MXR Phase 100 - a BJT as a variable R?

[cjlectronics]

I am working on a Phase 100 and have a question or maybe I should request clarification of the variable BJT portion of the circuit.

Looking at the schematic, in the lower left corner is Q1 and Q2 in a Darlington arrangement (I will call it this for a lack of knowing what this transistor arrangement really is) followed by a 20 ohm resistor. My question is, the LFO varies the base current of Q2 which in turn varies how hard Q1 is turned on...I think?
Why use two transistors, why not use just one?
Is it safe to say that Q1 is acting like a variable resistor?
What is the purpose of the 20 ohm resistor when Q1 varies the current?
Could you use a JFET in place of Q1 and Q2?

BTW, a big thank you goes out to JD Sleep for making his version of the Phase 100 schematic available to the world.

Thanks in advance
CJL

[Mark Hammer]

Thanks for that post CJ! It never really occurred to me before that Q1 and R54 simply provided variable current limiting on what is drawn from Vcc through the 3 LEDs in the Vactrol. One generally thiks n terms of a fixed current limiting resistance, and the voltage fed to it is varied by an LFO. Here, the LFO appears to vary the amount of current pulled through the LEDs, thus varying their brightness.

My guess is that R54 is there just to keep these pricey little Vactrols from completely frying.

[R.G.]

My question is, the LFO varies the base current of Q2 which in turn varies how hard Q1 is turned on...I think?
Why use two transistors, why not use just one?
Is it safe to say that Q1 is acting like a variable resistor?
What is the purpose of the 20 ohm resistor when Q1 varies the current?
Could you use a JFET in place of Q1 and Q2?

What's going on is that the darlington is enough current gain to make Q1/Q2/R54 be a voltage to current converter, kind of.

Q1 is not really a variable resistor, it's a current source driven eventually by the triangle wave-ish across C15, as buffered by Q4 and modified in position by the current drawn by Q3 to load it down and change its DC position.

It's a darlington to have enough gain so the 20 ohm can be driven by the available current from Q4.

Any time you see a transistor with a resistor from emitter to ground being driven by a control signal, one of the things you should consider is that the transistor/resistor is acting like a voltage controlled current source. I think that's what's happening here.

I like VCCSs a lot. One trick you can play is to not drive the base of a transistor with a current, but a current-limited voltage. The current in the collector of a transistor is an exponential function of the base voltage, so if what drives the base is a linear ramp, the collector current is an exponential ramp, to a high degree of accuracy. You have to hand the special case of the base input impedance being quite low, so the drive voltage must be lower still, and must not go higher than the base voltage at the max collector current, about 0.8V. But in many cases, this can be as simple as a single-supply opamp (324) driven by a voltage divider on its + input. Can't use an emitter resistor with this, it linearizes the base-collector relation too much.