Measurement of RFID antenna coupling with a coaxial testset

Question asked by maikhermanns on May 8, 2007
Hello,
At the moment I'm writing my diploma thesis on a RFID topic.
To confirm my simulations I have to measure the coupling between the RFID reader antenna and the tag antenna coil.
I already made some measurements, varying the coil distance in defined steps using my HP8753C and the85046 testset. The reader antenna is basically a square single turn loop of 250mm x 500mm. The tag antenna, connected to Port 2 is a coil and measures about 50mm in diameter.
I assume that the RFID link ist just magnetic coupled and use the "T" eqvivalent circuit to model the magnetic coupling.

Measurement:
I'm measuring all 4 S-Parameters from a single frequency measurement at 13.56MHz, store them on the PC, convert them into Z-Parameters and calculate the transformer coupling coefficient k (M/sqrt(L1*L2))
I've made a full 2 Port SOLT calibration, DUTs connected via Sucoflex 104 SMA cables, Pout=10dBm, CW Frequency 13.56Mhz.

Problem:
If I change the winding direction on one of the coils, my calculated coupling coefficient changes. It is, dependig on the antenna geometry, app. 20% greater or smaller than the simulated or mathematical calculated value.

If the measurement frequency is reduced to 500kHz these effects disappear, and the coupling coefficient calculated from the measured S-Parameters are as big as the simulated coupling coefficients.

I think the problem is the feed of the symmetric antennas with the asymmetric coaxial analyzer ports. I've already wound 3 turns of the test Port cable trough a ferrite core to test that, but no change at all.
An alternative would be other ground coupling problems.

In the measurements S12 is S21, if I change the winding direction, the absolute value of the reflection coefficient changes about app. 10% (not expected), the phase changes approx. 180degrees (as expected).

Are there any suggestions how to eleminate these problems or is the source of these effects well known.
Would it be a good approach to measure through two identical baluns (any cheap manufacturer / type suggestions ?)and how can the analyzer be calibrated.

regards
Maik Hermanns