid 5v. I looked at the schematic and saw there are TWO voltages derived from the 7905 output, and one is dedicated to the m50740-701sp. That chip produces signals for the drums, bass line, and signals for the programmable timer that puts out the frequency pulses for the chords. All I hear are basic clock ticks bleeding through.
I have found multiple people suggesting to re-solder the solder joints for the connecting pins between the two circuit boards. That could explain why some of the buttons have no effect on operation. It can also explain why the power to the microcontroller is bad- a high-resistance connection between the pins and the pcb could lower the voltage. The smart people at Folktek also suggest the re-soldering approach as a first step when repairing an OM-84.
I re-soldered all the pin connections- about 80! This fixed the power supply to the microcontroller (now ~5v) but had no other noticeable effects.
Check the circuits in the rhythm section. Maybe buttons are just not working due to bad connections? One bad solder joint repaired already- so may be more!
Now that all the chips have good power, time to try the O-scope on the circuit and see what I get from the timer and microcontroller and if any button pushes change the signals seen on the scope. That will help figure out if buttons are doing anything at all.
Look for burned resistors. Found none.
While inspecting the board again I found a very loose capacitor. The solder joints were cracked so I repaired them. This was in the control section. I found a couple more and repaired them.
Why one note? Maybe I have a shorted/stuck chord button? That would give one note on and prevent (I’m guessing) others from sounding. I disassembled the button pad, cleaned it, tested all the cables and verified and the molex adapter pins (2 cables, 4 plugs) are all good. This had no effect- still have one note.
New Random Problem. I turned the main volume now and got a click and sound cut out. That was weird. The knob has caps and either side and feeds right in to the audio amp.
The rhythm section is dead, except for clock tick. It has a 4011 in it, set up for use as 3 separate inverters (logical ‘not’). The schematic has one pin mislabeled. The Bass Drum feeds in to pins 1,2 and out through 3 (vice 5).
Enter the Logic Probe
This appears to have gone beyond a simple fix. Time to check signals and look for component failures down to the IC level. That is a job for a logic probe.
I now think it is playing only one note since only the sonic strings are playing. I have also confirmed that the time (8523) is only putting out two of 3 signals to the frequency divider circuit. Pins 10, 13 are OK. Pin 17 is NOT putting out a clock signal.
The logic probe confirmed the 4520 (which is fed from pins 10, 10 of the 8523) is good. You can even hear the octave change between stages on the probe as the clock pulses slow.
The 4040 is not receiving input so can’t test it (without a logic pulser!).
Audio IC and circuitry is good, since I have partial playback.
The clock circuit is good. I see all pulses where they should be, and turning the tuning trim pots changes the pitch of the sonic strum playback. The PIT (8523) and Micro-controller are getting clocked properly.
4556 Dual Binary Decoder
The logic probe and voltmeter show the correct truth table values so the 4556 is good.
4069 Hex Inverter
All the even pins are reading low (with logic probe and volts around 0). All the odd pins are reading 5v. This looks normal.
4050 CMOS to TTL converter (Non-Inverting)
Pins 11, 12 have a high signal. 12 has about 4.8v, 11 has 4.1. I would expect 11 to be lower than 12, since this takes CMOS to TTL levels, but why no signal anywhere else? (11 feeds 12)
74Ls05 Hex Inverter (TTL)
The inputs are the odd pins and the outputs are the even pins. All have inverted signals at their pins as expected.
4011 is a Quad NAND
I found that one of the NAND gates in a 4011 on the main circuit board is in an invalid state! That’s important. Pin 12 is high, but 11 and 13 are low. It is IC 5 on the schematic, which correlates to the 4011 closest to edge of main PCB. I am likely misunderstanding something, since IC5 is involved in the Tempo circuit and that is working properly. It sends a clock signal to pin 19 of the micro-controller. If you turn the tempo pot and listen to the clock pulse at IC 19, it goes up and down as expected.
IC6 is the 4011 next to it. Pins 1,2,3 and 4,5,6 are as expected. Note that pin 6 gets a clock pulse, causing a pulse at pin 4. Looks like over half of IC is all connected to same pin and not used. Not sure where pulse is from. the schematic does not appear accurate for these IC’s.
IC15 is another 4011. It is part of the rhythm circuitry and appears really messed up. None of the 3 NANDs in use are in the proper state. The rhythm circuit does not play at all.
Can’t figure this one out. Thought it might be connected to kill switch but see no response.
Battery Terminal Corrosion
They are thoroughly corroded, as expected. I soaked them in vinegar to remove some corrosion but honestly, a pain to work with so I am ignoring for now.
If all else fails, there are repair shops with parts and know how! Omnichord heaven is also on facebook and has already answered an initial question. Very helpful!
Circuit Description (from a web forum, with some additions by me)
The touch sensor (strum plate) is connected to the main circuit board (the one with all the IC’s) via a ribbon cable on the right hand side.
The 2 molex connectors are for the chord button connections. The buttons are just switches, so you could short across the points on the main circuit board the same way the buttons do, and get a chord/note to play. That is also a way to test if the buttons work or not- test continuity across the two points when you push the buttons to see if working or not. That is, while holding a button down there are two points on the circuit board with 0 resistance between them. Just have to find the points! Not hard. These is a group of 10 k resistors that feed 5v power through buttons into IC 7 (4050) and IC 9 (4069). Those are the points.
A keyboard is a keyboard, whether it is on your computer or on a synthesizer. All of them take a switch closure of some sort, and tell the CPU it happened. The CPU looks up in ROM what it is supposed to do about it. Your keyboard array selects the base chord/key it is playing in.
The CPU then produces the notes of the chord selected. There is an array of 20 dividers (Made by IC 11 (4520) and 12 (4040)) in upper right area, that produces individual notes. The dividers act on the frequency of the input from the Programmable Interval Timer (8253).
In between those dividers you can see the numbered feeds down to the touch sensor circuits, providing the correct frequency for the harp sound (strum plate).
The touch sensors then trigger the notes to come through the audio. Each sensor allows a cap to charge holding the base of each of the 13 transistors to an on condition, which gates the sound through. All 13 transistors are not draw, only the first and last. As the cap discharges, the gates comes closed. The sustain control is a adjustable voltage that can retard the discharge, thus lengthening the sound of the note. Discharge is also controlled by resistors R1- R13. The values fall into 4 groups, one for each section of the touch sensor.
Center right is a noise circuit – it generates white noise. Or at least something similar to white noise. Note upper right side of the CPU has many ports, notably ports P30-34. Note further they are also labelled BD SD CL CY HH, for bass drum, snare drum, claves, cymbal, and hi hat. These output the signals to the noise generator to produce the rhythm. The circuits around the noise generator gate the white noise to emulate percussion.
Look at the two noise generators, then to their right is a third transistor with an inductor in its collector lead and a 100k trimer in its emitter lead. That is the output from the noise generator. Note now to the left of the generator the HH SD and CY control signals all come in to gate that same third transistor. They use the same sound for each, the tone being a matter of the gate timing. Different size caps gate different length of time.
Don’t be confused by the drawing. Up on the CPU the port numbers are on the IC and the function – HH CY etc – is in a circle with an arrow. But down at the noise circuit, the port number is in the circle and the function is just nearby letters. These are the same things.
Below the generator is a row of three pulse generators for percussive hits, created by using the 4011 inverter and surrounding RC network for tone control. Each is triggered by pulses from the microcontroller. The left is BD, the middle CL, and the right is SD. Note too that the snare drum triggers not only some white noise but also has a percussive hit. That makes a more realistic snare drum hit.