Noise Voltage Measurement

1. Basic characteristics of noise

In computer science, noise refers to all unnecessary signals and various electromagnetic interferences other than useful signals. Noise is a random signal that needs to be described using statistical methods.

Figure 1 Noise voltage waveform

Light that contains all colors is called white light. Similarly, noise with equal power density at all frequencies is called white noise. True white noise should have infinite bandwidth and therefore infinite power, but the bandwidth of an actual system is always limited. As long as the noise has a flat power spectrum density within the frequency band under study, we can regard it as white noise.

2. Measuring noise voltage with an average value meter

Average value: ;

According to the definition of waveform coefficient, we get:

According to the principle of equal average value, the effective value of the noise voltage is calculated:

3. Measurement of Device and Amplifier Noise

A noisy amplifier can be equivalent to an ideal noiseless amplifier, and the actual output noise voltage uno is equivalent to the input of the noiseless amplifier. In Figure 2, us and rs are the voltage and internal resistance of the signal source respectively, uno is the output voltage of the signal source, and the voltage gain of the ideal device is au.

Figure 2 Equivalent input noise source model

Then the equivalent input noise voltage uni of the amplifier is defined as , where the output noise voltage uno includes the thermal noise of rs and the noise generated by the internal devices of the amplifier.

1. Measurement of equivalent input noise uni

(1) Sinusoidal signal method

Figure 3 Sine signal method to measure uni

Using a sine wave as the measurement signal, the voltage gain au of the amplifier is measured, as shown in Figure 3, which is defined as:

(2) Noise generator method

Figure 4. Measuring uni using the noise generator method

A noise generator can also be used instead of a sinusoidal signal source for measurement. Assume that the noise generator outputs a noise voltage of uns. When measuring, do not connect the noise generator first, connect only rs to the amplifier input, and measure the effective value of the amplifier output noise voltage un1. When measuring, connect the noise generator again, and measure the effective value of the amplifier output noise voltage un2.

According to the definition of uni, we have :

(3) Comparison of the two measurement methods

(4) Issues to consider when measuring noise

① Impact of bandwidth

②Average value and waveform factor

③The influence of measurement time

④ Since white noise has a Gaussian normal distribution, in order to avoid the noise peak being clipped, the meter reading must not exceed half of the full scale value during the measurement process.

3. Measurement of noise bandwidth

(1) Sinusoidal signal method

Figure 5. Calculation method of bwi

Figure 6b: Noise bandwidth measured using a signal generator

(2) Noise generator method

3. Signal-to-noise ratio test

Figure 7 Signal-to-noise ratio measurement principle

4. Noise figure measurement

There are many ways to express the noise factor. The most widely used one is to calculate the noise factor using the signal/noise ratio. In the amplifier circuit shown in Figure 8, and rs are the signal source voltage phasor and internal resistance respectively, and rl is the load resistance.

Figure 8 Noise coefficient schematic diagram

The noise factor of the amplifier is defined as the ratio of the input signal-to-noise ratio to the output signal-to-noise ratio of the amplifier at a standard temperature of 290o K, that is,

The noise coefficient f is usually expressed in decibels, which is also called the noise index.

(2) Sinusoidal signal method

The measurement principle of the sinusoidal signal method is the same as that of the noise generator method, except that the signal source uses a sinusoidal constant voltage source. This method is more complicated to measure and has larger errors. It is only suitable for low frequencies or situations where there is a lack of dedicated noise generators.