There are several sources of noise in electronic systems. Noises are unwanted signals polluting random or not signals that reduce overall signal quality.
Thermal Johnson/Nyquist noise
Two scientists, Johnson, discovered this noise who did the experiments and Nyquist, who developed the formula. Thermal noise is present at all frequencies (has a constant power density at all spectra) and is called white noise. This noise can only be reduced by reducing the temperature, resistance, and bandwidth. Let’s see the Noise voltage RMSformula:
where k – Boltzmanns constant (1.38·10-23J/K);
T – Absolute temperature (K);
R – Resistance (Ohms);
Δf – Bandwidth (Hz).
For example, if we have a 10kOhms resistor at 300K. Bandwidth 10kHz then we get VRMS=1.3μV.
This type of noise occurs because of random charges moving across the potential barrier.
where e- charge of electron (1.6·10-19C);
I – DC signal current (A);
In — Noise current (A);
Δf – Bandwidth (Hz).
Usually, Shot noise is more due to the high-frequency component. So reducing bandwidth may reduce high-frequency noise.
This noise increase due to decreasing frequency. So measuring sensitive measurements is better to use DC. Flicker noise origins aren’t known and not well understood. Usually, it is not important compared to noises over 1 kHz.
This type of noise comes from outside sources. This type of noise is often called interference. Interference may be mechanical or electrical. Electromagnetic interference (EMI) is most common that arise from:
- Radiation from switching devices;
- Radiation from AC power lines, rectifiers;
- Flashes of Lightning;
- Transmitter RF noises(radios, cell phones);
- Electronic devices working at high frequencies;
- Electrostatic discharges.
Reduction of environmental noises should start from noise sources. If it’s possible, the best practice is to remove it. It is not always possible to do so. Then the second step is to divert the noise signal to the ground using filters. If we talk about measurements, then it is important to remember that the major noises contribute to the first amplification stage. So preamplifiers should be differential and with good CMRR and located closer to the transducer. Cables between the amplifier and further stages should be shielded. Shields should be grounded to common pint to eliminate ground loops.
Signal to noise ratio
When we are talking about noises, there is a need to measure it. It is common to measure a relative magnitude of the noise given by signal to noise ratio SNR (dB). Signal to noise ratio (SNR) is the ratio of signal power over the noise power:
So the larger SNR is the better.
If expressed in decibels:
Vs and Vn are RMS values used. Noises are always present in signals. In the circuits with amplifiers, new noises can be added by the amplifier itself. This is called SNR degradation. This degradation from amplifier input to output is called Noise Figure (NF). NF less than 3dB is considered good:
SNRI – signal in the amplifier input;
SNRO – signal in the amplifier output.
Signal degradation increases NF of device raises because of thermal noise and flicker noise (< 1kHz) because of semiconductor resistance and randomness of the diffusion processes in PN junction.