Receiver Sensitivity Measurements

1. General

There are two basically similar ways of defining the sensitivity of a receiver which are used for most "typical" HF and VHF communications receivers - for a given RF signal input level, modulated or unmodulated depending on the mode under test, the minimum signal to noise ratio or SINAD ratio at the audio output is specified. Signal to noise is easier to measure as it only requires a calibrated audio output detector e.g. test meter, audio power meter etc., whereas to measure SINAD requires a calibrated distortion meter.

However, these methods are not ideal for extremely high sensitivity VHF and UHF receivers because it is difficult to prevent signal leakage from the signal generator affecting the measurements. A more accurate method is to use a noise generator instead of a signal generator although this is not suitable for FM only receivers. A noise generator consists of a special thermionic or solid state diode, good quality resistive termination and suitable power supply. Depending on the device, the noise output will be specified for a given fixed or range of diode current. Specialist suppliers can provide semiconductor noise diodes but in the UK there are also a range of fairly easily available thermionic noise diodes of which the CV2171 (B7G, 150MHz max) and CV2398 (wire ended B9A, 450MHz max) are examples.

2. Noise Generator

Most surplus noise generators contain the noise diode in the main case which can lead to measuring errors at VHF and UHF. If you are fortunate enough to be able to obtain one of these items, the first task is to remove the noise diode assembly and rebuild it in a small metal box connected to the power supply by a suitable length of multiway cable. Examples of thermionic noise generators are shown in the RSGB and ARRL Handbooks and are not difficult to build if you have some experience of home construction.

Thermionic noise diodes have a directly heated tungsten filament fed from a nominal 6.3v AC source via a low value wire wound variable resistor and a DC HT supply of about 150v at up to 20mA with the positive leg grounded.

The author has modified a surplus CT82 by rebuilding the noise generator section in a small die cast box approximately 1.25 x 1.4 x 3.4 inches with a BNC connector, resistive 50ohm load, associated decoupling components and cabling and connector to the PSU. The power supply, switching and metering components were all tested but left in the power supply assembly - carefully check all components, particularly any meters as they may become unreliable with age. The CT82 also contains a simple AF power meter and load but the author's model gave problems and was discarded in favour of an external unit that was known to work correctly. Click here for more CT82 details.

The measurement method is to connect the noise generator to the receiver under test with a suitable coaxial connector and the shortest possible length of coaxial cable (ideally just a back to back coaxial connector) and the receiver AF output to the AF power meter. Turn the receiver on and select SSB or CW, adjust the AF gain so that noise is audible and the RF gain so that the noise level is below the AGC threshold. You can use AM for this measurement but the non-linearity of the diode detector at low signal levels can (and does) produce innaccurate (too optimistic) results.

With the noise generator diode current turned off, set the receiver AF noise output to 0dB or a suitable power level - say 1mW, turn the noise generator on and adjust the diode current so that the AF output level has doubled i.e. +3dB or 2mW and note the noise diode current. Turn the noise generator current off between measurements as the noise diodes do not have a long working life and can overheat at high currents.

The noise factor F in dB is calculated using the formula:

F = 10 Log10(20 x I x R)

where the log is to base 10, the current I is in amperes and the resistor R is 50 ohms.

Typical HF communications receivers have noise factors in the range 10-15dB and good VHF converters/receivers 1-3dB.