discrete all-pass filter design

Thread Starter

jorngensen

Joined Nov 25, 2019
2
out of curiosity, does anyone happen to know of a discrete (single transistor?) all-pass filter design? i am curious how this might be done. thanks.
 

crutschow

Joined Mar 14, 2008
34,281
An all-pass filter changes the phase of a signal with frequency, without changing the amplitude.
It can be readily done with an opamp, but I don't offhand know how to do it with a single transistor.
 

KW_KW

Joined Dec 18, 2019
9
One simple all-pass filter was used in electronic music for decades. A single transistor, equal emitter and collector resistors, base biased at about 1/4 VCC (for maximum voltage swing at emitter and collector). Connect a capacitor to the emitter. Connect a resistor to the collector. These are your phase shift components, and the RC value will be the 90 degree phase shift point. The resistor should be a good order of magnitude greater than the collector resistor. Connect the other terminal of the resistor and the capacitor to a high-impedance buffer (which can be and often was another of the same stage). The phase varies from 180 degrees at low frequencies to 0 at high frequencies, with nominally unity gain. Typically R3 was a photoresistor, to make the phase shift variable. If you flip the R and C, and have a high enough supply voltage, you can run a handful of them in series and leave the biasing out for the later stages.transistor all-pass.JPG

I (we) have no idea whether it provides the DESIRED phase response, it's just AN all-pass filter. You can make fancier ones, with feedback and sharper phase transitions.
 

crutschow

Joined Mar 14, 2008
34,281
Below is the LTspice simulation of KW_KW's circuit.
As expected, the gain is essentially constant with frequency while the phase goes from 180° (inverted) at low frequencies to 360° (non-inverted) at high frequencies.


1576730767452.png
 
Top