LFO: tremolo, vibrato and phaser oscillators

The slow oscillator behind a tremolo, vibrato or phaser, on its own. Pick one, turn the speed, change any part, and see how fast it runs, how big it swings, what shape the wave is, and whether it starts at all.

Swipe sideways for the whole drawing. Tap a part to jump to its value. The red numbers are DC voltages to ground with the oscillator stopped. Tap a dot on a wire to watch that point on the scope. The ringed one is on the scope now.

Two cycles, running steady

Across the Speed knob

frequency, Hz (left)swing at the reference node, V peak to peak (right)where the signal comes back around the loop in step

Loop gain

gain round the loop, dB (left)phase, degrees (right)

Start-up

swing at the reference node, each cycle

Harmonics

evenodd dB below the fundamental, 2nd to 6th, at the node on the scope

All values

How a phase-shift oscillator works

A triode inverts its signal, which is half a cycle of phase shift. A ladder of caps and resistors runs from its plate back to its grid, and each section of the ladder shifts the phase a bit more. At one frequency the ladder’s shift adds up to the other half cycle. At that frequency the signal arrives back at the grid in step with itself, and the loop keeps itself going. That frequency is the speed. The Speed pot is one of the ladder’s resistors.

The ladder also weakens the signal a lot, by a factor of 29 in the textbook version, so the tube needs more gain than that for the oscillator to start. Once it starts, the swing grows until the tube runs out of room, with grid current on one side of the swing and cutoff on the other. That sets how big the swing is and how far the wave is from a sine. Clipping and grid current also shift the phase, so the real speed is slightly lower than the speed worked out from the loop gain.

How an op-amp LFO works

The op-amp version works by switching back and forth. Part of its output comes back to the + input, so the output snaps all the way high or low and stays. A resistor charges a cap from the output, and the cap feeds the − input. When the cap reaches the switch point, the output flips, the switch point moves, and the cap heads the other way. The output is a square wave, and the cap ramps between two switch points. The resistor and cap set the speed, and the feedback resistors set how far apart the switch points are. It starts by itself, and the switch points set its swing.

The ramps are curved because a cap charges more slowly as it gets close to the voltage it is charging toward. On the Phase 90 both switch points are above the middle of the output’s swing, so the cap climbs more slowly than it falls and the wave is lopsided. The figures list the switch points and the speed worked out by hand from the op-amp’s output limits and the divider. At the fast end the simulated speed is slower than that, because the small cap on the − input makes each flip slightly late.

Watching it run Everything moves at the oscillator’s speed. The round light by the waveform’s heading pulses with the node on the scope, like a tremolo pedal’s speed light. Each wire on the schematic glows brighter as its node swings high and fades as it swings low, scaled to its own swing (under 0.05 V stays dark). On the tube versions the glow moves around the ladder from plate to grid. Each node peaks slightly later than the one before. That delay is the phase shift, and around the whole ladder it adds up to half a cycle. On the op-amp versions the output’s wires snap on and off while the cap’s fade up and down. ⅛ speed slows it all down, and Still stops it.

Loop gain For the tube oscillators we cut the loop open at the grid, send in a small test signal, and measure what comes back. At the frequency where the returning signal is in step with the test signal, the oscillator runs if the loop gain is over 1 (0 dB). The further over 1 it is, the faster the oscillator starts and the harder it clips.

What each one drives In the blackface amps the second triode pulls current through a neon lamp inside a light cell. The lamp’s light changes the cell’s resistance, and that resistance turns the signal up and down. On this page the lamp is a fixed voltage drop, so it loads the oscillator as in the amp. The light and the cell aren’t modeled. In the 6G7-A Bandmaster a cathode follower feeds the vibrato circuit, which moves two tubes’ bias in opposite directions. The earlier 6G7 feeds it straight from the plate through 4.7 MΩ, barely loading the oscillator. The 6G16 Vibroverb’s “vibrato” is a tremolo: the follower’s swing goes through the Intensity pot onto the power tubes’ bias. In the Vibro Champ a follower feeds the Intensity pot, whose wiper moves a preamp tube’s cathode so that stage’s gain rises and falls (that tube is included so the pot is loaded by a real cathode). In the Phase 90 the cap’s ramp goes through 3.9 MΩ to four JFET gates, with the bias trimmer feeding the same line through 1 MΩ, so about a fifth of the ramp gets there. The gates swing a few tenths of a volt, moving each JFET’s resistance and the phaser’s notches.

The drawings’ voltages The drawings don’t say whether the oscillator was running when the voltages were measured. A meter on a swinging node reads its average. With the oscillator running, the plate’s average is well above its voltage at rest, so both are shown beside the drawing’s figure.

Warnings Red text flags an oscillator that doesn’t start, one still growing at the end of the run, or the solver not settling.