My submission (2nd one) was composed from a longform improvisational recording of about an hour of me playing my synthesisers in tandem. I have been doing these a lot, and I find them to be very constructive to understanding my equipment more and more, especially during the playback phase – when deep in the midst of the patch, ducking and weaving through patch cables like a jungle explorer slashing vines, it is difficult to get a full perspective on the tonal qualities of the work. On playback, there are plenty of moments where I go ‘whoah!’, and yet more where I think ‘more work needed on this part of the synth’. It also makes me confront my personality as a performer, which really comes through working with instruments of my own designing.

One thing that these projects have forced me to learn is that it is time to move on in my electronics prototyping journey – my projects are getting to large and complex for stripboard, and I think I will have to make the jump to designing PCB’s. Softwares like Altium, EAGLE, EasyEDA are all on my radar, and I need to figure out which best suits my interests. PCB’s seem cheap to have built, around £1 per board in some cases!

These are all shots from the planning and designing of the various synthesisers that I planned to make for this project. Only one actually made it through all of the trials and tribulations of testing, prototyping and making, which is shown below, still being adjusted and worked on, as it likely will be for the next couple of months as I play with it more and more.

After wrestling with very simple single transistor VCAs with not much success, I arrived at the idea of constructing a low pass gate instead for V2 of the synth. Having worked in the past with DIY vactrols, and then graduated onto transistors as a more elegant solution to the issue of adding voltage control to circuits that rely on the manipulation of resistances for control parameters (ie, feedback paths in opamps), swapping the resistors in a sallen-key filter for BJTs was an obvious leap.

This circuit worked well, although took a lot of refining to get it exactly where I want it.

A combination of stripboarding, breadboarding (both shown above) with a hearty dash of digital circuit simulation (shown below) has led me to arrive at a nearly complete design. This synthesiser is a direct descenant of my first prototype, with improvements and alterations that reflect my experience playing (and performing) with V1.

The main alterations are as follows:

• added noise generator
• added 3 attenuators
• entirely different oscillator design – this is an experiment, and I am not entirely sold on this being the best option, but these oscillators are now voltage controlled, which is a huge step forward.
• the LFO’s have been replaced with pulse/slope generators, with controls for rise/fall time rather than pitch.
• the VCA’s are replaced with lowpassgates, and these lowpassgates will have output sockets on the front panel rather than being hardpatched to the L/R buses.

Thre reasoning for these changes are as follows:

• There was very little room for subtlety in V1, as there was no way to attenuate signals, as well as the fact that it was mostly dealing with square waves at a low frequency level – on or off, 100% or 0%.
• No filters (in V1) had an interesting effect on the playing habits, but I did have some desire for filtering, to experiment with it from a soldering/electronics angle but also from a performance/playing angle. Hence the lowpass gates

I have procured a PDF of the above book, which is a technical approach to designing efficient, low distortion circuits for audio. It mostly details the studio side of things such as preamps, tone controls, gain staging etc, but I am consulting it in an effort to find an amplifier topology that is suitable for driving a small 3.2Ω speaker.

This book, and lots of forum lurking, has led me to the conclusion that 2 gain stages are likely required. I need a low output impedance to drive my speaker, but high current and somehwat medium voltage gain. Some testing (using a common collector topology) has shown that I can achieve very low volumes out of a speaker if I care little for distortion, but tables like the one shown below have made me realise that two stages is the way forward. A first stage that is suited to the output charicteristics of my VCA’s, and a second stage that is suited to the characteristics of my speaker cone.

The output impedance of my VCAs is currently an unknown – I am also considering using a lowpass gate instead. But, regardless, it looks like I need 1 stage to do voltage gain, and another to do current gain. There’s also lots of talk on forums of using a transformer to connect to the speaker, but I need to do more research and find out exactly why this keeps getting recommended. I have 1 or 2 transformers lying around from having ripped apart old portable tape decks… I wonder if those were used as speaker drivers….? Or, power conditioners…?

These shots show a square wave being conditioned to clock a S+H module, which will (hopefully) be a part of a solar powered synthesiser that I am making.

Looking into S+H modules, I found a great youtube video by Moritz Klein that details exactly how they work – however, he specifies that a JFET transistor would be needed rather than a BJT, as BJTs have limited (and directionally uneven) current flow between the collector and emitter. I do not have any JFETs, and don’t fancy learning about them yet – BJTs are tough enough for me at the moment. I think that the future of this solar powered synth is looking very free of any S+H module …

Another issue that I am starting to be aware of is the presence of opamps in the schematic, which do not perform as well as transistors at lower voltages – opamps will simply cut out at lower voltages, whereas transistorised circuits will die in a more interesting and dynamic manner. A redesign is needed.

A day later, I am now digging deep into the world of discrete circuits – opamps are easier to work with but it is common knowledge that, for audio purposes, well designed transistor circuits can outperform their IC counterparts. I’ve already done some research on transistor square wave oscillators (bistable multivibrators), but have just managed to successfully solder a phase shift oscillator, which outputs a sine instead – very lovely to work with sines after months (years…?) of blasting square waves out of my speakers.

Now, the real problem is designing a (relatively) clean output stage to drive a speaker, all at very very low voltages. More on that in the next post.