Vector Network Analyzer - Transformer Turn Ratio

Thread Starter


Joined Apr 24, 2017
Dear All,
it is my first post here on AAC, I hope I am not infringing any forum rule :)

I am trying to measure by an Vector Network Analyzer (Agilent n9914) the turn ratio of an air-core auto transformer in a frequency range 30kH - 30MHz, since I am looking for some series resonance and voltage overshoot (it may sounds weird but it is an exercise I am doing for another purpose).

The schematic is attached.

I am comparing the value I computed with transmission line model (see first data series in "result") with measured data.

I performed a first comparison applying V1 by a signal generator and measuring the gain V2/V1 by an oscilloscope, see data series "Meas. Oscill" in the same picture. The resonance was confirmed.

Next, I wanted to perform this tedious task by a network analyzer, to have a nice and smooth characteristic swept over the whole frequency range of interest.

The idea is computing the scattering matrix of the auto transformer, seen as a double port element. After performing the matrix, I converted it to the ABCD matrix, since V2/V1 is equivalent to 1/A. The result that I get is the dark orange data-series in the picture "results".
The trend is followed at higher frequency (I get a peak and a valley), despite the agreement is poor (but I already saw by the oscilloscope that I compute the overvoltage well). I am doubting of the results I am getting from the oscilloscope, since the 1/A increases at lower frequency, which does not make sense at all (increasing turn ratio at low frequency?? I do not expect any resonance there).

Do you have any comment, suggestions?
Of course I performed the calibration of the network analyzer before performing my test. Besides, I using the same COAX connection both for the oscilloscope and the network analyzer measurements, so the coil terminals should not matter.