Calibrating a cheap eBay RF power meter

Calibrating an RF power meter

I needed a rough and ready measurement of RF power at the -40 to 0dBm range, at frequencies from a few hundred kHz to VHF. I’ve got interested in some radio projects again, specifically making some small signal amplifiers. To check how well the amplifiers work, I need to know what input signal I’m driving them with. I can easily measure their relative gain using a VNA and calibrated attenuators, but had no way to know accurately how much signal I was driving them amplifier with.

The meter has sat unused for over a year because it came with zero documentation. Turing it on just yield a frequency selection and somewhere to set the offset of the device. It’s powered from USB and does appear as a USB device on the computer, but there’s no detail on how to talk to it.

My RF signal generators are still in calibration for frequency, but I have not been able to test them at the low signal levels I’m interested in. My oscilloscope can’t measure very small signals, but can give some indication down to about -30dBm.

I decided to check the output of my nanoVNA (which is specified at ~1 MHz to be -15dBm) with my oscilloscope.

The nanoVNA was set to produce a continuous signal at 1.5MHz and the output was taken to the 50Ω terminated input of the oscilloscope. The output of the nanoVNA is a square wave, rather than the more useful (for this purpose) sine-wave.

The measured peak to peak voltage is 232 mV, which into 50 Ohm is -8.7 dBm, assuming a sine-wave. As this isn’t a perfect sine-wave the power is different, and will depend on the duty cycle of the square wave. This output seems to be very close to 50% duty cycle, so the actual power will be larger than the calculated value (square waves contain lots of harmonics, each with some power, so all the additional harmonics add to the power in the signal).

I built a simple filter to remove the harmonics and measured again. This time, including insertion loss from the filter, the measured peak-peak voltage was 204mV or -9.8 dBm.

The signal was then applied to the power meter, and the meter indicated -39 dBm, revealing an undocumented 30 dB attenuator in the input. This can be calibrated out with a menu item in the meter.

I ended up using the oscilloscope to calibrate my signal generator, and use this to start to calibrate the power meter. In the end I bought a tinySA spectrum analyser, and have had much more success with this, as it shows (pretty accurately) the power and frequency components of a signal.

I’d not really recommend anyone buy one of these meters expecting it to just work, the documentation is nonexistent, so there’s no useful information regarding frequency ranges, linearity or accuracy. Save up a bit longer and get a genuine tinySA instead.