[Image1]
Hey it's a me again @drifter1!
Today we continue with my mathematics series about Signals and Systems to talk about the Properties of the Z Transform.
So, without further ado, let's dive straight into it!
The Z Transform is a generalization of the Discrete-Time Fourier Transform. Basically, the complex exponentials that are used as building blocks for signals are replaced by more general exponentials with complex base. This allows the Z Transform to converge for signals/sequences that normally don't converge using the Fourier Transform, and can be thought of as some kind of exponential weighting.
The Fourier and Z Transform are related in the following manner:
Similar to all other Transforms that have been covered throughout this series, the Z Transform also satisfies many properties. These properties are basically the same as the Discrete-Time Fourier Transform (or even the Laplace Transform), but have a few differences in the math. Of course, similar to the Laplace Transform, all these properties also affect the ROC, and so the complete definition of the outcome, algebraic equation and ROC, will be the main topic of this article.
Consider the following two Z Transforms of Discrete-Time signals:
The linear combination of those signals leads to the following Z Transform:
So, the ROC of the result is equal to the intersection of the two individual ROCs. Let's note that the ROC might also additional regions, and so this ROC is basically the least for which the Z Transform converges. More specifically, if the linear combination leads to additional zeros or poles, these will of course have to be included in the ROC.
Let's now consider the following Z Transform:
The time-shifted (sample-shifted) signal x[n - n0] has the following Z Transform:
The ROC is basically the same as the original signal, but additional regions might be included at z = 0 or infinity. This depends on how the scaling by z -n0 affects the pole-zero pattern.
In Discrete-Time, because samples are only taken at integer indexes, time-scaling leads to the introduction of zeros between the samples. As such, scaling by 1 / k introduces k - 1 zeros between the individual samples of the original signal. Mathematically:
Now for the Z Transform. The algebraic expression and ROC are:
So, if z is a point in the original ROC then z 1 / k is a part of the resulting ROC. The same is also true for poles and zeros. As such, if z = a is a pole (zero) then z = a 1 / k will also be a pole (zero) in the resulting pole-zero pattern.
Reversing a signal in time, also leads to an interesting result. The Z Transform of the signal x[-n], is as follows:
So, if z0 is a part of the original ROC, then 1 / z0 will be part of the resulting ROC.
The equivalent of frequency shifting in the case of the Z Transform is known as z-Domain Scaling. Of course frequency scaling in the Time-Domain leads to scaling in the z-Domain.
So, Multiplying a given signal x[n] by z0n yields the following result:
So, if the original ROC contained z, then the resulting ROC will contain |z0| z. The same also applies for the poles and zeros.
A special case is z0 = e jω0, which leads to a rotation of the ROC by an angle of ω0. In other words, the pole-zero pattern is rotated by an angle of ω0. Mathematically:
And if we get even more generic with a z0 = r0 e jω0, then in addition to being rotated by ω0, the poles and zeros are also scaled in value by r0.
The Z Transform also satisfies conjugation. As such, the Z Transform of the conjugate x*[n] is:
If the signal x[n] is real, then its possible to conclude that:
Which means that if the original ROC has a pole (or zero) at z = z0, then there will also be a pole (or zero) at the conjugate point z = z0*.
Next up is convolution, which is a very useful property. This property allows us to turn the complicated convolution sum into a simple multiplication in the z-Domain.
Mathematically, the Z Transform of the convolution of two signals is as follows:
So, the ROC is equal to the intersection of the two individual ROCs, with the possibility of including additional poles-zeros depending on the result of the multiplication.
Differentation in the z-Domain is another quite useful property, which is basically the multiply-by-n property.
Algebraically, the resulting Z Transform is:
which is a direct consequences of differentiating both sides of the z Transform equation.
The Z Transform also satisfies similar initial and final value theorems to the Laplace Transform.
The initial value theorem states that for any causal signal x[n]:
As such, this theorem is useful for finding the initial value of a signal without taking the inverse Z Transform.
Whilst, the final value theorem states that for any causal signal x[n]:
Similarly, this theorem is useful for finding the final value of a signal withotu taking the inverse Z Transform.
Mathematical equations used in this article were made using quicklatex.
Block diagrams and other visualizations were made using draw.io and GeoGebra
And this is actually it for today's post!
Next time we will get into how the Z Transform is useful in LTI System Analysis...
See Ya!
Keep on drifting!
