Maker Pro
Maker Pro

How does shifting phase across frequencies work on the physical level

00135599

Aug 23, 2014
4
Joined
Aug 23, 2014
Messages
4
How does shifting phase across frequencies work on the physical level? When filter or probe has this frequency charakteristic, where in one graph there is a gain for every frequency and in the second one there is phase for every frequency. What makes different frequencies to be transmitted with different gain and phase?
 

Harald Kapp

Moderator
Moderator
Nov 17, 2011
13,700
Joined
Nov 17, 2011
Messages
13,700
The elemental factor is the frequency depended characteristic of reactive components (inductors, capacitors).

You may start by reading and understanding this tutorial.
 

Merlin3189

Aug 4, 2011
250
Joined
Aug 4, 2011
Messages
250
One way of getting an idea about this is to think of an R-C circuit. If you apply a signal to a resistor in series with a capacitor and measure the Voltage across the capacitor, it obviously takes some time for the capacitor to charge up to the applied Voltage. (The actual time may be very small if R and C are small.)
Now if the applied signal is very low frequency, the output will be just this small time behind the input, which will be a small difference in phase.
As the signal gets to higher frequencies, this small delay becomes a bigger and bigger fraction of one cycle, which is a bigger phase difference. Say at 3.3Hz one cycle takes 300mSec and a delay of 1mSec is about 1 degree of phase, but at 33 Hz one cycle takes about 30 mSec so a delay of 1mSec is about 12 degrees of phase. (Phase shift in one RC stage never gets above 90 degrees, so this is oversimplifying a bit, or a lot!)
Also, because the output is delayed, the input changes from increasing to decreasing before the output reaches the full input level. And similarly starts increasing again before the output has reached the minimum. So the output level is also "trailing" the input in amplitude: it is attenuated in amplitude. Again the effect gets worse as the frequency increases, because the signal changes direction sooner.
That is a very crude qualititive way of looking at it and does not predict the correct values.
You can see one of characteristics of these graphs, that, for an RC section, as long as the frequency is below a certain value (for any given R & C) the attenuation is negligible and phase shift is small, then above that, when the delay is becoming a significant part of one cycle and the phase shift is becoming bigger, then the attenuation is increasing at 6 dB per octave (ie. double the frequency and you double the attenuation.)

Reversing the RC you can get the high pass configuration, where high frequency signals don't have time to put any significant charge onto the capacitor, so appear as full unattenuated output across the resistor with no phase delay. But as you reduce the frequency and the capacitor has time to charge up, some of the signal is "lost" in doing that and it causes again an increasing phase delay.

The inverted bathtub frequency response you see for many amplifiers represents the summation of parallel capacitances attenuating the high frequencies and the serial capacitances attenuating the low frequencies, with a passband in between.

You can do a proper mathematical analysis of RC circuits (and RL or even LC circuits) and when you solve the differential equations, the formulae and graphs appear just as in the textbooks and, to be fair, much as we see them in real circuits. And knowing that it can be done, I happily use the formulae to calculate circuit values. But I like to have this rough intuitive notion of what's going on when I look at a circuit and think about what will happen if something is changed.
 
Top