Two-formant vowel filter with fuzz, 1970 style

Pedal and tube amp circuits, solved part by part the way SPICE (the circuit simulator engineers use) does it, while you listen and watch. Turn the knobs, change any part, or tap a wire to put a probe on it.

How to use the bench

Every part is simulated. Nothing you hear is a recording.

  1. 1Pick a circuit

    Choose one from the menu, or type a name, maker or year in Find a circuit. The buttons under the menu switch between versions.

  2. 2Set the knobs and the input

    Turn the circuit’s knobs on the front panel. In the Input box, choose Guitar or Signal generator, and the note or chord.

  3. 3Press Play

    Press Play through the circuit to hear a picked note or chord. Play the string is the guitar by itself.

  4. 4Tap a part

    Tap any part on the schematic to change its value, or swap the transistor, tube or diode. The red numbers are DC voltages, and they update as you go.

  5. 5Tap a wire

    Tap a wire to put the probe there. The Probe graph and the spectrum show the signal there.

  6. 6Try a change

    Press Try it on a card under Try this to hear one change, before and after. Modify this circuit, below the schematic, adds or removes parts.

Good to know

Modeled after the Ludwig Phase II Synthesizer.

A talking wah from two resonant filters tuned like the first two formants of a voice, after the Ludwig Phase II Synthesizer. A rocker pedal moves them along one of three paths, and a square-wave fuzz and a slow oscillator can go in front.

Input

Add

Peak voltage from the pickup for each note. Strum hard and a single coil puts out a few hundred millivolts. When you play a chord, its strings share this level.

How it’s played
Tap a part to change its value, or a wire to probe it. To add or take out parts, open Modify this circuit, below. Red numbers are DC voltages to ground, not warnings. Scroll or drag sideways to see the whole circuit.
Modify this circuit Add parts, take them out, or put one in series with another, then hear what it does. Tap to open.

Pick what to do, then tap the schematic. Your changes show in purple. Any change can go on a switch beside the knobs, from the menu in its row under Your changes.

Add a part

Tap where each leg goes. Tap both ends of a part to put the new one across that part.

Change a part

With Remove a part on, tap any part of the pedal or amp to take it out, and tap it again to put it back. Tapping a purple part you added deletes it.

Turn on Add in series with a leg and tap a part. In that part’s box, below the schematic, pick which leg to lift and what goes in the gap, such as a resistor between a cap and ground or a stopper in front of a grid.

Change a wire

With Cut into a wire on, tap any wire. Pick a part from the menu and it goes in the gap, wired in. Volume pot puts a pot there with the input on one end, the output on the wiper and the other end to ground.

With Run a wire on, tap a leg, then where the wire goes. Drag any piece of a wire you ran, or of an added part’s lead, to square it up or move it.

Keep and share

Your changes stay in this browser, on this device, for your next visit. Nothing is sent to us. Anyone using this browser can see them. They’re lost if you clear your browsing data or use a private window, so save a file to keep a copy. A shared link carries the circuit after the #, a part browsers never send to a server.

Test tone for the graphs
The graphs use a steady tone. Play sounds a plucked string that dies away.

Input

pickup, unloadedat the pedal input

Output

at the amp input

Probe

Spectrum

Frequency response

What the curves mean

The ink curve is the response to a very quiet signal, from the generator to the probe. The blue curve is measured with pink noise at the level you set, like an audio analyzer. When stages clip, the blue curve pulls away from the ink one and the gray coherence line (right-hand scale) drops below 1. Then no single curve tells the whole story. Both use the signal generator, whichever Input tab is picked.

Output

Bias

every voltage and current at rest

What to listen for. QF3 sums the inputs and feeds a square-wave fuzz (QF5, QF6) and a clean amp (QF22). The Fuzz switch picks one for the filters. Each filter is a loop of two integrators made from discrete transistors. Its tuning comes from two differential pairs that steer how much of a big cap each integrator sees, so one control voltage moves both integrators together, from about 180 Hz to 5 kHz. Filter 2 has the smaller cap and sits higher. Their outputs are mixed with opposite signs, as in a speech synthesizer. The rocker makes three control voltages: the rocker itself, its inverse from Q5, and a V from D1 and D2 that is lowest in the middle. In Vowel, filter 1 follows the V and filter 2 the rocker, so heel to toe the formants trace ee, ah, oo. A UJT oscillator and a flip-flop can wobble the rocker’s voltage or pulse the fuzz. The supply is shown as a 35 V battery.

Download.

Parts and what they do

Stock values. Tap a part on the schematic to change it.

RF16, 6.8 kΩ
From the supply to the input stages’ 10 V rail, 6.8 k.
CF6, 10 µF electrolytic
Filter cap on the 10 V rail, 10 µF.
RF6, 1 MΩ
The Hi-Z input’s load, 1 M.
CF3, 10 nF
Hi-Z input coupling cap, .01 µF.
RF7, 680 kΩ
QF2’s bias from the 10 V rail, 680 k.
RF8, 680 kΩ
QF2’s bias to ground, 680 k.
RF9, 1 kΩ
Base stopper on QF2, 1 k.
QF2, 2N3904
The Hi-Z input’s emitter follower. No voltage gain.
RF10, 4.7 kΩ
QF2’s emitter resistor, 4.7 k.
CF4, 100 nF
From QF2 toward the mixer, 0.1 µF.
RF11, 10 kΩ
The Hi-Z input into the mixer, 10 k. Against QF3’s 100 k it sets the guitar’s gain at about 10.
RF4, 4.7 kΩ
The Lo-Z amp’s collector resistor, 4.7 k. The Lo-Z input has nothing plugged in here.
RF2, 33 kΩ
The Lo-Z amp’s collector to base, 33 k.
RF3, 10 kΩ
The Lo-Z amp’s base to ground, 10 k.
QF1, 2N3904
The Lo-Z input’s amp, for a microphone or a low-impedance pickup. Idle here.
CF2, 100 nF
From the Lo-Z amp toward the mixer, 0.1 µF.
RF5, 22 kΩ
The Lo-Z input into the mixer, 22 k.
RF14, 4.7 kΩ
The mixer’s collector resistor, 4.7 k.
RF12, 100 kΩ
The mixer’s feedback, 100 k. It holds QF3’s base at a low impedance, so each input’s gain is set by its own resistor.
RF13, 10 kΩ
QF3’s base to ground, 10 k.
QF3, 2N3904
The mixer, a one-transistor summing amp. Its output feeds the fuzz, the clean amp and the bypass.
RF20, 22 kΩ
From the mixer to the fuzz, 22 k.
CF10, 4.7 nF
Drives QF5’s base, 4.7 nF.
CF11, 4.7 nF
Drives QF6’s base, 4.7 nF. Two caps keep the 470 k bias between the bases.
RF27, 470 kΩ
Between the fuzz’s two bases, 470 k: the only DC path, so it biases both barely on.
QF5, 2N3906
Top half of the fuzz, a PNP. With QF6 it works like a CMOS inverter: so much gain that the output swings from one rail to the other.
QF6, 2N3904
Bottom half of the fuzz, an NPN.
QF21, 2N3904
The fuzz’s supply, an emitter follower. The square wave’s height is its emitter voltage, so this is the fuzz’s volume control.
CF12, 220 nF
Holds QF21’s base, 0.22 µF. Fuzz Mix charges it slowly, and the Percussion pulses kick it.
CF7, 100 nF
From the mixer to the clean amp and the bypass, 0.1 µF.
RF21, 22 kΩ
Into the bypass balance pot, 22 k.
VR2, 10 kΩ pot, linear taper
Bypass Balance, 10 k: the dry level in bypass.
CF8, 100 nF
Into QF22, 0.1 µF.
RF19, 4.7 kΩ
QF22’s supply resistor, 4.7 k.
RF23, 680 kΩ
QF22’s bias from its supply, 680 k.
RF24, 47 kΩ
QF22’s bias to ground, 47 k.
RF22, 4.7 kΩ
QF22’s collector resistor, 4.7 k.
QF22, 2N3904
The clean amp, gain about 10. It brings the clean guitar up to about the fuzz’s level.
RF25, 470 Ω
QF22’s emitter resistor, 470 Ω.
CF9, 100 nF
Out of the clean amp, 0.1 µF.
RF26, 47 kΩ
The clean amp’s output to ground, 47 k.
RF28, 3.3 kΩ
Top of the Lambda divider, 3.3 k.
RF29, 150 Ω pot, linear taper
The Lambda trim, 150 Ω. It sets the fixed bases of all four pairs, about 4.2 V, and with them where both filters sit.
RF30, 330 Ω
Foot of the Lambda divider, 330 Ω.
CF13, 10 µF electrolytic
Holds Lambda steady, 10 µF.
CF16, 10 nF
Coupling cap into filter 1, .01 µF.
RF36, 68 kΩ
Into filter 1’s first integrator, 68 k. It also damps the loop.
RF37, 4.3 kΩ
QF9’s emitter resistor, 4.3 k from the supply.
RF38, 4.7 kΩ
QF9’s base to emitter, 4.7 k. With the 22 k from the second integrator it sets how much of that voltage reaches the base.
RF39, 22 kΩ
From the second integrator back to QF9’s base, 22 k. This closes filter 1’s loop.
QF9, 2N3906
Turns the second integrator’s voltage into a current back into the first, which closes the loop.
RF32, 4.7 kΩ
QF7’s collector resistor, 4.7 k.
CF14, 2.5 µF electrolytic
Holds QF7’s collector at AC ground, 2.5 µF. The share of the big cap’s current QF7 carries goes here.
RF33, 150 kΩ
QF7’s base from its own collector, 150 k. The self-bias widens the control voltage’s window from a few tens of millivolts to a few volts.
RF35, 47 kΩ
QF7’s base to ground, 47 k.
CF15, 10 µF electrolytic
Smooths QF7’s base, 10 µF.
QF7, 2N3904
Control side of the first integrator’s pair. The more of the tail current it carries, the more of CF17 the integrator sees, so the lower filter 1 tunes.
QF8, 2N3904
Fixed side of the first integrator’s pair, its base on Lambda. Its collector is the integrator node, so its share of CF17’s current goes straight back and cancels.
RF40, 1 kΩ
Tail resistor of the first pair, 1 k.
CF17, 330 nF
The first integrator’s steered cap, 0.33 µF, to the pair’s emitters. The pair’s split sets how much of it the integrator sees.
CF18, 10 nF
The first integrator’s fixed cap, .01 µF. It sets filter 1’s top end.
QF10, 2N3904
Follows the first integrator. Its emitter is the band-pass output, and its 10 k emitter resistor turns that voltage into a current pulled from H.
RF43, 10 kΩ
QF10’s emitter resistor, 10 k.
RF42, 12 kΩ
H’s resistor from the supply, 12 k.
RF45, 47 kΩ
H’s resistor to ground, 47 k.
QF11, 2N3906
Turns H’s voltage into a current into the second integrator.
RF46, 4.7 kΩ
QF11’s emitter resistor from the supply, 4.7 k.
RF48, 4.7 kΩ
QF11’s emitter to the second pair’s control collector, 4.7 k, an AC ground. With the other 4.7 k it sets QF11’s gain.
CF33, 10 nF
The second integrator’s fixed cap, .01 µF.
CF20, 330 nF
The second integrator’s steered cap, 0.33 µF, to the second pair’s emitters.
QF12, 2N3904
Fixed side of the second integrator’s pair, its base on Lambda. Its collector is the second integrator node.
RF47, 1 kΩ
Tail resistor of the second pair, 1 k.
QF13, 2N3904
Control side of the second integrator’s pair. It moves with QF7, so both integrators tune together.
RF50, 150 kΩ
QF13’s base from its own collector, 150 k.
RF49, 47 kΩ
QF13’s base to ground, 47 k.
CF21, 10 µF electrolytic
Smooths QF13’s base, 10 µF.
CF22, 2.5 µF electrolytic
Holds QF13’s collector at AC ground, 2.5 µF.
RF34, 22 kΩ
The control voltage into QF7’s base, 22 k plus part of the balance trim.
RF51, 22 kΩ
The control voltage into QF13’s base, 22 k plus the rest of the balance trim.
RF77, 50 kΩ pot, linear taper
filter 1’s balance trim, 50 k. Its wiper takes the control voltage and splits it between the two pairs, so it sets how well the two integrators track, and with that filter 1’s Q across the sweep.
RF41, 500 Ω pot, linear taper
filter 1’s offset trim, 500 Ω from Lambda to the fixed bases. The base current through it shifts the filter’s tuning.
CF19, 100 nF
Out of filter 1, from QF10’s emitter, 0.1 µF.
RF44, 22 kΩ
Filter 1 into the mix point, 22 k.
CF25, 10 nF
Coupling cap into filter 2, .01 µF.
RF57, 68 kΩ
Into filter 2’s first integrator, 68 k. It also damps the loop.
RF58, 4.3 kΩ
QF16’s emitter resistor, 4.3 k from the supply.
RF60, 4.7 kΩ
QF16’s base to emitter, 4.7 k. With the 22 k from the second integrator it sets how much of that voltage reaches the base.
RF61, 22 kΩ
From the second integrator back to QF16’s base, 22 k. This closes filter 2’s loop.
QF16, 2N3906
Turns the second integrator’s voltage into a current back into the first, which closes the loop.
RF56, 4.7 kΩ
QF14’s collector resistor, 4.7 k.
CF23, 2.5 µF electrolytic
Holds QF14’s collector at AC ground, 2.5 µF. The share of the big cap’s current QF14 carries goes here.
RF53, 150 kΩ
QF14’s base from its own collector, 150 k. The self-bias widens the control voltage’s window from a few tens of millivolts to a few volts.
RF54, 47 kΩ
QF14’s base to ground, 47 k.
CF24, 10 µF electrolytic
Smooths QF14’s base, 10 µF.
QF14, 2N3904
Control side of the first integrator’s pair. The more of the tail current it carries, the more of CF26 the integrator sees, so the lower filter 2 tunes.
QF15, 2N3904
Fixed side of the first integrator’s pair, its base on Lambda. Its collector is the integrator node, so its share of CF26’s current goes straight back and cancels.
RF59, 1 kΩ
Tail resistor of the first pair, 1 k.
CF26, 150 nF
The first integrator’s steered cap, 0.15 µF, to the pair’s emitters. The pair’s split sets how much of it the integrator sees.
CF27, 10 nF
The first integrator’s fixed cap, .01 µF. It sets filter 2’s top end.
QF17, 2N3904
Follows the first integrator, and its 10 k emitter resistor turns that voltage into a current pulled from H. Its collector is this filter’s output.
RF64, 10 kΩ
QF17’s emitter resistor, 10 k.
CF29, 220 pF
Across QF17’s emitter resistor, 220 pF.
RF63, 12 kΩ
H’s resistor from the supply, 12 k.
RF66, 47 kΩ
H’s resistor to ground, 47 k.
QF18, 2N3906
Turns H’s voltage into a current into the second integrator.
RF67, 4.7 kΩ
QF18’s emitter resistor from the supply, 4.7 k.
RF69, 4.7 kΩ
QF18’s emitter to the second pair’s control collector, 4.7 k, an AC ground. With the other 4.7 k it sets QF18’s gain.
CF30, 150 nF
The second integrator’s steered cap, 0.15 µF, to the second pair’s emitters.
QF19, 2N3904
Fixed side of the second integrator’s pair, its base on Lambda. Its collector is the second integrator node.
RF68, 1 kΩ
Tail resistor of the second pair, 1 k.
QF20, 2N3904
Control side of the second integrator’s pair. It moves with QF14, so both integrators tune together.
RF70, 150 kΩ
QF20’s base from its own collector, 150 k.
RF71, 47 kΩ
QF20’s base to ground, 47 k.
CF31, 10 µF electrolytic
Smooths QF20’s base, 10 µF.
CF32, 2.5 µF electrolytic
Holds QF20’s collector at AC ground, 2.5 µF.
RF52, 22 kΩ
The control voltage into QF14’s base, 22 k plus part of the balance trim.
RF72, 22 kΩ
The control voltage into QF20’s base, 22 k plus the rest of the balance trim.
RF55, 50 kΩ pot, linear taper
filter 2’s balance trim, 50 k. Its wiper takes the control voltage and splits it between the two pairs, so it sets how well the two integrators track, and with that filter 2’s Q across the sweep.
RF62, 500 Ω pot, linear taper
filter 2’s offset trim, 500 Ω from Lambda to the fixed bases. The base current through it shifts the filter’s tuning.
CF28, 100 nF
Out of filter 2, from QF17’s collector (the inverted side), 0.1 µF. Its load on QF17 lags the loop slightly, so RF65 sets filter 2’s Q as much as its level.
RF65, 25 kΩ pot, linear taper
Filter 2 into the mix point, a 25 k trim. Filter 2 comes in with the opposite sign to filter 1, so the two add between the formants. Turned toward 0 Ω it raises filter 2’s level and lowers its Q.
RF73, 4.7 kΩ
The mix point’s load, 4.7 k. With the 22 k and 25 k into it, this pad drops the level a lot.
QFA, 2N3904
The make-up amp, added on a terminal strip after the boards were made (Keen: QF+1). A shunt-feedback stage that brings the mix point back up, about 20 times.
RFA1, 10 kΩ
The make-up amp’s collector resistor, 10 k.
RFA3, 68 kΩ
The make-up amp’s collector to base, 68 k.
CFA1, 100 nF
Out of the make-up amp, 0.1 µF.
RFA2, 47 kΩ
The effect output to ground, 47 k.
R31, 25 kΩ pot, linear taper
The rocker’s range trim, 25 k from the supply to the rocker’s top. It sets how high the control voltage goes.
PED, 10 kΩ pot, linear taper
The rocker pedal, 10 k.
RTAP, 33 kΩ
The tapering resistor, 33 k across the rocker’s top half. It bends the rocker’s curve.
R75, 10 kΩ
From the rocker’s wiper to the control node, 10 k.
C6, 1 µF electrolytic
Smooths the control node, 1 µF.
R14, 6.8 kΩ
The rocker voltage toward filter 2, 6.8 k.
R13, 100 kΩ
Filter 2’s rocker voltage to ground, 100 k.
D1, 1N914 / 1N4148
With D2, picks the higher of the rocker voltage and its inverse: high at both ends of the rocker, lowest in the middle, a V. That turns the first formant round in the middle of the sweep.
D2, 1N914 / 1N4148
The inverse rocker’s half of the V.
R16, 100 kΩ
Into the inverter Q5, 100 k.
R18, 120 kΩ
Q5’s feedback, 120 k.
R19, 6.8 kΩ
Q5’s base toward the centre trim, 6.8 k.
R20, 10 kΩ pot, linear taper
The inverter’s centre trim, 10 k. It sets where the inverse crosses the rocker, so where the V bottoms out.
Q5, 2N3904
The rocker’s inverter. Its collector falls as the rocker rises.
R21, 15 kΩ
Q5’s collector resistor, 15 k.
R17, 8.2 kΩ
The inverse rocker voltage out, 8.2 k.
R25, 4.7 kΩ
Filter 1’s control, 4.7 k.
R24, 47 kΩ
Filter 1’s control to ground, 47 k.
C9, 10 µF electrolytic
Smooths filter 1’s control, 10 µF.
R27, 4.7 kΩ
Filter 2’s control, 4.7 k.
C10, 10 µF electrolytic
Smooths filter 2’s control, 10 µF.
R15, 10 kΩ
The animation into the control node, 10 k.
VR6, 10 kΩ pot, linear taper
FFM Amount, 10 k: how far the animation moves the filters.
R30, 2.2 kΩ
Foot of FFM Amount, 2.2 k.
C12, 50 µF electrolytic
From flip-flop side A to FFM Amount, 50 µF.
R2, 100 kΩ
Soft start, 100 k from the supply. The Animation switch grounds its far end to stop the oscillator.
R3, 1 MΩ
Soft start, 1 M into C1.
C1, 2.5 µF electrolytic
The soft start cap, 2.5 µF. With R3 it ramps the oscillator’s supply over a few seconds, so the animation swells in and dies away.
Q1, 2N3904
The soft start’s emitter follower, the oscillator’s supply.
R4, 500 kΩ pot, linear taper
The Rate range trim, 500 k. Set at 0.9 (assumed).
VR5, 500 kΩ pot, linear taper
Rate, 500 k: the UJT’s charging resistance.
R5, 10 kΩ
Into the timing cap, 10 k.
C2, 1 µF electrolytic
The timing cap, 1 µF. It ramps up until the UJT fires and empties it.
R6, 10 kΩ
The UJT’s base 2 resistor, 10 k.
R7, 220 Ω
The UJT’s base 1 resistor, 220 Ω.
Q2, 2N2646
The 2N2646 unijunction transistor. When C2 reaches its peak point, about 10 V, it fires, empties C2 through base 1 and resets. Each firing pulls base 2 down several volts.
C3, 10 nF
Couples the UJT’s base 2 pulse onto the flip-flop’s load line, .01 µF.
R10, 4.7 kΩ
Feeds the flip-flop’s load line, 4.7 k.
R11, 4.7 kΩ
Q3’s collector resistor, 4.7 k from the load line.
R40, 4.7 kΩ
Q4’s collector resistor, 4.7 k from the load line.
Q3, 2N3904
Flip-flop side A. Each UJT pulse pulls down whichever collector is high, which turns the other transistor off, so it toggles once a firing.
Q4, 2N3904
Flip-flop side B.
R8, 47 kΩ
Cross-coupling, Q3’s collector to Q4’s base, 47 k.
C4, 4.7 nF
Speed-up cap across R8, 4.7 nF.
R12, 47 kΩ
Cross-coupling, Q4’s collector to Q3’s base, 47 k.
C5, 4.7 nF
Speed-up cap across R12, 4.7 nF.
C13, 220 nF
From flip-flop side B to Percussion, 0.22 µF.
R31B, 4.7 kΩ
Into Percussion, 4.7 k.
VR3, 10 kΩ pot, linear taper
Percussion, 10 k: how hard each flip-flop edge kicks the fuzz level.
R29, 2.2 kΩ
Foot of Percussion, 2.2 k.
D6, 1N914 / 1N4148
Lets each flip-flop edge charge CF12 quickly. The charge then bleeds away, so each edge is a burst of fuzz that dies out.
R1, 47 kΩ
Top of Fuzz Mix, 47 k from the supply.
VR4, 10 kΩ pot, linear taper
Fuzz Mix, 10 k: the fuzz’s level.
R28, 2.2 kΩ
Foot of Fuzz Mix, 2.2 k.
R32, 470 kΩ
Fuzz Mix into CF12, 470 k: a glide of about 0.1 s.

About this version

Keen’s redraw
R.G. Keen’s redraw of the Ludwig Phase II, checked against the factory sheets. 2N3904 and 2N3906 in place of Ludwig’s own part numbers. Trims set toward a voice’s vowels.

Sources

How this tool works. Every part is an equation, and the whole circuit is solved several times per audio sample: 192,000 samples a second while you drag a knob, 384,000 once you let go. Transistors use Ebers-Moll with gain, leakage, Early effect, base resistance, junction capacitance and temperature. Their charge storage is set at the bias point, close for small signals and rough while they switch hard. Tubes use Koren’s equations fitted to each datasheet, plus grid current. Diodes are exponential junctions with series resistance. Presets are typical of each type. Caps are ideal, and the battery or HT is a voltage behind a resistance.

Spot an error, or want a circuit added? Email info@swellamps.com. A schematic or a clear photo of one helps. Anything you send may be used to improve these models and our desktop app.

© 2026 Swell Amplification LLC. All rights reserved. The simulator, its circuit models, drawings, presets and text are the property of Swell Amplification LLC and may not be copied, reproduced, redistributed or used to build another product without written permission. The original circuits are historical designs documented in public sources, and no claim is made to them. Everything here is provided as is, without warranty of any kind, for study and personal use. Results are simulations of typical parts, not measurements of any real unit. Don’t rely on them for the safety of any build or repair. Tube amplifiers carry lethal voltages.

Product and company names are trademarks of their owners. They are used here only to identify the circuits these models are based on; this site is not affiliated with or endorsed by any of them.

From Swell Amplification, swellamps.com.