VU3SQM directional wattmeter build – #2

VU3SQM directional wattmeter build – #1 laid out the first steps in design review and build of a directional wattmeter.

This article canvasses the issues of the display.

Intention is a digital based display (though not to exclude an analogue meter or bar graph type displays).

So, the output of the AD8307 needs to be digitised.

Let’s first consider the nature of the AD8307.

It is a log detector, so it provides a ‘DC’ voltage proportional to the log of the input signal, but the ‘DC’ voltage can vary very quickly.

The chart above from the AD8307 datasheet shows that the unfiltered response to a burst of RF has a rise time of well under 1µs.

Peak Envelope Power

If we were to reliably capture the Peak Envelope Power of an SSB telephony signal, we must sample it sufficiently often to capture the peaks.

The instantaneous power of a SSB telephony signal varies with the speech frequencies, and sampling theory tells us that to fully capture it, it must be sampled at greater than 8kS/s, which implies and ADC conversion time of 125µs.

This is beyond the capability of low cost microcontrollers, and low cost ADC chips.

An alternative approach is to use a rather conventional peak holding amplifier using analogue techniques between the AD8307 and ADC.

A combination approach is likely to be effective, a conventional peak holding amplifier using analogue techniques with fast capture time and moderate decay time between the AD8307 and ADC, and extension of the decay time performed digitally. This reduces the size of the integration capacitor, improving the rise time response, whilst still giving a usable hold time to the display of PEP.

Average Power

The log detector provides a means of calculating Average Power. (Note that hams often refer to power meter readings as Average Power, but in most cases it is not Average Power.)

It is probably not practical to sample at sufficient rate to calculate true average power.

Nevertheless smoothing the AD8307 outputs for calculation of Return Loss should give an accurate result with minimal measurement noise.

Return Loss & VSWR display

The question also arises whether it is better to calculate the Return Loss digitally, or to us an op amp to create a Return Loss signal from the Pfwd and Prev detectors (avoiding temporal issues of sequential ADC). The advantage of the latter is that with a little filtering, it is probably easier to create a glitch free Return Loss display and VSWR display as it is calculated directly from Return Loss.


Alarms can be derived digitally, and give visual and / or sound alerts, and optionally be used to inhibit a transmitter under exceptions.

Emerging architecture

The architecture emerging is the VU3SQM directional coupler followed by some linear stages using op amps to develop:

  1. fwdPEP signal, fwdSMOOTHED, revSMOOTHED, and fwdSMOOTHED-revSMOOTHED; or
  2. more simply, simply fwdPEP signal, revPEP, and fwdPEP-revPEP

signals which are digitised in four or three channels of ADC.


… more as it unfolds…