S21 measures transmission
S11 describes reflection at the input; S21 describes transmission from port 1 to port 2. To investigate a passive filter, connect its input to the source port (usually CH0 or PORT1) and its output to the receiver port (CH1 or PORT2). The NanoVNA provides the test signal, so no external transmitter is needed.
Calibrate the complete test setup
- Select a frequency range that includes the expected passband and enough of the stopband to see the filter's behaviour.
- Fit the test cables and adapters you will use. Follow your model's full calibration sequence at the device connection planes, including open, short and load as required.
- Join the cable ends with the appropriate through connection and record THRU. Follow the model instructions for any isolation measurement; not every NanoVNA has a separate isolation step.
- Finish calibration and check the through connection. S21 log magnitude should be close to 0 dB over the calibrated range.
- Insert the passive filter in place of the through. Select S21 in LOGMAG format and use markers to inspect its response.
The procedure depends on the hardware and firmware; see the NanoVNA V2 calibration reference or the original NanoVNA instructions. Keep the cables in the same arrangement after calibration.
Read insertion loss correctly
The display uses S21(dB) = 20 log10|S21|. A reading of −1 dB means insertion loss of 1 dB under the measurement conditions. In a matched system, −3 dB corresponds to roughly half the transmitted power and −20 dB to one hundredth. Keep the sign clear: more negative transmission means greater attenuation.
As a worked example, suppose a band-pass filter peaks at −1.2 dB. Its bandwidth measured 3 dB below that peak uses the two −4.2 dB crossings, not the −3 dB absolute level. Subtract the lower crossing frequency from the upper one. State whether your reported bandwidth is relative to the peak or uses an absolute transmission limit.
Recognise the filter response
A low-pass filter passes low frequencies and attenuates higher ones. A high-pass filter does the reverse. A band-pass filter has a window of low loss between two attenuated regions, while a band-stop filter rejects a band or narrow notch. Check the passband ripple as well as the cutoff or notch frequency; one marker cannot describe the entire response.
Know the measurement limits
Very deep notches may reach the instrument's noise or leakage floor. If a stopband trace stops falling, that does not establish the filter's true attenuation. Repeat with a narrower span, suitable averaging and a careful cable layout, and compare with the through check. Use enough sweep points to resolve narrow features.
These measurements use the instrument's nominal 50-ohm environment. A filter intended for another impedance may behave differently in its actual circuit. S21 loss can include both dissipation and mismatch; inspect S11 as well when assessing the input match. This guide covers passive filters, not powered amplifiers or transmitter outputs.
For the input-side view, continue with reading the Smith chart.