Noise figure confusion

Hi all

As I recall the NOISE FIGURE of an amplifier is defined as : Signal to Noise at output divided to Signal to Noise at Input .
How exactly that NOSE at input and at output is measured

Thanks
Elico

BarNash said:

Hi all

As I recall the NOISE FIGURE of an amplifier is defined as : Signal to Noise at output divided to Signal to Noise at Input .
How exactly that NOSE at input and at output is measured

Thanks
Elico

Click to expand...

Your definition is only one possible out of more definitions.
Noise figure is what an amplifier or another device adds to the thermal noise of a matched (to a transmission line) resistor at a nominal ambient temperature.
The thermal noise is defined by noise power,

Pn = kTB,

where k is the Boltzmann constant, T is the temperature in Kelvins, and B is bandwidth in Hz.

An attenuator in a transmission line has a noise figure equal to its loss. Amplifiers add usually more.

Measuring noise figure of an amplifier or a receiver requires a calibrated power meter connected to the output, and two noise sources connected one after another to the input.

The change of output noise power due to the two input levels is named "Y-factor", and the measured ratio indicates to amplifier noise figure.

The detailed explanation is rather long, please google or find a textbook on noise figure measurement. Agilent issued long ago a good white paper, "noisecom.com" and other websites offer detailed procedures, too.

Thanks
Elico

There is an equation adopted by IEEE for the Noise Figure of a Device:

Device NF[dB] = 10*LOG [(Na + k*To*B*G) / (k*To*B*G)]

where: Na=Added Noise Power [W], To =290K, B=Bandwidth[Hz], G=Gain[Linear] , k=Boltzmann constant.

Regarding your question about Noise Figure accuracy, most of the RF systems can accept +/- 0.1dB noise figure accuracy.

"Na=Added Noise Power [W] " Whats that exactly ?

Elico

BarNash said:

"Na=Added Noise Power [W] " Whats that exactly ?

Elico

Click to expand...

Again it it the noise power a device adds over the resistor thermal noise.

Have you tried to study the basics? Otherwise you will stay confused (like I was till I started to study)

Thanks I will devote time to study the basics .
Elico

Be aware that noise figure isn't always a useful indicator of system performance

For example a reverse-biassed PIN photodiode driving a transimpedance amplifer will have a very low noise even thought the "noise figure" of the amplifier is extremely high. The quote marks indicate that where a high-impedance source is driving a low-impedance load, noise figure is meaningless.

Noise figure is defined in systems where devices are matched to a common line impedance. The case with PIN photodiode is different also by that there is no specified input noise source to be used (light noise).

jiripolivka said:

Noise figure is defined in systems where devices are matched to a common line impedance.

Click to expand...

Precisely.

The case with PIN photodiode is different also by that there is no specified input noise source to be used (light noise).

Click to expand...

Not really, since 1W->0.5A, and since the dark/noise current is defined, so there is a direct linear relationship, which is sufficient.