How to Measure Coax Cable Loss with a NanoVNA

[Pavel]

Short answer: Use a calibrated two-port S21 measurement with the cable connected between ports, or a one-port return-loss method with a known open or short when two-port connection is impractical. Calibrate at the cable ends so test leads and adapters are excluded.

Before you start​

• Check maximum RF input level and DC tolerance before connecting a transmitter, powered line or unknown source to test equipment.
• Warm up the instrument, use known cables and adapters, and calibrate at the connector plane where the device under test will be attached.

Step-by-step method​

1. Step 1: Set the frequency span of interest and perform full two-port calibration at the two free test-cable ends. Verify a direct through connection reads close to 0 dB within expected instrument accuracy.
2. Step 2: Insert the coax under test without changing adapters. Display log-magnitude S21 and record loss at each target frequency; cable loss normally increases with frequency.
3. Step 3: Reconnect and flex only within the cable's safe bend radius. A changing trace can reveal a damaged connector or intermittent shield.
4. Step 4: For a long installed cable measured from one end, terminate the far end with a known open or short and use the round-trip loss method, remembering measured loss represents two passes through the cable.

Concrete example​

If two-port S21 is -1.8 dB at 100 MHz and -5.2 dB at 1 GHz after through calibration, those are approximate one-way insertion losses at those frequencies. In an open-end return method, divide the round-trip loss contribution by two after accounting for the reflection standard.

How to judge the result​

The result is reliable when reconnecting produces nearly the same trace and a known short through adapter verifies the calibration plane. Record cable length, type, connectors and temperature.

What to record​

• Save frequency span, calibration state, attenuation, reference level and connector plane with every result.
• Repeat the measurement after reconnecting the device; a large change often points to a poor adapter, cable or calibration.
• Use a VNA for reflection and network measurements, and a spectrum analyzer for signal energy versus frequency.

Common mistakes​

• Changing several hardware, software or RF variables at once, which removes the controlled comparison needed to identify the cause.
• Treating one autoscaled screenshot or one unusually good result as proof without recording the settings and repeating the test.
• Subtracting traces taken with different adapters or calibration planes can create a loss value that belongs mainly to the test setup rather than the coax.

Final check​

Reconnect the standards or device under test and repeat the measurement without changing the calibration plane. Save the span, calibration, attenuation and uncertainty with the result.