How does it work? The transistors are germanium. The distortion comes from keeping the base (input) of the transistor close to the ground voltage. This slices off some of the bottom half of the guitar signal. The other half of the guitar signal gets clipped by the next transistor. If the base voltage is too close to the ground the transistor stops amplifying until the guitar signal pushes it higher. So the quiet parts of the guitar sound disappear. This acts as a noise gate, but kills the sustain.

Q1with R3 is an input buffer. If Q2 base was connected to the input via C2 without Q1 and R3, you would have your pickup going to ground via the 0.01 input cap if the Fuzz pot is on zero. So that would act like a tone control on zero, since the input is connected to the guitar even with the effect off. Killing your guitar tone completely with the effect off wouldn't have been acceptable. Q1 isn't very high impedance, so it would affect you guitar tone to some extent.

Q2 has the base attached to a potential divider R6/R4-R5-Fuzz pot. Q2 has internal leakage which affects this too. So the bias voltage on the base of Q1 changes as you turn the Fuzz knob. With the fuzz pot at 50k, the voltage on the Q1 base might be about 0.1V, turn it the other way and the voltage will be about 0.01V. A 0.01V base, Q2 will be turned off, and won't amplify. When you hit the strings hard, the voltage from you pickups should turn the transistor on enough to amplify. But a 100mV guitar signal might only push the Q2's base up to a few tens of mV, so the bottom half of you guitar signal gets cut off, or clipped. This gives you a very bad noise gate which cuts the sustain of your guitar down. It also give you severe clipping. Germanium doesn't clip quite so harshly as Silicon here, but it is still brutal. Turning the Fuzz towards 50k will give you a cleaner sound with less noise gating, as Q2's base will be biased to allow the full signal to be amplified. But individual transistors will bias differently in this circuit, so some will be more gated than others. So if you build this with the transistors you have, you may have to fiddle with resistor values to get a sound you like. You would need an original unit to comare it to, if you wanted a clone.

Q3 has a 10k resistor pulling its bias down a bit. Leakage current (due to Germanium not being ideal) acts something like an invisible resistor going from base to power -1.5V. So, the bias on Q3 is going to cause clipping and gating too, but it depends on the individual transistor. Some transistors simply won't work in this circuit. Q2 inverted the signal, so Q3 clips on the other side of the waveform. Tube preamps do the inverted clipping thing too, but it isn't so harsh.

So then on to the volume control and there you have it, your guitar signal severely clipped, with some noise gating. With the Fuzz pot at 0k, you get more severe clipping on Q2, and probably less clipping on Q3 because the signal going into it is smaller. The clipping will be different on each unit, because of the variability of germanium transistors.

So this isn't what someone would design today, because you would need to hand tune each unit because of the variability of the germanium transistors. Plus the severe gated distortion is not to everyone's taste.

Hurst upped the battery voltage to 9v. This would increase the voltage on the transistors and give you more sustain immediately. He may have hand tuned resistor values to get the bias and sustain just right. The alleged Tonebender MkI schematic comes next.