Digisound 80 style Analog Synthesizer

 

Goal : Construct analog synthesizer design based on Digisound 80 with custom PCB and chassis

Specifications/Research :

    When searching for a project for IEEE's quarterly project, a colleague recommended an analog synthesizer and we concurred that it sounded like a rewarding challenge. He had found a website which included the full schematic along with a lengthily description of its construction and functionality. All that was left to do was collect the components and assemble it. The design was based off the Digisound 80 and included the following features: Square, sawtooth, and triangle outputs; CV, PWM, FM, and Hard/soft sync inputs; and a total frequency range of about 0.1Hz to 50kHz. Commonly referred to resources:

1. Synth base
2. Datasheet for AS3340
   

Progress log: 

• 02-28-23: I ordered most of the parts in one pass from Digikey; only having to source the main oscillator chip(AS3340) and the polystyrene capacitors elsewhere. Once the parts arrived with some solder-able permanent breadboard PCB's, I carefully started laying out the components. I double checked every resistor with my multimeter before soldering it into place. I placed the oscillator chip on one breadboard, the output op-amp(TL074CN) on another, and the 1/8" output jacks onto a square proto-board. For this iteration, we had chose 1/8" jacks over the more common 1/4" jacks for both cost and space efficiently; also we had the ends for both so conversion cables could be made. This was also around the time I used the laser cutter to simple chassis for the synthesizer.  
  
  
  
• 03-01-23: Once the main boards were completely assembled and checked for accuracy, I installed the pre-assembled boards into the chassis and began wiring the potentiometers. I started with the -15V, 0V, and 15V connections before doing the rest. After a final check for correct component layout, it was time to power the synth for the first time with a set of DC power supplies. Unfortunately, the square wave output didn't do anything but the ramp and triangle outputs came through perfectly clean on the oscilloscope. To see how it was sounding, we sent the outputs to a little test speaker ripped from an older project. While extremely quiet, we finally were able to hear our synthesizer do some synthesizing. From what we could tell, the frequency controls worked as intended and the synthesizer produced frequencies all across the audible spectrum. When questioning the quietness of the speaker, some research led us to the conclusion that in reality, it was working as it should. The synth outputs 0-10V waveforms which the speaker sees as a 5Vpp wave with a 5V DC offset. Speakers are not designed to handle signals that do not oscillate around 0V(their negative connection) and we likely destroyed the speaker as soon as we plugged it into our synth.  Over the next couple months, I would develop a circuit to attenuate and level shift the signal in order to output my synthesizer signal to my guitar effects and/or amplification. 
  
  
  

  

  Sawtooth output when testing level-shifter/attenuator

  
  
• 09-25-23: For a variety of reasons, I decided it was time to rebuild my synthesizer completely. Firstly, I regretted my usage of permanent breadboards because they still have many of the same parasitic effects of normal breadboards. To remedy this, I decided to design a custom PCB. I also regretted using 1/8" jacks because of the ubiquity of the 1/4" jack in audio. Next, I could never figure out why the square wave didn't work and I had to assume something was wrong with my wiring or with the chip itself. Finally, I wanted to modifying the output stage to include my level-shifter. I started by making a copy of the synth's schematic in Eagle and exporting it to KiCAD for PCB design. At this point, I had strongly considered using a multi-PCB technique I had seen in a guitar amplifier: the potentiometers, jacks, and main component board were on separate boards connected mostly by removable connectors. This makes assembly and disassembly much faster; while also frankly looking much cleaner. Considering the ratsnest that was my first iteration of the assembled synth, I initially placed my attenuator on the jack assembly. I would move this to the main board when I realized there was going to be an empty space with just the oscillator and output IC's. 
  
  
  
• 10-03-23: It was this re-placement of my level-shifter and the fact that the pots I already had were not made for mounting to PCB's that led me to scrap the multi PCB idea; at least for now. Instead, I assumed that both the pots and output/input jacks would have to be hand-wired again. To account for this, I placed as many of the ports on the outer perimeter(~7cm x 9cm) as I could. I also added four M3 holes on the corners for mounting. After a final schematic check and slight re-arrangement of the text on the board's surface, I ordered a set of five PCB's along with enough components to assemble two boards. 
  
  
  
  
  
• 10-20-23: Propelled by the excitement of seeing my PCB's for the first time, I went to work assembling the boards. Unfortunately, I was short a couple of capacitors and would have to wait for them to arrive until I was able to fully assemble the boards. During the wait, I redesigned my first chassis to sit upside down and include all the necessary ports and switches. I prepared this new chassis for assembly by installing the potentiometers/switches and adding the necessary GND/-15V/+15V wiring.

  

                                                    Unassembled and assembled synth boards                                                                                         
                          New chassis with added line/instrument level switch and 1/4" centered output

New chassis with added banana power inputs and power switch

• 11-7-23: With the rest of the parts in hand, I finished the assembly of the two boards. To install one into my new chassis, I developed a much better method than the first time. Instead of one at a time, I soldered all of the needed wires, with extra length, to the board first and then fixed the board to the chassis with standoffs; all that was left was to trim each wire to its final length before routing and soldering it where it needed to be. For some reason, my output jacks were shorting the ground and output together so I left some of the I/O wires loose to test them. While unloaded, the outputs looked perfect and everything was working as expected except for the square wave. When the instrument level output was connected to a guitar amplifier input, significant waveform distortion started occurring. After some consideration, I realized the culprit was excess output impedance of my module. In addition to reducing the resistor values to achieve a ~5k output resistance, I added 10uF coupling capacitors to the line and instrument level outputs to isolate them from any DC. With these changes, the loaded versus unloaded behavior became nearly identical. The last addition to the synth was a MIDI to CV module which uses an Arduino nano to convert MIDI instrument signals to the proper control voltage for corresponding frequency synthesis. A link to that project's originals source can be found under specifications/research. 

Test setup with labeled I/O wires

Sawtooth 0-10V and instrument level output

Triangle 0-10V and instrument level output

Final assembly with midi2cv module installed

Conclusion :

    Though I can admit I wasn't the most excited about the premise of this project, the creation of this synthesizer has turned out to be extremely fun and fulfilling. It was the first time I designed a circuit from scratch and the first time I had a custom PCB made, so the project felt like a good mix of known and unknown facets. This project also introduced me to the wonderful world of synthesizers, which I now find to be an extremely fascinating topic. This project has inspired a plan for my next project to be a digital synthesizer, something more compact and with access to a greater variety of sounds and interfaces.