The S11 method.The single port adaptor efficiency evaluation (SPAEE) was developed at the National Institute of Standards and Technology (NIST) around year 1990. The method is described here: www.nist.gov/calibrations/upload/arftg-95.pdf The method has been used by the NIST Noise Calibration Service to measure noise sources for which primary standards or radiometers were not avilable.
In this paper from 1999 Comparison of Adapter Characterization Methods J. Randa, Wojciech Wiatr, Robert L. Billinger compare three different methods for the characterization of precision adapters. In this study the efficiencies are reported to agree within about 0.005 for efficiencies near one.
For a typical phase or impedance shifter on 1296 MHz the loss is in the order of 0.03 dB which means that the efficiency is in the order of 0.993 which is very close to one. An uncertainty of 0.002 which is the lower limit reported would mean that the efficiency would be between 0.991 and 0.995 or between 0.0218 and 0.0393 dB.
The test objects.The test objects, DUTs,are presented here. There are three more DUTs:
DUTA A 0.25 wl cable
DUTB A cable that transforms from 50 to 70 ohms.
DUTC A 0.125 wl cable.
Raw data.Table 1 gives raw data, return loss, for the DUTs when the calibration kit "open" and "short" are used with a DUT inserted. The average of the losses in both cases gives the loss for a signal that passes through the DUT twice so the loss is the average divided by two. The network analyzer is calibrated to show return loss zero when the calibration kit open or short is connected to the test port. (That means that the instrument is configured for lossless "open" and "short".)
Dev RL Open RL Short Loss DUTA 0.052 0.087 0.03475 DUTB 0.064 0.069 0.03325 DUTC 0.032 0.100 0.033 DUTA+DUTC 0.169 0.085 0.0635 DUT1 0.067 0.075 0.0355 DUT3 0.023 0.075 0.0245 DUT8 0.070 0.066 0.034
|Table 1.Return losses with open and short and the dissipative loss of the DUT from measurements by SM0ERR using the S11 method.|
The sum of the losses of DUTA and DUTC does not quite agree with
the loss measured on the two connected together.
The difference is 0.00425 dB which is in line with the uncertainties
reported in the NIST paper.