UNVERIFIED. Drawn by my own layout script and checked in Copper Bottom, not built. The netlist matches the board, but the editor's checks do not come back clean on this one: it reports three pin-shorts, one per stage, because each stage ties an inverter's output to its own input and to the output of the stage before it. That is the circuit, not a fault - a self-biased inverter is the resistor here - and the checker cannot tell a trick from a mistake. Nobody has soldered this board or heard it.

CMOS Ladder Filter

One chip, one transistor, a cutoff knob that starves it · By Onion Madder · Mess O' Pedals

A lowpass filter with no resistors in the signal path. A CD4069 is six inverters, and an inverter with its output tied to its input sits at half the supply and behaves like a resistor to the middle, a few hundred ohms to a few thousand depending on how much voltage the chip is given. So each stage here is an inverter driving a node, a second inverter tied to itself to be the resistor, and a 47nF cap to ground. Three of those in a row is an 18 dB per octave ladder. The cutoff knob is the chip's supply: a transistor follows the pot and feeds the 4069 anywhere from about two and a half volts to about seven, the inverters get slower and softer as it drops, and the corner moves with it. Starving a chip on purpose, with a knob.

The layout file is the guide

Like every layout on this shelf, the build document for this one is the file itself: ↓ Download CMOS_Ladder_Filter.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.

Ground is the bottom strip. GND, GND2 and the cutoff pot's BOT pad all sit on strip j, and the three filter caps and the divider's bottom resistor return to it. At the very bottom of the cutoff knob the chip's supply dips under the three volts the datasheet gives as a 4069's floor, so expect the last of the travel to choke it, which may be the point. Copper Bottom will report three pin-short errors on U1. Read them before you dismiss them: each names one stage's node, where a driving output, a self-tied inverter and the next inverter's input all share a strip on purpose. If it reports a fourth, that one is real.

The netlist

My script wrote this from the design, not from the copper, and the board was checked against it in Copper Bottom: every connection it names is on the board and nothing on the board is missing from it. The lines to find are Y1, Y2 and Y3 - one per stage, each carrying two outputs and two inputs of the same chip, which is the whole trick - and VDDF, the chip's own supply pin hanging off the transistor's emitter.

The off-board wiring

Eight pads leave the board: the cutoff pot takes three, the level pot two, the two jacks and the supply the rest.

How it works, in one breath

The signal comes in through C7 to the first inverter, which R3 holds in the middle of its swing so it works as an amplifier. There is no input resistor, so its gain is set by whatever is plugged in, and with a guitar that is a lot: expect this stage to bite. Its output drives node Y1. On Y1 sits the second inverter with its output tied to its own input, which parks it at half the supply and makes it fight anything that tries to move the node, gently, like a resistor to the middle. With C1 to ground that is one lowpass pole. The third inverter reads Y1 and drives Y2, where the fourth inverter is the resistor and C2 the cap, and the fifth drives Y3 with the sixth as the resistor and C3 the cap. Three poles, three inversions, out through C8 to the level pot.

How strong that resistor is depends on the chip's supply voltage, because an inverter's transconductance does. That is the cutoff control: the pot is a divider between R1 and R5, its wiper goes through R2 into Q1's base, and Q1's emitter is the 4069's pin 14 with C4 to keep it steady. Turn the pot down and the chip runs on three volts, the inverters go weak and slow, the poles drop and the whole thing gets darker and quieter. Turn it up and it opens. The rest of the board runs on the full nine volts; only U1 is starved. That is a read of the circuit, not of a build, and the exact range and how it sounds at the bottom are the two things only a board can answer.

Things to try

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

• Change the three 47n caps together. 100n moves the whole range down, 22n up. Change one and the poles spread instead
• Put a resistor in series with the input. 10K to 100K between C7 and pin 1 turns the first stage into a proper inverting amplifier with a gain you chose, if the bite is too much
• Swap R1 and R5 for other values to move the ends of the cutoff range. The chip will not do anything useful below about three volts, so keep R5 in
• Feed it a square wave. Any 40106 voice on this shelf into IN, and the filter does what a filter does to a square wave. The XOR Cross-Mod is the obvious partner

Using a 4069's inverters as resistors and starving the chip to move the cutoff is an old trick - the EDP Wasp built its filter that way in the seventies - and it belongs to whoever first noticed that an inverter tied to itself sits in the middle. The arrangement here, three stages on one chip with a follower on the cutoff pot, and the layout, are mine and so is anything wrong with them.