Group Delay Format |
Group delay is a measure of phase distortion. Group delay is the actual transit time of a signal through a device under test as a function of frequency. When specifying group delay, it is important to specify the aperture used for the measurement.
Group delay is:
A measure of device phase distortion.
The transit time of a signal through a device versus frequency.
The derivative of the device's phase characteristic with respect to frequency.
Refer to the graphic below for the following discussion:
The phase characteristic of a device typically consists of both linear and higher order (deviations from linear) phase-shift components.
|
Linear phase-shift component: |
Higher-order phase-shift component: |
|
Represents average signal transit time. |
Represents variations in transit time for different frequencies. |
|
Attributed to electrical length of test device. |
Source of signal distortion. |
Refer to the graphic below for the following discussion:
In a group delay measurement:
The linear phase shift component is converted to a constant value (representing the average delay).
The higher order phase shift component is transformed into deviations from constant group delay (or group delay ripple).
The deviations in group delay cause signal distortion, just as deviations from linear phase cause distortion.
The measurement trace depicts the amount of time it takes for each frequency to travel through the device under test.
Refer to the following equation on how group delay is calculated:
Phase data is used to find the phase change (-df).
A specified frequency aperture is used to find the frequency change (dw).
Using the two values above, an approximation is calculated for the rate of change of phase with frequency.
This approximation represents group delay in seconds (assuming linear phase change over the specified frequency aperture).
During a group delay measurement, the analyzer measures the phase at two closely spaced frequencies and then computes the phase slope. The frequency interval (frequency delta) between the two phase measurement points is called the aperture. Changing the aperture can result in different values of group delay. The computed slope ( -delta phase / delta frequency) varies as the aperture is increased. This is why when you are comparing group delay data, you must know the aperture that was used to make the measurements.
Refer to the graphic below for the following discussion:
|
Narrow aperture: |
Wide aperture: |
|
Provides more detail in phase linearity. |
Provides less detail in phase linearity because some phase response averaged-out or not measured. |
|
Makes measurement susceptible to noise (smaller signal-to-noise ratio) and analyzer phase detector resolution. |
Makes measurement less susceptible to noise (larger signal-to-noise ratio). |
It is important to keep the phase difference between two adjacent measurement points less than 180° (see the following graphic). Otherwise, incorrect phase and delay information may result. Undersampling may occur when measuring devices with long electrical length. You can verify that the phase difference measured between two adjacent points is less than 180° by adjusting the following settings until the measurement trace no longer changes:
Increase the number of points (resolution)
Narrow the frequency span
The frequency response is the dominant error in a group delay test setup. Performing a calibration significantly reduces this error. For greater accuracy, perform a 2-port measurement calibration.
Particularly for an amplifier, the response may vary differently at various temperatures. The tests should be done when the amplifier is at the desired operating temperature.
Preset the analyzer.
If your DUT is an amplifier, it may be necessary to set source power to Low. If needed, use an external attenuator so the amplifier output power will be sufficiently attenuated to avoid causing receiver compression or damage to the analyzer port 2.
Connect the DUT.
Select an S21 measurement.
Select the settings for your DUT:
frequency range
number of measurement points.
format: delay
scale: autoscale
Remove the DUT and perform a measurement calibration.
Reconnect the DUT.
Scale the displayed measurement for optimum viewing.
Use the markers to measure group delay (expressed in seconds) at a particular frequency of interest.