DrQuack - how does the transistor bit work?

[Paul Perry (Frostwave)]

Looking at the DrQ and derivatives, I can see (thanks RG!) that the transistor forms a variable resistor in the twin T filter section.
But, the transistor seems to have 0v on the collector and emitter (I havn't measured it, but looking at the schematic for the DrQ and derivatives, I dont see the collector connecting to anything but ground).
Now I can't see the transistor doing any amplifying under these circumstances, so could it just be replaced by a pair of diodes?

And this is a completely unrelated question, but in any circuit where you want to amplify & rectify a signal, could you deliberately misbias a transistor or op amp so that only one half cycle of the signal going in is amplified?

[amz-fx]

The transistor is essentially acting as a current controlled resistor...

Can you do something similar with diodes? Sure...  look at this Paia bandpass filter schematic from my site (posted with Paia's permission, I might add):

http://www.muzique.com/schem/2720-3b.gif

could you deliberately misbias a transistor or op amp so that only one half cycle of the signal going in is amplified

Yes you can though the other half of the signal will be clipped off.  A distortion could be made by putting a pair of these in sereis...  clip one-half of the waveform and then the inverted output goes into a second identical circuit where the other side is clipped.

regards, Jack

[R.G.]

The transistor is essentially acting as a current controlled resistor...

It's not widely known but bipolar transistors act like variable resistors if they are fed a tiny signal centered on a zero volt collector-emitter bias. It is critical that DC be kept off the collector and emitter for this to work well, and that the signal be small. The linear variable resistance region of a bipolar is much smaller than it is for JFETs, and is not well suited to linearization like JFETs are, so they don't get much play. They are also subject to an offset voltage caused by the current fed from the base to modify the collector/emitter conductance. The signal needs to be quite small to make this work well.

Now I can't see the transistor doing any amplifying under these circumstances, so could it just be replaced by a pair of diodes?
...
Can you do something similar with diodes? Sure...

Diodes have a nonlinear curve of current versus voltage. That is, if you feed a diode a DC bias current and a smallish signal current at the same time, the smallish signal acts as though the diode is a resistor of different values for each different DC bias current.  As long as the signal current is negligible compared to the DC bias current, it's like a current controlled resistor.

This diode current controlled resistor is quite different from the bipolar transistor one. The transistor is like a resistor controlled from a separate terminal, the base; it is firmly tied to ground, and operates for small signals symmetrically around ground. The diode must be fed from a current source (or other quite high impedance) bias source, and you have to couple the signal onto the diode's voltage.

The diode's voltage varies with bias current, so you have to arrange to get signal onto and off it without absorbing the DC bias current-caused voltage into your signal. The PAIA circuits did this; a slightly more sophisicated scheme was used in the Thomas Vox amps for a tremolo. A diode bridge was fed with a variable DC at the + and - points, and signal was fed into it at the AC signal points.

Given a nicely balanced bias, the signal flowed through two diode-resistances in series, and the bias cancelled itself out.  This is one version of a diode modulator that had wide use in the 50's and 60's.

So - the mechanisms of changing the currents are fundamentally different, as are the circuits that are there to take advantage of them.

As a final note, the bipolar transistor CCR will produce acceptable results with slightly bigger signal voltages than the diode version. The diodes really need to be limited to 25mv peak of signal, and the transistor one can get up to maybe 100mv before the distortion gets too bad.

As a final comment, some bipolar transistors work better as both switches and CCRs if you use them in inverted mode - that is, swap the collector and emitter. Depends on the device.

[Paul Perry (Frostwave)]

THis has to be the best of times.. type in a FX electronics question before going to bed at 3AM, wake up at 10AM & there are the answers, from Jack & RG!! thanks!