A better analog 「BIT CRUSHER」

Maybe we should call it a sample rate reducer

This one comes with a Disclaimer. Some of the potentiometer values shown in the schematic sections are not the values used in the final design. Some tapering or stopper resistors may be needed. The design should work as shown but anyone building one of these will want to tweak the range of particularly RV1. The Vref arrangement is also not ideal and doesn't reflect the final version used.

Some of these choices might seem a bit odd

Background

This was originally designed in collaboration with another musician, who did not have a background in electronics. As such some of the design choices may seem slightly odd at first glance. The goal of the design was to create a 「BIT CRUSHER」 type pedal with low noise and a clean blend. Part way through the design process the collaborator decided that they would like a transformer output stage and specifically mentioned the Hudson Broadcast as something that they felt gave them a good clean boost and uses an output transformer. This is an unusual use of the Hudson Broadcast, which is typically used as a fuzz, the transformer use in this is also the subject of many online debates regarding its tonal effects. It is now prudent to discuss the output stage.

Designing a transformer output stage out of spite

This stage was designed in collaboration with aotmr, whose contributions to the DIY effects scene are greatly missed.

The specific output stage mentioned is the Hudson Broadcast (schematic below). Experienced DIYers will be able to tell from a glance that the transformer in this circuit does not contribute meaningfully to the overall tone. The talented Dylan159 has designed an excellent alternative to this schematic and has demonstrated here that the output stage is purely for isolation purposes. Isolation is not required for guitar and bass effects so the only logical conclusion is that the transformer is included here to differentiate the broadcast from the vast number of other boutique fuzz pedals that are currently available. With the exception of the output stage the Hudson Broadcast is an interesting design and produces an exceptionally good sound. It is disappointing and somewhat disheartening that a somewhat deceptive use of a transformer is required to add weight to a circuit that would otherwise on its own be noteworthy.

With that in mind it was decided that we would include a stage that would actually make use of such a transformer. The design is straightforward and is shown below. C10 is connected to the output of the previous stage. U1A is a simple phase inverter and is used in a push/pull configuration with U1B to produce approximately 3dB of signal gain. A demonstration of this stage can be found here. Refer to this series of articles by Elliot at sound-au for more information regarding transformers and the specifics of how this stage functions.

An excerpt from the full schematic containing just the transformer based output stage

What is a「BIT CRUSHER」and how does it work

Some of the schematics featured below are heavily inspired by work done by Parasit Studios. Fans of strange guitar fuzzes and synths should definitely visit their pages.

「BIT CRUSHER」 is often used to describe the sound of what is actually a sample rate reducer. 「BIT CRUSHER」is a more fun name so will be used going forwards. 「BIT CRUSHING」 involves two distinct sections, an audio rate LFO and a sample and hold (usually JFET based). The sample and hold is toggled at a frequency in the audio band (roughly 20Hz - 20kHz), which converts a smooth sine based source signal to a stepped signal and produces an audible effect similar to aliasing effects found in early digital audio equipment.  Many common 「BIT CRUSHER」designs will introduce a significant amount of clock noise into any signal chain even when bypassed. When the effect is in use this is not an issue as the clock noise forms part of the central 「BIT CRUSHER」sound. When the effect is bypassed however the added noise is highly undesirable. This noise was hypothesised to be due to the current draw associated with the constantly running LFO. An additional envelope not typically found in a「BIT CRUSHER」is included to reduce this noise

Sample rate reducers and how they「BIT CRUSH」

Sample and hold circuits are well documented partly as a result of their popularity in modular synthesisers.  Many articles exist discussing the finer points of their design.  This section will be a brief overview aimed at the lazy readers.  A useful youtube video discussing the design of these stages by Moritz Klien was used heavily as a reference during the initial stages of this project. 

An excerpt from the full schematic containing just the sample and hold section and input stage

• H2 is a pad for wiring an offboard switching assembly.  The design of this assembly is left as an exercise for the reader. 

• U1C functions as a simple input stage that can be found throughout audio electronics, with RV1 used to control the overall input gain. 

• RV2 provides the ability to blend the unmodified signal with the modified signal to increase the range of available tonal options.  

• U1D is used as a buffer as the output from Q1 and C7 will not be sufficient to drive the rest of this circuit.

• Q1 is used as a switch in this arrangement. 

• When Q1 is open, the output of U1C is connected directly to R8 and C7, which causes the input of U1D to simply be this signal.  In this configuration C7 is charged by the output from U1C to the signal voltage.

• When Q1 is closed however, the output from U1C is no longer connected directly to C7.  C7 is then discharged through R8. Provided that the capacitor is large enough to not significantly discharge this will result in a constant voltage at the input of U1D that is equal to the signal voltage at the time of Q1 switching

• By toggling Q1 repeatedly using an LFO it is possible to create the aliasing like effect at the heart of every「BIT CRUSHER」.   

Audio rate LFO and envelope

The Parasit Studios Sonic reducer and Dylan159s' Enveloping Timbre heavily inspired this section. Both are designing interesting new effects fairly regularly so a visit to their respective pages is always worth it.

Any oscillator capable of producing a square wave between 20Hz and 20kHz is acceptable.  Oscillator design is discussed extensively by Eliott at sound-au in this article.  A comparator oscillator was chosen to avoid wasting a second opamp stage as the remainder of the circuit features 4 opamps and so could be built using a single TL074 chip.  Comparators are also reported to demonstrate increased stability and a reduction in unwanted power supply coupling compared to traditional opamp based oscillators.  

RV3 is used to set the rate of the oscillator.  The values shown will allow for a range of approximately 30Hz - 1.5kHz.  This range was chosen through experimentation and produced a wide variety of sounds.  RV4 is used to adjust the duty cycle of the square waves generated to control the resolution of the sample and hold.  Setting the duty cycle such that the JFET is open for as short as possible results in a very sharp stepped wave as U1D is only exposed to the signal for very short periods of time.  Increasing the duty cycle will reduce the sharpness of the steps and will result in a less intensely 「BIT CRUSHED」sound.

C6 is connected directly to the output of U1C and serves to connect the envelope network formed by Q2 and Q3 to the main audio signal.  The envelope used is a simplified version of the setup described by Dylan159 in the article above, which contains a detailed write up of the design and function of sidechained envelopes.  Here  the envelope is used to pull the comparators negative input to ground when no output from U1C is detected.  This causes the oscillator to stop when it is not required for「BIT CRUSHING」. This has successfully eliminated the noise from the bypassed state, which proves the initial hypothesis was correct. 

Final Words

This schematic will function as presented. As mentioned previously, individual values for potentiometers and one or two less important resistors changed somewhat during the final build process. This pedal was built in autumn 2024, and is no longer in my possession, so the final value for RV1 is lost. Prior knowledge of opamp gain stages is assumed and therefore it should be trivial to organise a suitable substitution. RV3 can also be varied to taste if a different or narrower range of aliasing frequencies is desired.

This design would not be possible without the help of aotmr, Dylan159, many other members of the Audio Electronics DIY discord server, Parasit Studios and sound-au.com I am truly standing on the shoulders of Giants.