home | tumblrblog | musics | electronics | ask | about | more
So I’ve built a couple pedals, and here’s some info about them.
Soon I will have photos but for now you’ll just have to be content with schematics and a video at the end.
The Optical Tremolo Pedal
I had the power supply/volume control pedal from my Guitorgan project just lying around, unused, so one day I decided to do something with it. Obviously it would be easy to mod it into a standard guitar volume pedal since that’s basically what it was, but I already have two, so I don’t need another. I decided to try something new: a pedal-controlled optical tremolo circuit.
The basic idea here is that an light-source such as an LED blinks, and the light from this light-source hits a photoresistor, which changes its resistivity depending on how much light hits it. An ideal photoresistor has an infinite resistance when no light hits its surface and zero resistance when a certain fixed amount of light hits its surface. The guitar’s signal passes through this photoresistor in a voltage-divider scheme, which is basically the same way a volume knob works. In essence, the blinking LED controls the guitar signal’s amplitude.
Morley optical volume pedals work in the same way. Instead of relying on a potentiometer to turn the guitar signal down—which can wear out and get scratchy with time—the pedal simply controls how much light emitted from an LED hits the surface of a photoresistor. Even the famous LA-2A compressor works on much the same idea; it turns down the incoming signal based on how much light hits a sensor, and the light is controlled by an approximately average value of the input.
First up was to design an oscillation circuit that would control the voltage through an LED. The easiest to build are basically square-wave oscillators: they output either a high or a low value. I researched a number of topologies: relaxation oscillators, voltage controlled oscillators…. I eventually settled on probably the simplest oscillator design: the astable multivibrator. It basically consists of two filtered gain stages that are fed back into each other. The feedback makes either gain stage want to change the other, and the filtering prevents them from changing too fast; in other words, the filter controls the rate of oscillation.
One of the simplest gain stages is a common-emitter BJT amplifier, and one of the simplest filters is an RC network. With an RC filter, the corner frequency is inversely proportional to the time constant RC, so changing R will affect the rate of oscillation. In an astable multivibrator, there are two RC filters going on because of the necessary complimentary feedback. In order to control the rate of oscillation in an even way (both “halves” of the square wave have equal pulse width), I would need a dual potentiometer to change the R value of the RC filters. Luckily, I had a 100k dual pot, so I designed the oscillator circuit around that to oscillate at low (LFO) frequencies.
The astable multivibrator circuit doesn’t “play well” with other circuitry, though. What I mean by that is that the impedance of outside circuitry easily affects the RC value of whichever leg of the oscillator you connect the circuitry to. To prevent this, either only high-impedance circuitry may be used with it (like a MOSFET, for example), or it must be buffered with a high-input-impedance buffer. Voltage-follower op-amp configurations are ideal for this; op-amps are designed to have high input impedances, and are usually very stable. I used a dual-op-amp chip, since then I could use the second op-amp as a signal filter to turn the square-wave oscillation into more of a triangle-type wave.
The filter topology I chose was a Sallen-Key first-order Butterworth lo-pass filter in a non-inverting configuration. Saying all that makes me feel all fancy. It’s actually a pretty standard op-amp filter configuration. This final op-amp filter drives the LED that is coupled to the photoresistor, as well as another status LED mounted to the top of the pedal.
The actual guitar-signal circuitry is very straightforward: either the signal goes through the photoresistor when the pedal is on, or it doesn’t when the pedal is off. This is called “true-bypass” switching, which is preferred for those who build their own pedals. Manufactures tend to use JFET switching because it’s cheaper to add a handful of surface-mount components at the mass-production level than to implement a 3PDT switch that most “true-bypass” switching needs. I, however, did not have a 3PDT switch on hand; the original pedal used a simple SPST switch, so I needed to come up with an alternative. I had two relays sitting around from an Arduino starter kit I got a while back, so I decided to use that for the switching. When the pedal is powered up, current flows through the electromagnetically-controlled switches of the relays, switching the guitar signal to flow through the photoresistor. Powered down, the relays switch off. Simple, and effective.
However, relays are much less efficient power-wise than a “passive” switch. They require around 200 mW apiece to switch, for 400 mW total for the pair. This maye not sound like much, but it’s more than this sort of pedal should need. Maybe that’s just me trying to make things more efficient; I don’t know.
Anyway, when I assembled everything and began testing, I was worried that the LED wouldn’t get bright enough to let the photoresistor get to a low enough resistance so that it wouldn’t attenuate the guitar signal when it was suppose to let it pass in the tremolo cycle. To my surprise, however, I found the opposite was true: the photoresistor actually never reaches a very high resistance when the LED is off, despite being wrapped up in electrical tape and in a sealed pedal. What I failed to realize is that this particular photoresistor has a sluggish transient response. Probing the voltage across the LED to ground and the voltage of a test DC constant signal across the photoresistor to ground showed that the photoresistor sort of lagged behind the LED as it didn’t “release” as quickly as I had hoped. What this meant is that it’s like a tremolo effect set to a low intensity; it lightly modulates the incoming guitar signal but doesn’t dramatically turn it on and off.
In the end, I enjoy the subtle, smooth pulsing effect it creates, especially before a distortion stage because it allows the guitar signal to move in and out of the threshold of saturation. Coupled with my simple single-transistor distortion pedal, I’m happy with the way it sounds. I might try changing the oscillator’s signal filter because it may be unnecessary, as the sluggish response of the photoresistor filters the amplitude modulation anyway, and that might make the effect more dramatic and deeper.
A Simple Fixed-Threshold Distortion Pedal
I built this pedal using parts from an earlier project kit that failed. Basically, I tried building it before I really knew what I was doing. The transistor that I was sent with the kit was bad, but I never realized it, and I instead assumed the other parts were bad and tried changing them around without actually understanding what was going on. I’ve since learned a lot more about transistor amplifier design, and this one is technically an original design, but is based off the typical common-emitter BJT scheme and a Big Muff Pi tone stack.
The distortion comes from the “incorrect” biasing on both sides of the BJT to give the output an uneven waveshape, and the gain is set so that the amplification factor is greater than the available voltage (+/- 4.5V around the bias point). Additionally, the input impedance—ideally around 1MΩ—is approximately the value of R2. A high input impedance ensures maximal voltage transfer between two stages, which is ideal since we use voltage as a reference in most pre-amplified audio electronics. R2 is definitely nowhere near this, which affects the way the pedal interacts with the guitar (the original Big Muff Pi has a similar input impedance). Anyway, since the distortion comes from a single transistor, the distortion threshold and gain is fixed slightly above low-level guitar signals but below typical maximum signals. This allows for the distortion amount to be dependent on the incoming signal. Some distortion schemes either clip so low or after a gain stage that they distort any amplitude of signal coming into it. My design allows for a more dynamic distortion response while maintaining simplicity. With an enclosure that has only two knobs to work with, there’s only so much that can be controlled.
The tone stack is taken from the Big Muff Pi. The idea with this passive filter arrangement is that a single knob can attenuate either the highs or lows, and has a steep crossover notch in the mid-range. It is basically a lo-pass and a hi-pass filter with their shoulder frequencies overlapping to create the notch, with a potentiometer that can sweep between the two extremes.
In my original design stage, I failed to take typical guitar characteristics into account, and kept the original Muff filter parameters because it seemed to work well with my design. Once I actually built the circuit and listened to it, the tone stack was much more extreme than I had originally planned. At the low end of the knob, highs were cut considerably, and at the high end, lows were almost non-existent. This is because tone stacks are essentially an extension of the guitar pickups’ intrinsic response. They have an inductance (obviously; they’re inductors) and capacitance, which creates a bandpass-like response. Without having a basic idea of how they interact, a tone design will not accurately reflect how it will operate in the real world. Additionally, the Big Muff Pi’s tone stack is after many gain and clipping stages, so it will interact differently than if directly coupled to a guitar, for example.
That being said, I actually am happy with the way the low-end of the tone sounds, even if it was not what I was expecting. I plan to rework the tone stack eventually, but I’m content with the way it turned out. Also, the exact response of my pedal is due to the transistor I used. It is an unknown PNP-type transistor, found in a large bin of junk parts from my electronics lab. Early breadboarding of the circuit allowed me to try a number of “junked” transistors, and I chose the one that sounded the coolest. I wish I could tell you what it is, but unfortunately it is unmarked in an old generic container.
You can check out this video of me demonstrating the pedals working together:
https://youtube.com/watch?v=C8oLL-AlaoY