Noise figure is an important parameter in many RF applications. Per its definition, Noise Figure (NF) is a measurement of the degradation of the signal-to-noise ratio (SNR) which is caused by components in a signal path. It is a number by which the performance of an amplifier or a radio receiver can be specified. The lower the values, the higher the performance. As an example, a DUT (Device Under Test) with a Noise Figure of 3 dB will degrade the SNR of a signal by 3 dB when the system is kept at 290 K.
There are several methods of performing noise figure measurement which vary for different applications. Some applications have high gain and low noise figure, some have low gain and high noise figure (mixers), some have very high gain and a wide range of noise figure (receiver systems). Measurement methods must be chosen carefully. This blog will focus on the noise source/spectrum analyzer based “Y-factor method.” Often it may not be feasible to find a proper noise figure meter and, in these circumstances, it is important to know how to measure noise figure with a widely available spectrum analyzer.
The Y-factor method is a popular way to measure noise figure and is often accepted as being equally as accurate as that of a specialized noise figure meter. It requires a noise source with accurate ENR (Excess Noise Ratio) data and a spectrum analyzer as a high gain sensitive receiver. The setup is shown in Figure 1:
If the DUT’s operating frequency is beyond the spectrum analyzer’s upper limit, a mixer and a LO reference source can be inserted after the DUT to bring down the frequency, as shown in Figure 2.
The noise sources offered by SAGE Millimeter require a DC voltage supply from +18 to +28 VDC and cover full waveguide band. Each noise source has its own personalities, i.e., dedicated ENR table. Below is the example ENR table showing very different ENR for two different models.
Table 1. Example of ENR of Noise Heads
|Frequency (GHz)||NF (dB)||NF (dB)|
By turning the noise source “on” and “off” (by turning the DC voltage on and off), the output, which is the noise power density difference (Y-factor) of the DUT, can be measured with a spectrum analyzer. The formula to calculate noise figure is:
If the ENR of the noise source is known, then the Noise Figure can be obtained.
The recommended measurement procedure using the Y-factor method is shown here:
- Power on the DC power supply and spectrum analyzer. On the power supply, set the voltage and current limits for the DUT (for this example, an amplifier) and for the noise source (+28 VDC for SAGE Millimeter standard STZ Noise Source Family)
- On the spectrum analyzer, choose a central frequency at which to perform the measurements. Set the central frequency to a value within both frequency ranges of the noise source and the amplifier. For higher frequency, use measurement setup II.
- Set the proper averaging on the spectrum analyzer.
Note: The average button is essential for noise measurements with a spectrum analyzer. Without averaging, the power reading changes frequently and doesn’t represent the precise data. The more points used in the average, the greater the accuracy. However, more points also require more time. The user should find a balance between reasonable accuracy and time. Do not change the resolution bandwidth (RBW) or span between the first and second measurements. This will cause inaccurate results.
- Turn on the amplifier’s DC biasing while leaving the noise source off – as a “cold” state. The spectrum analyzer’s reading should increase due to the added amplifier noise floor.
- Perform an average of the data and record the value once all points are collected.
- Turn on the DC supply for the noise source – as a “hot” state. The spectrum analyzer’s reading shall go high.
- Perform the second average and record the number. Calculate the difference in dB between the first and second measurements. This is the Y-factor.
- Look up the ENR value for the frequency that the Y-factor measurement was performed at. Input the ENR and Y-Factor values into the above equation to solve for noise figure.