« Back to DIY

UNVERIFIED. This layout is drawn and checked in Copper Bottom - the DRC is clean and the netlist below matches the copper - but nobody has built it, me included. Every number on this page that is not a part value is arithmetic off the datasheets, and the page says which is which. Build it, fix anything that fought back, and the banner comes off.

Analog Bit Crusher

A sample-and-hold with the clock on a knob · By Onion Madder · Mess O' Pedals

This is a bit crusher that crushes the sample rate and not the bit depth, which is the half of the sound most people actually mean. It takes a snapshot of the signal, holds it flat until the next snapshot, and the knob decides how often that happens. Turn it down far enough and a guitar comes out the other side as a staircase. There is no code in it and no converter: the sampler is one 2N5457 JFET used as a switch, the memory is a 1nF capacitor, and the clock is half a TL072 chasing the other half.

The Analog Bit Crusher layout drawn as a stripboard diagram, titled Analog Bit Crusher, 16 by 14 stripboard. Fourteen orange copper strips run left to right, lettered a to n down the left side, with sixteen columns numbered above them. Two pale rectangles sit one above the other in the middle of the board, the upper marked IC2 TL072 and the lower IC1 TL072, each with a column of gray X marks through the strips between its two rows of pins. Outlined resistors stand upright between the strips across the whole board, with small green film capacitors, pink electrolytics and three outlined diodes. A small transistor outline marked Q1 sits at the left on the middle strips with its legs lettered G, S and D. Gray X marks show every cut strip. Holes carrying wires off the board are labeled down the right margin: FREQ_2, GND, FREQ_1, LED+, 9V and OUT, with IN alone on the left. Beside the board the panel parts are drawn with a wire to every hole they land in: a potentiometer marked POT1 B100K rate, a red LED, a DC jack marked 9 V and two quarter-inch jacks. A parts list runs underneath and a credit line reads Made with Copper Bottom. Below it the same board is drawn a second time from the solder side, mirrored, with the column numbers running right to left, every solder point as a dark dot, the cuts as X marks, both chips as dashed outlines with pin 1 boxed and the pads labeled on the margin they now sit nearest.
The board as Copper Bottom draws it: the component side on top, then the same board from the solder side with the columns running the other way, so a cut is the same hole in both pictures. The panel parts are drawn beside it with a wire to every hole they land in.

What you'll need

tap to expand / collapse

Thirty-one parts on the board and five off it. Nothing here is exotic; the JFET is the only part worth buying carefully.

  • 2× TL072 dual op-amp, DIP-8 (IC1, IC2) - spot it: eight legs, TL072 printed on top, a notch at one end. One is the signal path, the other is the clock. A TL082 or an NE5532 drops straight in; do not use an LM358 here, its output will not swing far enough for the clock to behave
  • 1× 2N5457 N-channel JFET (Q1) - spot it: TO-92, three legs, one flat face, 2N5457 on the flat. This is the sampler. A J201 or a 2N5458 works. It is not a 2N3904 and a 2N3904 will not do anything useful here
  • 4× 1M resistor (R1, R2, R7, R8) - spot it: bands brown-black-green. Two hold the jacks at ground, one biases the input, and R7 is the one that lets the clock drive the JFET's gate without loading it
  • 4× 100K resistor (R4, R5, R6, R9) - spot it: bands brown-black-yellow. Two make the half-supply reference, two set how far the clock swings
  • 1× 22K resistor (R3) - spot it: bands red-red-orange. It sits in series with R4, and the pair of them is deliberate - see the netlist note
  • 1× 100R resistor (R10) - spot it: bands brown-black-brown. Easy to mistake for the 1K family; the third band is the difference. This is the clock's fast path and the floor under the knob
  • 1× 2K2 resistor (R11) - spot it: bands red-red-red. The LED's series resistor, and the only part on the board that does nothing for the sound
  • 2× 1 µF electrolytic (C1, C2) - spot it: a small can with a stripe down one side. In and out coupling; the stripe faces the jack on both of them
  • 1× 10 µF electrolytic (C4) - decouples the half-supply reference. Stripe to ground
  • 1× 100 µF electrolytic (C3) - the supply reservoir. Stripe to ground
  • 1× 1nF film or ceramic (C5) - spot it: printed 102. This is the memory. Use film if you have it: a cheap ceramic leaks and droops, and droop here is the sound sagging between samples
  • 1× 3n3 film or ceramic (C6) - spot it: printed 332. The clock's timing cap, and the part to change if you want a different range of sample rates
  • 2× 1N914 diode (D1, D2) - spot it: tiny glass body with a black band at one end. A 1N4148 is the same thing. They point in opposite directions on purpose and the layout shows which way
  • 1× 1N5817 Schottky (D3) - spot it: fatter black body, band at one end. Reverse-polarity protection on the supply. It stands on end - see the note under the layout file
  • 10× wire links (J1 to J10) - offcuts of component leg. J10 is the one diagonal, under IC2

Off the board

  • 1× B100K linear pot - the rate knob. Wired as a variable resistor: wiper tied to one outer lug, that pair to FREQ_1, the remaining lug to FREQ_2. R10 on the board is its floor, so this one needs no extra resistor in series
  • 1× LED and nothing else - anode to LED+, cathode to ground. R11 is already on the board
  • 2× 1/4″ jacks - in and out. 3.5 mm works just as well if this is going in a synth box rather than on a pedalboard; nothing in the circuit cares
  • 1× DC jack, 9 V, center negative or center positive as you like - D3 means a backwards plug does nothing at all
  • No bypass and no footswitch. There is none in the layout file, so there is none on the drill template either. It is always on; wire a true-bypass loop around it if you want one

The layout file is the guide

Like every layout on this shelf, the build document for this one is the file itself: ↓ Download Analog_Bit_Crusher.json - import it into Copper Bottom and the editor gives you the board picture, the placement walkthrough, the electrical checks and a printable build sheet, all from the same file. No coordinates are written out on this page on purpose: the editor is the viewer, and a number typed twice is a number that can drift.

Two parts stand on end, and that is deliberate. D3, the 1N5817, has two strips to cross and a DO-41 body is longer than that, so it stands up rather than lying flat. R11 does the same between its two strips. Standing an axial part up is ordinary and it works; all it changes is how tall the finished board is, which is what catches you when the lid goes on.

The netlist

Quoted verbatim from the layout file. It was written from the board, so checking the board against it can only pass - what it is for is carrying the circuit somewhere else, and telling you what each strip is supposed to be doing while you stare at it.

R3_A is a net with nothing on it but two resistor legs, and it is not a mistake. R3 (22K) and R4 (100K) sit end to end, so what the clock actually sees is 122K. Two parts where one would do, because 122K is the number the hysteresis wants and 122K is not a value you can buy. If you would rather fit one resistor, 120K goes in R4's holes with a link where R3 was.

The off-board wiring

Seven pads leave the board: one knob, one LED, two jacks and the supply.

PadGoes to
INInput jack tip. The sleeve goes to GND
OUTOutput jack tip. The sleeve goes to GND
FREQ_1 / FREQ_2The rate pot, B100K. Tie the wiper to one outer lug and run that pair to FREQ_1; the remaining outer lug goes to FREQ_2. A variable resistor, not a divider - and unlike the pitch pots on the Cicada NAND it needs no resistor of its own in series, because R10 on the board is already doing that job
LED+LED anode. The cathode goes to GND. R11 is on the board, so the LED wires straight to the pad
9V / GNDThe supply. DC jack positive to 9V, negative to GND, and GND is also both jack sleeves and the LED cathode. D3 is already in the way of a backwards plug

How it works, in one breath

The signal comes in through C1, gets biased to half the supply by R8, and lands on the first half of IC1 wired as a follower - output tied straight back to its own inverting input, so it copies the signal and drives the JFET without loading anything. Q1 is the switch. Its drain sees that copy; its source sits on C5, the 1nF. When the clock pulls Q1's gate up, the JFET conducts and C5 charges to whatever the signal is doing at that instant. When the gate goes back down the JFET turns off and C5 is left holding that voltage, with nowhere to leak to: the second half of IC1 is another follower and a TL072's input draws essentially nothing. That held voltage leaves through C2. So the output is the input, redrawn as a staircase, and the width of each step is the whole effect.

IC2 is what sets that width, and it is the oldest trick in the op-amp book: a comparator and an integrator chasing each other. The second half integrates - C6 across it, current in through R10 - so its output ramps. The first half watches that ramp against the half-supply reference through R5 and the R3 plus R4 pair, and flips its output the moment the ramp goes far enough. The flip reverses the current into the integrator, so the ramp turns round, and round it goes. The two diodes are why the knob works. D1 lets the charging current take the short way through R10 alone; D2 sends the discharge the long way, out through the pot and back. So one half of every cycle is fixed and fast, and the other half is as long as you ask for.

The arithmetic, and it is arithmetic and not a measurement. With C6 at 3n3 and the pot wide open at 100K, one half-cycle works out around 0.5 ms, so the sample clock lands near 2 kHz - a heavy crush, roughly a telephone made of gravel. With the pot at zero the only thing left in that path is R10, and the clock goes as fast as a TL072 can be bothered to go, which is well above hearing: that end of the travel is the effect switched off. The one number worth measuring first is the sampling window - the fast half-cycle is short, and C5 has to charge through the JFET inside it. If it comes out dull or quiet at the crushed end, that is the window being too short, and the fix is a bigger R10 (try 1K) or a smaller C5.

Things to try

None of these have been tried, because none of this has. They are the obvious knobs.

  • Change C6. It sets the whole range. 10n takes the slow end down to a few hundred hertz, which stops being a crusher and starts being a ring modulator with delusions; 1n moves the useful part of the sweep up where it is gentler.
  • Put a pot in series with R7. The gate drive decides how hard the JFET switches, and softening it blurs the edge of each step rather than the rate.
  • Take the clock out to a jack. IC2 pin 1 is a square wave with the rate on a knob and nothing else using it. That is a usable LFO or gate for anything else in the box.
  • Feed the clock from somewhere else. Lift R7 from IC2 pin 1 and drive it from another oscillator instead, and the crush follows whatever that is doing.
  • A wet/dry blend. The dry signal is already buffered at IC1 pin 1. A pot between there and the output, against C2, mixes the staircase back under the real thing.
Don't eyeball the holes. There is a 1:1 drilling template for the 125B - the face of the box lying flat, 66 by 121 mm, tape it on and punch through it. Four holes: the LED and the rate knob on the center line, in and out along the bottom.
The DC jack is not on the sheet - it goes on a wall. Two 1/4″ jacks and a barrel will not sit inside the lid lip: a jack wants 16 mm of room and the barrel 12, and there are only 42 mm of face to share between them.
↓ Or with the jacks on top
Print at 100% / “Actual size”, never “Fit to page” - check the bar measures exactly what it says before you touch a drill. The dashed line is the keep-out for the lid lip and the corner bosses.
Nobody has drilled this one yet. The hole sizes come from the parts the layout file names; where they sit is this format applied to them. Offer the sheet up to your own box and your own parts before you make a hole.

The sample-and-hold is as old as the JFET and belongs to nobody - it is the front half of every analog-to-digital converter ever made, used here without the digital half. The comparator-and-integrator clock is the same. The board is Onion Madder's, drawn in Copper Bottom.