Op-amp filters
Tone controls and filters built around one op-amp. Turn the knobs on a Baxandall or the Tube Screamer’s tone stage, or enter a frequency and Q for a two-pole filter and get the parts.
Knobs
Work out the parts
Parts
Swipe sideways for the whole drawing. Tap a part to jump to its value.
How op-amp filters work
An op-amp’s gain is very large, so with feedback its output moves to whatever voltage keeps its two inputs equal. Put caps in the feedback or in the network ahead of it, and the gain changes with frequency. The op-amp’s low output resistance drives the next stage without loading the filter.
The Baxandall The bass network and the treble network each run from the input to the output, and each pot’s wiper connects to the minus input. With a knob at noon its network is balanced, and the gain is 1 at every frequency. Turned toward the input end, more of the signal reaches the minus input than comes back from the output, so that band is boosted. Toward the output end it’s cut. C1 and C2 short the bass pot above the lows, so the bass knob only works down low. C3 and C4 let only the highs into the treble pot. The most boost or cut is set by the resistors on each side of the pots.
The TS tone stage R7 and C5 cut the treble a little ahead of the op-amp. The Tone pot runs from the plus input to the minus input, and R8 and C6 hang on its wiper. Toward the plus input the branch shunts the treble to ground before the op-amp, a cut. Toward the minus input the branch pulls more treble current through R10, so the op-amp’s gain rises in the treble, a boost. With the knob at noon the stage still cuts the top a little, part of the mid hump the pedal is known for. The same network, with the clipping stage ahead of it, is on the tone stack page.
Two poles A filter with two caps cuts twice as steeply as one resistor and one cap: 12 dB per octave beyond the filter’s frequency. Q sets the shape of the response near that frequency. At a Q of 0.71 (a Butterworth filter) the response is as flat as it can be up to the cutoff, and it is 3 dB down at the filter’s frequency. With Q under 0.71 the bend is more gradual. With Q over 0.71 the level rises to a peak before it falls. A band-pass filter passes a band whose width is its center frequency divided by Q.
Sallen-Key The op-amp is wired non-inverting. One part connects from the output back to the point between the two input parts, and that feedback raises the response near the filter’s frequency. With a gain of 1, the ratio of the two caps sets Q in a low-pass, and the ratio of the two resistors sets it in a high-pass. Adding gain (Rf and Rg) raises Q too, and the band-pass form needs it. In the band-pass form Q climbs fast as the gain nears 3, so a 5 % resistor there moves Q a lot. With a real op-amp, a Sallen-Key low-pass stops cutting at very high frequencies and the level comes back up. The feedback cap passes the input straight to the output, and up there the op-amp no longer has the gain to cancel it.
Multiple feedback The op-amp is inverting, with its plus input on ground, and two parts feed back from the output into the network. The gain is set by the parts, −R2 / R1 in a low-pass, −C1 / C2 in a high-pass. The stage inverts the signal. As a low-pass it keeps cutting at very high frequencies better than a Sallen-Key does.
The op-amp’s bandwidth An op-amp has less gain at higher frequencies. Its gain-bandwidth product is the frequency where its gain is down to 1, about 3 MHz for a TL072 or a JRC4558. When a filter needs nearly as much gain as the op-amp has left at that frequency, the response no longer matches the ideal one. The gray dashed curve shows the same parts with an ideal op-amp. At guitar frequencies most op-amps have gain to spare, but a high-Q filter, a high gain or an old part like the 741 can show a difference.