Building A Modular Synthesizer

July 2022

I am going back to my roots, when I built a PAIA synthesizer from a kit in the 70's. My old synthesizer looked something like this:

I recently bought the Behringer 2600, a patchable mono/duophonic semi-modular synthesizer:

I have wanted a big modular synthesizer since Carlos and Tomita made synthesizer transcriptions of Bach, Debussy, Mussorgsky and other composers. OK, the deciding factor: I bought the 2600 when I read that it was used in Jive Talkin'... Here is a Bach organ fugue I sequenced on the 2600:

When I found the pages below, I decided to make my own modules:

Digisound 80 pages

Yusynth Modular Synthesizer pages

Instead of building from kits, I will make my own circuit boards and front panels, and I will source parts from Digikey, Amazon, Tayda Electronics and ebay. My plan is to build a complete synthesizer voice: voltage controlled oscillator (VCO), voltage controlled filter (VCF), voltage controlled amplifier (VCA) and voltage controlled envelope generator to control note attack, decay, sustain and release (ADSR). If I like what I make, I'll make more.

Circuit boards

I want to etch my own boards. I bought FR4 single-sided copper clad boards from Amazon, 10 boards for $7.

Many sites recommend ironing toner from laser printer images onto circuit boards. I don't have a laser printer so I had to come up with my own method. I have a 3d printer. I made a simple pen holder to hold an extra fine Sharpie marker and secured it to the 3d print head. I used the printer as a plotter, had it draw circuit diagrams directly on the copper boards.

I worked up a method for converting PDF files to 3d printer files. I converted PDF 's to svg's using Inkscape, imported them into Blender, turned them into meshes and exported them as stl's for the printer. All of these processes I learned from watching YouTube or reading websites.

The Sharpie uses waterproof ink. This prevents the aqueous etching solution from etching the copper where Sharpie marker is present. I made etching solution by mixing 30 mL of full strength muriatic acid with 60 mL of 3% hydrogen peroxide solution. It is a dangerous mixture, releases corrosive gases as it works. Do not do this if you are not comfortable with concentrated harsh chemicals.

Here is my first try, a test on a scrap of circuit board:



Here are my first printed circuit boards (PCB's): VCA, ADSR, VCO and VCF:

I inspected each board carefully to make sure all copper traces were isolated from each other. A utility knife easily cuts the copper if there are any issues.

I drilled the small holes at my old drill press, had to find a drill bit that the press could hold. I found a set of "jewelry" drill bits at Harbor Freight. I screwed the boards to a wooden block to steady them while drilling.

I downloaded diagrams of parts layout from the synthesizer websites above. I used the diagrams to identify the locations of each component. This VCA board is ready for parts.

Resistors mounted:


All parts on. This is the DC only version, for audio.

Matching transistors

Some of the circuits require matching transistors, Q1 and Q2 on the board above, for example. I didn't know this was a thing, but I schooled myself on the web. I used the information on this website to make a simple circuit to identify matched transistors. I made a +/- 12 volt power supply using old wall warts and 2596 variable voltage converters. I got a matched pair on my first try. Here is my set-up, including the circuit diagram:

Front panels

I got aluminum plate from a local metal supplier (Fazzio's).

I based my VCA front panel design on the Yusynth design, arranging pots and jacks in similar positions. Here is a template I made using Inkscape.

I printed a paper copy of the template and taped it to a 2x5" piece of aluminum plate, drilled holes until I had all holes the right size. I had to counterbore the holes for the jacks to allow space for the mounting rings because the plate is 1/8" thick. I added mounting holes on the top and bottom.

October, 2022, the Behringer Model D went on sale for $261 plus tax. I got one.

I wanted a CV keyboard to control these synthesizers without needing to boot up the computer, and I later found it was important for independent testing of the modules as I built them. I chose the highly recommended Keystep 32 by Arturia.

Two mono synthesizers are not enough. The purchase of the Model D did not stop me from building my own modular synthesizer.

October 2022. I built a MIDI-to-CV converter from a kit from Midimuso.

A vintage style modular synthesizer requires control voltages (CV) to modify VCO and VCF frequencies, and to trigger VCA and ADSR. This MIDI-to-CV module converts pitch and gate voltages from a MIDI keyboard.

I bought the kit before attempting any home made circuit boards. Assembling the kit helped me develop a procedure for making my own kits.

The front panel:

November, 2022. I cobbled together a bench power supply, a case, and a means to distribute power to the modules. I decided to use a Eurorack style power supply, + and - 12 volts. For the bench power supply, I used wall warts and a voltage regulator. A 12v DC wall wart provides +12v to a 10 pin male connector on a bread board PCB. A 15v wall wart is decreased to 12v with a 2596 voltage regulator. The output wires of the 2596 are reversed (+ to ground, -12v to the negative end of the 10 pin connector).

Here is a power distribution bus built on pcb project boards:

I mounted this in an old DVD player case:

Before finishing the VCA, I decided I needed to build the VCO. How else could I test the VCA?

Collecting parts for the VCO:

VCO wired up:

Testing the VCO using a $24 oscilloscope. The lights are LED's on 2596 voltage regulators:

Nice sawtooth wave!

The 1/8" thick aluminum plate is too thick so I switched to 1/16" aluminum, better for mounting jacks.

Now back to the VCA. I completed the assembly including knobs I designed and printed with the 3d printer:

I had to track down problems with the circuit:

The black square shows where I found a break in the copper trace invisible to my eye but found with the continuity setting on the volt meter, beeps when a circuit is good.

In general, problems include "dry" solder joints that need to be rewetted with a hot soldering iron, burned out components (smoke!), short circuits, miswired panel jacks. I encountered all these problems while building the first three modules.

December, 2022. I eventually got the VCA to work. I tested it by feeding the gate output of the Keystep to the control input of the VCA (white wire from the Keystep to the VCA), audio output from the VCO (short green wire) into the audio input of the VCA. The VCO is getting control voltage from the Keystep (long green wire).

Here is a video demonstration of the set-up above, the first sounds from my synthesizer:

Building the ADSR:

January, 2023. After some frustrating troubleshooting, I got the ADSR working. Here it is in the test rack:

To complete the picture, a trace from the oscilloscope of the classic ADSR envelope.


VCF next. PCB:

Position of parts on the PCB:

VCF parts going on the board:

Filter front panel:

I made several variations of knobs to fit the various shafts of the potentiometers. I made many till I had tight knobs. The printer is printing one here:

Completed filter, worked the first time but still needs tuning :-D

Testing the four-module synthesizer:

Here is a video demo of the 4-module synthesizer:


I moved the 4-module synthesizer to the studio. I tuned up the filter according to instructions on the Yusynth website. I checked the tuning of the VCO. It is still tracking well across at least three octaves.

I cut and drilled a new front panel for the ADSR because it didn't quite fit in the rack.

Now it fits in the rack. I printed new smaller knobs to fit the new panel design.

I arranged the modules in a tighter fit across the top of the rack. It's ready to play. Hear's a nice simple horn patch:

Here is the horn with a couple of acoustic guitars: horn demo

A modular synthesizer is never complete. OK by me. I have started building Yusynth's Voltage Controlled Low Frequency Oscillator. Here is the drilled PCB partially labeled.

LFO working! Now I can add vibrato, tremolo, make proper theremin-like music, other generic space sounds. Here is the LFO in the rack with the oscilloscope showing the sawtooth output of the LFO.

Except for the filter, none of these modules worked the first time. Each took several days of trouble shooting. One or more solder connections to the short pins of the DIP sockets were often dry, had to be resoldered. On the LFO PCB, there was a hair-thin copper trace from trimmer T3 to the +12v bus that caused a 10 ohm resistor (R1) to light up like a little holiday light when I adjusted the trimmer. Before I found the problem, I assumed I had shorted out something by accident and replaced the resistor, only to have it smoke up again. See the peeling paint.

When I found and scraped off the hair line short circuit to T3, the resistor problem disappeared.

Here I test the circuit for correct voltages at the DIP sockets, an easy test to do before plugging in the IC's.

The 3d printer method of making circuit boards is cool, but it is tedious. I will try drawing a circuit freehand after drilling the holes for components. Then I can "connect the dots." Here is a paper mask of the ADSR circuit sized to match the original PCB schematic:

Here is the PCB after drilling holes through the mask, mask removed. The black shape is fine steel wool to clean the copper surface. I also wiped the copper surface with 90% isopropyl alcohol:

The circuit diagram drawn freehand with a fine point Sharpie. I wore nitrile gloves to prevent getting any oil from my skin on the clean board.

After etching:

As a side project, I will do more tests to improve coverage of the copper with the Sharpie ink...

A new "fine point" Sharpie works better to cover the copper. I tried it on a small circuit board for a buffered multiple. I designed the circuit board using a schematic from Eddy Bergman's project 25. I sketched a rough diagram, then on the computer I cut and pasted bits from other diagrams to get hole spacings correct for the quad op amp and power connector.

I got uniform copper traces without any breaks:

Here is the finished pcb with components mounted:

The buffered multiple allows me to route the pitch CV from the Keystep to four outputs while maintaining one volt per octave accuracy. Works!

February 2023. I wanted a +/- 12 volt power supply to test modules on the bench. I have a box of old 6-volt wall warts from now-defunked Edmund Scientific. I used two pairs of these in series to give me two 12-volt supplies, ran the outputs through two 2596 voltage regulators to give me steady 12-volt outputs.

I built a Dremel press using some parts from a Thingyverse Dremel press, here: I modified the pipe bearing unit to fit the parts I had:

I made a wooden base and support, added a return spring. This works very well for drilling pcb's.

I finished the second ADSR, gave me more troubleshooting practice. Here it is connected to the bench power supply with a ribbon cable I scavenged from an old computer disk drive cable.

I spent days trying to find problems with the second ADSR. I wondered if the timer chip (7555) was bad, found a simple testing circuit that would make LED's blink to demonstrate proper function here:

The timer is working:

I finally traced the problem to a tiny bit of solder splatter that bridged two copper traces. Here is the completed ADSR-2.

Next, I build a second VCO. Here I start wiring the front panel:

March 2023. Second VCO working.

I had to pull the first VCO out of the case to make point-to-point comparisons with the volt meter to track down problems with VCO-2. I finally isolated the problem at the two diodes. After I replaced the diodes and cleaned up the PCB in that area, the VCO worked.

Here is the synthesizer with the second ADSR and VCO installed.

Mixer next, to mix the outputs of the two oscillators.

I drew the circuit for a 3-input mixer from Yusynth with a "fine" Sharpie to get a good layer of ink. Then I touched it up around the holes with the ultra-fine marker.

Here is the drilled circuit board and the drilled front panel. I followed the same procedure as previously described to etch the board. "Fine" Sharpie wins. The copper is uniform, not streaked as in earlier attempts.

I install the IC after checking for correct voltage at the power pins.

The empty holes in the circuit board are for components in the four-input version of the mixer which uses the same circuit board.

Here is the mixer with its 10-pin connector, ready to test:

The mixer works "straight away" as Yves Usson promised it would. Merci, Yves.

Next, the Yusynth Noise Source/Sample and Hold module.

April 2023. Here is the Noise/S&H module, ready to try:

 White noise and pink noise work, but need level adjustment. The S&H circuit is not working yet.

I can't work on that module because my Windows 10 hard drive fumbled. I decided to switch to Linux. It's taking several weeks to convert because I want Blender, LibreOffice,Waveform DAW, Scarlett Audio Interface, photo editor, HTML editor, browsers, email. I'm trying BlueGriffon HTML editor now as I type this.