Method for Measuring the Noise Figure of a Device Under Test with a Network Analyser

Abstract

A method for measuring the noise factor (FDUT) of a device under test, which requires exclusively a network analyzer. The noise factor (FDUT) is calculated from the internal noise (NNWA) of the network analyzer determined in a calibration process, the power amplification (GDUT) of the device under test determined by measuring the S-parameters of the device under test, and the measured value (PNOISE) of the noise output (NNWA) applied at a first gate of the device under test without exciting the device under test with a noise signal.

Description

The invention relates to a method for measuring the noise figure of a device under test with a network analyzer.

The most important electrical parameters of a device under test (DUT) include the S-parameters and the noise factor. The noise factor is generally measured with a noise source and a spectrum analyzer. A network analyzer is generally used to measure the S-parameters.

A method and a device for determining the noise figure of electronic devices under test is known, for example, from DE 103 02 362 A1.

As will be shown in detail below, the power amplification of the device under test, which is dependent upon the S-parameters, is also required for a measurement of the noise figure of a device under test. A network analyzer, for determining the S-parameters for the power amplification of the device under test and a spectrum analyzer, for measuring the output noise power of the device under test, which is additionally required in the determination of the noise figure, are therefore disadvantageously both required for the measurement of the noise figure of a device under test.

The object of the invention is therefore to develop a method for measuring the noise figure of a device under test, which requires a reduced number of measuring instruments.

For this purpose, the invention uses only a network analyzer, which, according to the prior art, provides an input sensitivity corresponding to the input sensitivity of a spectrum analyzer and which is therefore suitable for a measurement of the noise power. A network analyzer of this kind can be used to measure the S-parameters of the device under test, required for the determination of the associated power amplification, and also to measure the output noise power at one of the two ports of the device under test, which are connected respectively to a port of the network analyzer in each case without exciting the device under test with a noise signal generated by a noise source and without substantially changing the measurement structure.

The input impedance of the measuring port of the network analyzer is generally matched with a 50-Ω resistor. In the case of an error matching of the input impedance of the measuring port of the network analyzer, the measured output noise power must be corrected with an error-matching factor.

An inaccuracy between the correct measured value of the output noise power applied at the port to be measured of the device under test and the actually-measured, incorrect measured value of the output noise power applied at the port to be measured of the device under test and the internal noise of the network analyzer must be determined within the framework of a calibration process by exciting the measuring port to be measured of the network analyzer with the noise from a noise source, which is coupled to the network analyzer and switched on and then switched off again.

An embodiment of the method according to the invention for measuring the noise factor of the device under test with a network analyzer is explained in greater detail below with reference to the drawings. The drawings are as follows:

FIG. 1 shows a flow chart of the method according to the invention for measuring the noise figure of a device under test with a network analyzer;

FIG. 2 shows a block-circuit diagram of an arrangement for the implementation of the calibration process of the network analyzer; and

FIG. 3 shows a block-circuit diagram of an arrangement for the implementation of the measuring procedure of the noise-figure measurement with a network analyzer.

Before the method according to the invention for measuring the noise figure of a device under test with a network analyzer (NWA) is explained in greater detail below with reference to FIGS. 1 to 3, the mathematical context necessary for an understanding of the method according to the invention will be described in the following paragraphs.

The noise figure F_DUT of a device under test is defined as the ratio of the total output noise power N_OUT at the output of the device under test relative to the output noise power N_OUT at the output of the device under test, which results exclusively from the noise power applied at the input of the device under test. The total output noise power N_OUT at the output of the device under test is obtained according to equation (1) from the “noise-power sum”, which is derived from the noise power N_OUT generated by the device under test itself and the thermal noise N₀ of a 50-Ω resistor system always present in the case of a matched measuring circuit, amplified with the power amplification G_DUT of the device under test, at a room temperature of 290 K.

N _OUT=(N_DUT+N₀)·G_DUT  (1)

In the case of a device under test, of which the input is not excited by a noise source, the output noise power at the output of the device under test N_OUT, which results exclusively from the noise power present at the input of the device under test, corresponds according to equation (2) to the thermal noise N₀ of a 50-Ω resistor system at a room temperature of 290 K amplified with the power amplification G_DUT of the device under test.

N _OUT ′=N ₀ ·G _DUT  (2)

Accordingly, the noise figure F_DUT according to equation (3) is obtained:

$\begin{array}{cc}{F}_{\mathrm{DUT}}=\frac{N_{\mathrm{OUT}}}{N_{\mathrm{OUT}}^{\prime }}=\frac{N_{\mathrm{DUT}}+N_0}{N_0}& \left(3\right)\end{array}$

The noise factor NF_DUT of a device under test represents the log of the noise figure F_DUT of the device under test as shown in equation (4)

NF _DUT=10·log(F_DUT)  (4)

The measured value P_NOISE of the output noise power N_OUT of the device under test to be measured by the network analyzer is superposed with the internal noise N_NWA of the network analyzer according to equation (5).

P _NOISE =N _NWA +N _OUT  (5)

If equations (1), (3) and (5) are combined, the formula contained in equation (6) for determining the noise figure F_DUT is obtained.

$\begin{array}{cc}{F}_{\mathrm{DUT}}=\frac{P_{\mathrm{NOISE}}-N_{\mathrm{NWA}}}{G_{\mathrm{DUT}}·N_0}& \left(6\right)\end{array}$

The power amplification G_DUT of the device under test is obtained with a 50-Ω system from the forward-S-parameter S₂₁ of the device under test according to equation (7):

G _DUT =|S ₂₁|²  (7)

If the input impedance of the network analyzer does not provide the value 50Ω as a result of an error matching, the power amplification G_DUT must be multiplied by an error-matching factor M according to equation (8), wherein, in this context, Γ_NWA is the input-reflection coefficient of the network analyzer and S₂₂ is the output-reflection coefficient of the device under test.

$\begin{array}{cc}M=\frac{1-{{\Gamma }_{\mathrm{NWA}}}^2}{{1-S_{22}·{\Gamma }_{\mathrm{NWA}}}^2}& \left(8\right)\end{array}$

The thermal noise N₀ of a 50-Ω resistor system is obtained from the room temperature T of 290 K and the Boltzmann constant k according to equation

$\begin{array}{cc}4·{10}^{-21}\frac{W}{\mathrm{HZ}}=-174\frac{{\mathrm{dB}}_m}{\mathrm{Hz}}.\mathrm{as}& \left(9\right)\\ N_0=k·T=4·{10}^{-21}\frac{W}{\mathrm{HZ}}=-174\frac{{\mathrm{dB}}_m}{\mathrm{Hz}}& \left(9\right)\end{array}$

The internal noise N_NWA of the network analyzer, which is additively superposed on the output noise power N_OUT of the device under test, and an inaccuracy k of the network analyzer, which is multiplicatively superposed on the correct measured value P_NOISE according to equation (10) and leads to an incorrect measured value P_Mess, are determined in a calibration process of the network analyzer.

P _Mess =k·P _NOISE  (10)

For this purpose, the port to be measured of the network analyzer is connected to a noise source, which is switched off. The incorrectly-determined measured value P_MessOFF at the port of the network analyzer corresponding to the correct measured value P_NOISEOFF is obtained from the internal noise N_NWA of the network analyzer with the noise source switched off and the thermal noise N₀ of a 50-Ω resistor system at a room temperature of 290 K in the case of a matched measuring circuit according to equation (11).

P _MessOFF =k·P _NOISEOFF =k·(N_NWA+N₀)  (11)

According to equation (12), the incorrectly-determined measured value P_MessON corresponding to the correct measured value P_NOISEON with the noise source switched on is obtained in a corresponding manner from the internal noise N_NWA of the network analyzer, the thermal noise N₀ of a 50-Ω resistor system at a room temperature of 290 K, in a frequency range of 1 Hz in the case of a matched measuring circuit, and the excess noise N_NS of the noise source.

P _MessON =k·P _NOISEON =k·(N_NWA+N₀+N_NS)  (12)

The excess noise N_NS of the noise source can be determined according to equation (13) from the ENR factor (excess-noise ratio) specified by the manufacturer of the noise source, which indicates the ratio of the excess noise of the noise source relative to the thermal noise N₀ of 50-Ω resistor system at a room temperature of 290 K.

$\begin{array}{cc}{N}_{\mathrm{NS}}={10}^{\frac{\mathrm{ENR}}{10}}·N_0& \left(13\right)\end{array}$

From equations (11) to (13), the inaccuracy k is obtained according to equation (14), and the internal noise N_NWA of the network analyzer is obtained according to equation (15) in each case by mathematical transformation.

$\begin{array}{cc}k=\frac{P_{\mathrm{MesON}}-P_{\mathrm{MessOFF}}}{{10}^{\frac{\mathrm{ENR}}{10}}·N_0}& \left(14\right)\\ N_{\mathrm{NWA}}=\frac{P_{\mathrm{MesOFF}}}{k}-N_0& \left(15\right)\end{array}$

Starting from equation (6) with the inaccuracy k calculated according to equation (14) taking into consideration equation (10), the noise figure F_DUT for a device under test to be measured is obtained according to equation (16) from the incorrectly-determined measured value P_Mess of the output noise power N_OUT of the device under test:

$\begin{array}{cc}{F}_{\mathrm{DUT}}=\frac{\frac{P_{\mathrm{Mess}}}{k}-N_{\mathrm{NWA}}}{G_{\mathrm{DUT}}·N_0}& \left(16\right)\end{array}$

The method according to the invention for measuring the noise figure of a device under test with a network analyzer begins with the calibration of the network analyzer. For this purpose, in a calibration arrangement as shown in FIG. 2, the noise power of a noise source to be measured is connected to the first measuring port 3 of the network analyzer (NWA) 4. During the implementation of the method according to the invention for measuring the noise figure, the arrangement is operated at the constant room temperature T of 290 K, and the noise power spectrum to be registered is measured with a certain measurement bandwidth of the network analyzer 4 in order to obtain a thermal noise N₀ of the matched measurement circuit, which corresponds to a 50-Ω resistor system, with standardized scaling

$\mathrm{value}-174\frac{{\mathrm{dB}}_m}{\mathrm{Hz}}.$

In the first procedural stage S10 of the method according to the invention for measuring the noise figure of a device under test with a network analyzer according to FIG. 1, with the noise source 6 switched off, the measured value P_MessOFF at the first measuring port 3 of the network analyzer 4 is measured according to equation (11). In a similar manner, with the noise source 6 switched on, the measured value P_MessON at the first measuring port 3 of the network analyzer 4 is measured according to equation (12).

In the next procedural stage S20, with the two measured values P_MessOFF and P_MessON buffered, the inaccuracy k and the internal noise N_NWA of the network analyzer 4 are calculated according to equations (14) and (15) and buffered.

Following this, the calibration arrangement is converted into a measuring arrangement according to FIG. 3, in which the second port 5 of the device under test 2 is connected to the second port 7 of the network analyzer 4.

In this measuring arrangement according to FIG. 3, the individual S-parameters of the device under test 2 are determined according to known methods in procedural stage S30 using the network analyzer 4, and from the measured forward-S-parameter S₂₁, the power amplification G_DUT of the device under test 2 is determined according to equation (7) and buffered. If the input impedance at the first port 3 of the network analyzer 4 is not matched to the value 50-Ω, the output reflection coefficient S₂₂ of the device under test 2 is multiplied by an error-matching factor M according to equation (8) in order to achieve a matching of the measuring arrangement in this manner.

In procedural stage S40, while the second port 5 of the device under test 2 connected to the second port 7 of the network analyzer 4 is not excited by the network analyzer 4 with a generator signal in the form of a noise signal, the measured value P_Mess of the output noise power applied at the first port 1 of the device under test 2 is measured at the first port 3 of the network analyzer 4.

With the measured value P_Mess of the output noise power applied at the first port 1 of the device under test 2 determined in the preceding procedural stage S40, the power amplification G_DUT of the device under test 2 determined in procedural stage S30, the two parameters—inaccuracy k and internal noise N_NWA of the network analyzer 4—determined and calculated in procedural stages S10 and S20—and the standardized parameter of the thermal noise N₀ of a 50-Ω resistor system, the noise figure F_DUT is calculated in the final procedural stage S50 according to equation (16).

Claims

1. Method for measuring the noise figure (FDUT) of a device under test using only a noise source and a network analyzer, said device under test characterized by power amplification (GDUT) and forward-S-parameter (S21) values and comprising first and second ports, said network analyzer comprising first and second measuring ports and being characterized by inaccuracy (k) and internal noise (NNWA) values said method comprising: determining the inaccuracy (k) and the internal noise (NNWA) of the network analyzer from a measured value (PMessOFF), measured in a calibration process with the noise source switched off, at the first measuring port of the network analyzer connected to the noise source and a measured value (PMessON), measured with the noise source switched on, at the first measuring port of the network analyzer connected to the noise source;determining the power amplification (GDUT) of the device under test from a measured forward-S-parameter (S21) of the device under test, while the first port is connected to the first measuring port of the network analyzer and the second port is connected to the second measuring port of the network analyzer;measuring, at the first port of the network analyzer, a measured value (PMess) of output noise power applied at the first port of the device under test, while the second port of the device under test is connected to the second port of the network analyzer and is not excited by the network analyzer with a generator signal in the form of a noise signal; andcalculating the noise figure (FDUT) from the measured value (PMess) of the output noise power, the power amplification (GDUT) of the device under test, the inaccuracy (k), and the internal noise (NNWA) of the network analyzer, and thermal noise (N0) of a 50-Ω resistor system.

2. Method for measuring the noise figure of a device under test according to claim 1, wherein the first measuring port of the network analyzer measuring the output noise power (NNWA) applied at the first port of the device under test has an input impedance of 50Ω at a room temperature of 290 K.

3. Method for measuring the noise figure of a device under test according to claim 1, comprising calculating the noise figure (FDUT) of the device under test according to the following formula:

4. Method for measuring the noise figure of a device under test according to claim 1, comprising in the case of an error-matching of the input impedance of the first port measuring the output noise power (NNWA) applied at the first port of the device under test, multiplying an output-reflection coefficient (S22) measured at the first measuring port of the network analyzer by an error-matching factor M, which is calculated according to the following formula:

5. Method for measuring the noise figure of a device under test according to claim 1, comprising calculating the measured value PNOISE of the output noise power NNWA applied at the first port of the device under test to be measured at the first measuring port of the network analyzer by dividing a measured value PMess of the output noise power NNWA applied at the first port of the device under test, incorrectly measured by the network analyzer, by an inaccuracy k of the network analyzer determined by calibration according to the following mathematical relationship:

6. Method for measuring the noise figure of a device under test according to claim 5, comprising calculating a measured value PMessOFF of the output noise power NNWA applied at the first port of the device under test, incorrectly measured at the first measuring port of the network analyzer in a calibration process with the noise source switched off and connected to the first measuring port of the network analyzer according to the following formula: PMessOFF=k·(NNWA+N0),whereinN0 denotes the thermal noise of a 50-Ω resistor at a room temperature of 290 K; andk denotes an inaccuracy of the network analyzer 4.

7. Method for measuring the noise figure of a device under test according to claim 6, comprising calculating the measured value PMessON of the output noise power NNWA applied at the first port of the device under test, incorrectly measured at the first measuring port of the network analyzer in a calibration process with the noise source switched on and connected to the first measuring port of the network analyzer according to the following formula:

8. Method for measuring the noise figure of a device under test according to claim 7, comprising calculating the inaccuracy (k) of the network analyzer according to the following formula:

9. Method for measuring the noise figure of a device under test according to claim 8, comprising calculating the internal noise NNWA of the network analyzer according to the following formula: