In this post I’ll share some more NanoVNA measurements, this time of a T-match series-C/shunt-L/series-C antenna tuner, designed for 1-30 MHz. This off-the-shelf manual tuner + SWR meter accepts PL-259 connections for the transceiver and two antennas; a fourth SO-239 RF connector is available to attach a 50 ohm dummy load. The loads are selectable with a wafer switch and can connect through the tuned circuit or can be bypassed straight from transceiver to selected antenna. The tunable match is achieved by two meshing air capacitors on the Antenna and Transceiver/Transmitter side, with a tapped coil shunted between them, again selecting taps with a wafer switch.

## T-Matching Network: Theory

Matching networks are commonly used to adapt antennas to the impedance of the transmitter/receiver/transceiver. Matching networks may be used to achieve a narrowband impedance match between two networks to reduce the losses caused by reflections between mismatched impedances. Some references 1, 2

## Measured Self-Impedance

The 50 ohm load presented by port 2 of the NanoVNA means that for any measurement that switches through the T-match circuit of the antenna tuner, the match is *further* from ideally terminated. To see patterns in the data, I wanted to view the self-impedance seen at the transmitter with the thru-port 2 connected; this varies slightly from the Z-parameter definition but may be helpful for a first look.

### Data Across Fixed Antenna (C2) Setting

#### Observations

Sweeping the inductor across fixed C2 setting shows a shift in frequency with each inductor. The data appears to show B as the lowest frequency, and A as the highest; this might indicate the dial is one click “off” with respect to the inductor letter label.

At antenna setting of Tx10, a strong antiresonance appears for each of the inductor measurements.

### Data Across Fixed Transmitter (C1) Setting

#### Observations

Between Tx0 and Tx5 the resonant frequency is seen to increase. For transceiver setting Tx10, the inductors don’t matter much at all, and the impedance magnitude data is unitary in this frequency range.

## Data Across Fixed Inductor (L) Setting

#### Observations

Not much is concluded from the data- the plots are mostly differentiated by the sharp antiresonance that appears at high Antenna setting.

## Next Steps

Here I’ll outline the possible directions I see this analysis going. Send me some mail if you think I should look at the data in a different way. I’ll update with results in a future post (lol).

### Calculating C1, L, C2 From Data

Another way to view the measured data is by creating an equivalent circuit of the tuner based on the settings used. Using the known load impedance and the measured S11, S21 data, it should be possible to model an approximated tuner circuit with L and C elements. Although this circuit is best characterized with a full 2 port measurement, the T/R capabilities of the NanoVNA restricts the port 2 return loss data available without further measurements and enhanced calibration methods.

## Modeling as T-Parameter Matrix

Better would be to model the circuit using scatter transfer parameters (T-parameter) network (or, alternatively, as an ABCD matrix) [*Pozar, 4th ed. Ch 4.4*]. Using a T-parameter matrix makes sense to me as it can relate the incident wave to the transmitted wave under a constant load.

### Calculating S-Parameter Determinant from Existing Data

The measurements across settings can be spliced to replicate a two-port measurement. From this data, the determinant of the S-matrix can be calculated and used in the s-parameter to t-parameter conversion. This would approximate the return loss of both sides of the circuit, assuming similar properties for variable capacitors C1 and C2.

### Retake the Data with Flipped DUT

Another solution similar to above would retake the reverse of the T/R data taken previously to get (an approximation of, without respect to the port two load match) the port two return loss. This would allow for a view into the circuit from both sides, and is used with the forward return loss and insertion loss to identify the elements in the matrix.