Measurement of noise contributions

This is also referred to as white noise measurement. Both the thermal noise contributions and the intermodulation noise contribu­tions can be measured directly with one test. This allows rapid field evaluation of installed systems and bench evaluation of individual circuits.

In Figure 20.14 the system is loaded with Gaussian noise from a generating source that has a flat frequency spectrum of noise, bandwidth limited by the filter F.

A single defined channel at frequency f(c) is removed from the spectrum of noise by a filter F before the noise band is transmitted into the system.

A measurement on the same channel at frequency f(c) selected by filter F at the receiver will therefore measure noise that has been generated only within the system itself.

The system is loaded with noise power P that represents the expected multichannel load, as in Equation 20.30, where P The total power of the signal applied to the system, P is the multichan­nel load for N channels, and T is the transmission level point in dBm.

For instance, the figures for a 4MHz system at a -3()dBm trans­mission level point will be P = +14.8dBmO (assuming the normal channel loading of-15dBmO). P will evaluate to-15.2dBm at T = -30dBm.

Two measurements are taken. The first is with the channel stop filter at the transmitter by-passed with the switch S in the closed position. This gives the power due to simulated traffic in a single channel referred to the meter calibration. Measured figure = N.

The second measurement is with the channel stop filter at the transmitter in circuit with the switch S open. This gives the noise power generated from the system referred to the same meter cali­bration. Measured figure = N.

The Noise Power Ratio (NPR) is given by 101og(Ns/Nn) dB.

The total noise power falling into the channel can be determined from Equation 20.31, where N is the total weighted noise contribu­tion over 4kHz in the specified channel in dBmOp, k = B/4N where B is the bandwidth of the system bandwidth limiting filter in kHz, and N is the number of channels.

The term l0 log k is to correct for the fact that the actual traffic load has gaps between the channels and is not continuous. The total noise power is also given in pWOp by Equation 20.32.

Several channels are selected for measurement over the band by varying the frequency f(c) of both the channel stop filter in the transmitter and the channel pass filter in the receiver.

Also of interest is to vary the channel loading factor L (nominally -15dBmO) and determine the NPR for various conditions of traffic load. This is a measure of the system's robustness to peak loads and transmission level changes outside the normal design guides. The normal expected NPR curve is shown in Figure 20.15.

As the transmission load is lowered then the basic or Gaussian noise predominates. The NPR increases linearly with decreasing traffic power. As the transmission load is increased then intermodu-lation noise predominates and the NPR rapidly rises as more higher order intermodulation products contribute.

The optimum working point of the system is easily determined from Figure 20.15