This project, based on an article published in the February 2008 edition of Radcom (the magazine of the Radio Society of Great Britain) is for a meter with a 50 ohm input impedance for measuring very small RF powers. Many RF power meters become inaccurate at low powers, but this meter is able to measure down to very low levels. From a theoretical standpoint, the meter is only able to accurately measure the power of steady, sine wave voltage RF - it would not be suitable for square wave signals for example. In practice, however, the most common requirement is to measure the power of sine wave signals and to compare power readings (when peaking up IF stages, for example). The reason that the meter is best suited to sine waves is that the circuit actually measures the peak-to-peak voltage in order to infer the power.

The Radcom article explains how many RF 'sniffer' type projects rely on detecting RF using a diode. This diode will require a certain amount of voltage to bring it into forward conduction, and this is the source of inaccuracy at low powers for more conventional designs.

The original article gives details for a meter based on PNP transistors. I am more used to understanding circuits based on NPN transistors, so I first redrew the circuit diagram with NPN transistors and the diodes and PSU reversed. Once I had a good understanding, I realised that the design can be broken down into two sections; I call these the RF section and the DC voltmeter section.

The RF Section

The RF part of the circuit is basically a dummy load with two detecting diodes to get the peak positive and peak negative voltages. A DC blocking capacitor is connected at the input, which acts as a short circuit at RF frequencies. What makes this design different is that the detecting diodes are driven into conduction by a small dc current. In the circuit diagram from the original article, the power supply (a 9V battery) is shown as a 0V earth and a +9V supply line, and a +4.5V reference is derived for the earth of the RF section. To make the understanding of the RF section easier, however, I have chosen to show this as a 0V RF and DC earth, with +4.5V and -4.5V supply lines. The difference is academic. The circuit diagram is shown below:

RF Section Circuit Diagram.

The input signal passes through C1 and into the dummy load formed by R1. The diodes D1 and D2 then rectify the negative and positive peaks respectively of the signal across R1. The rectified voltages appear across C3 and C4, minus the forward voltage drop of the diodes. To bring the diodes into conduction and improve linearity, a small DC current of a little under 0.5µA passes from the +4.5V supply line, through R4, D1, D2 and Rx to the -4.5V supply line. Note that in the Radcom article, Rx was not used and instead the current flowed into the DC measuring circuit input impedance.

The DC blocking capacitor C1 will have a reactance of about 1.6 ohms at 1MHz, so this device will give less accurate results below 1MHz unless the value of C1 is increased.

The DC Voltmeter Section

This part of the circuit measures the difference between the peak negative and peak positive voltages. The original circuit had two halves, each consisting of two transistors connected in a Darlington confiuration, to give an emitter follower of high input impedance. As the emitter followers are referenced to the negative supply line, this gives no problem in measuring the peak negative voltage of the input RF.

Voltmeter Section Circuit Diagram.

In order to generate the 0V for the RF section, a potential divider circuit and decoupling capacitor is used, as in the following circuit diagram.

Potential Divider Circuit Diagram.