Monday 1 November 2010

MFOS Sequencers

The MFOS SoundLab MiniSynth is a voltage controlled synthesizer using a voltage per octave scale. To play it you need a musical keyboard with a CV (control voltage) output. You can buy them, but they're rather expensive. You can also DIY them by salvaging a keybord from a defunct electronic organ. I might do that sometime. Another way of "playing" a synthesizer is by feeding it control voltages through a device called sequencer. That's a device with a load of knobs by which you can set up and then repeatedly step through a number of voltages (usually 8 or 16) that you can feed into a synthesizer in order to get those repetitive bass lines and arpeggios that electronic music is famous for.

A sequencer is actually a fairly simple device to DIY, and there are simple designs galore on the internet, some using only very few components. When I get to it, I'll definitely design and build my very own sequencer, but to play with the SoundLab I wanted something simple fast. So I ordered the PCB of Ray Wilson's 10 step sequencer from Music From Outer Space (actually, I ordered and stuffed the PCBs of Ray's 16 step sequencer before I bought the smaller 10 step one. It's still unfinished but that's another story...). With my soldering skills steadily improving, stuffing the PCB was quick and painless. As usual with my DIY projects it was the housing that slowed me down. I finally decided to craft a small wooden box to house my sequencer. The front panel was to be a thin wooden board that I shielded by glueing a piece of aluminium sheet to the back. The labelling on the front consists of a printout of my design that I laminated, cut, and then stuck to the front (self-adhesive lamination sheet). It ended up way better than I expected.

When I tested the sequencer I found that the muted steps somehow leaked through to the synthesizer: I could still hear notes that were supposedly turned off. After debugging and signal tracing for hours I couldn't find a fault with the sequencer. I just hadn't played with the synthesizer long enough to discover that there's a dedicated knob to set by how much the AR envelope generator affects the volume of the sound. If, for example, set to mid-level the AR (and hence the sequencer's gate) control only the top half of the volume range while the bottom half is just sounding on, all the while changing notes according to the sequencer's changing voltage. Who could have known!

Friday 1 October 2010

PAiA FatMan

While looking for a synthesizer that I might be able to assemble I stumbled across the FatMan by PAiA Electronics. It looked impressive but I didn't want to spend some US$ 250 on a DIY project that might as well end in the waste basket. I shouldn't have dismissed it so fast. First, when I built the MFOS Mini-Synth I eventually ended up spending at least as much on parts and stuff. But second, the FatMan is a complete kit with all the necessary parts and excellent documentation. All in all it makes a great first SynthDIY project. I liked the look of the synth, and also needed some additional experience in Synth DIY. Their summer sale made it hard to resist, too, so I ordered a kit plus desktop case.

The FatMan is a complete monophonic synthesizer controlled via MIDI. It features two square wave oscillators that sound in tandem but can be detuned against one another up to several octaves, one VCF, one VCA, and two attack sensitive envelope generators (an ASR for the filter, and an ADSR for the amplifier).

Construction is straight forward. You start by cutting and soldering wire jumpers (actually a great way to practice your soldering skills: wires are very hard to fry... ;-), then resistors (a lot of them), then the rest in order of increasing delicacy (tendency to overheat and destruct).

The kit contains sockets only for the microprocessor and the EEPROM. All other ICs are soldered directly onto the board. I'm not the most advanced solderer, so in order to avoid overheating the more delicate components I tend to heatsink transistors using a paperclip and otherwise insert several components at once and then solder one lead of each component in a round robin way, thus giving them a little time to cool off. Mounting and wiring the panel controls was a chore, but I knew that already from the SoundLab, which had even more controls.

When I finally turned it on for what they call "smoke test" (go figure), I was surprised that the FatMan worked right from the beginning! All it required now was calibrating it to a proper octave measure. I did it "by ear" and used my DX7 to compare it with.

The FatMan has a nice warm analog sound especially well suited for fat bass lines or wild spacey solos. It's a neat little beast.

Thursday 19 August 2010

MFOS SoundLab Mini-Synth PLUS

The first Synth DIY project that I more or less successfully completed was a SoundLab Mini-Synth PLUS from Ray Wilson's site Music From Outer Space. I decided to purchase the necessary PCBs after having researched other options, in particular the Simplesizer project on a German Synth DIY site. Unfortunately, that project was abandoned a few years ago, and PCBs are not available anymore.

While waiting for the PCB to arrive and checking various suppliers for the necessary electronic parts I found out about an enhancement to the original design (the actual "PLUS" part). The PLUS enhancements carried the warning that they'll involve several kludges on the board and therefore require having some additional skills that I quite probably lacked. I boldly went for it anyway.

Building it was time consuming because I needed to learn basic skills such as soldering "on the spot" (both figuratively and literally). The desoldering pump was in frequent use. Also, I couldn't quite decide on the cabinet for the synthesizer which slowed things down for a while. Furthermore, I wanted a different panel layout than the one suggested by Ray Wilson. This required that I modified the wiring layout (i.e. the connections between the various controls on the panel).

When I turned on the Mini-Synth for the first time, it appeared to work! Mostly, anyway. The second VCO was silent. That spelled trouble in capital letters. I had no equipment to speak of, and was supposed to signal-trace and find a problem on a circuit that I basically didn't understand. To pinpoint the problem, I reasoned, it might be helpful to tune both VCOs to about the same frequency so I could check each of the circuits and compare the corresponding voltages (I had one functioning VCO to compare the malfunctioning one with). It took me a lot of studying circuit diagrams and a night's sleep to eventually notice that the same wire cannot have 0 volts at one and 7 volts at the other end. A close inspection with a looking glass revealed a dodgy soldering joint at a transistor.

After having fixed that, the SoundLab worked perfectly. I hadn't calibrated it to a proper Volt/octave scale yet. Nonetheless, turning knobs and throwing switches and listening to the noises that it produced was a lot of fun. Most importantly, though, I was hooked.