AnsweredAssumed Answered

measurements using TRL cal have unexpected (larger) phase

Question asked by HiFreq on May 9, 2007
Latest reply on May 9, 2007 by dgun
I'm using the ICM fixture & calibration kit whose ICM
part number is A0138389, and I'm measuring SiC MESFETs
from CREE whose model number is CRF24010, in the flange

The S-parameters of this device, measured using the ICM
fixture and TRL standards, using an Agilent PNA, differ
significantly from those in the CREE data sheet.  While some
differences are to be expected due to variation from device
to device, and from different bias conditions, my measurements
are puzzling.  In brief, the phase of the measured S-parameters
changes more than would be expected as the frequency in increased.

One way of characterizing these differences is to make a small-signal
model of the device which is adjusted to reproduce the measured
S-parameters.  When I do this, the model seems to require the
addition of series transmission lines at the gate and drain of the
transistor.  I've made models for four transistors of the same type.
The average length of these added transmission lines is 6.3 millimeters,
and the average transmission line characteristic impedance is 28 ohms.

These results seem to suggest that the result of the TRL calibration
places the measurement reference planes at a significant distance
from the gate and drain terminals of the device.  Because the
impedance of these transmission lines differs from 50 ohms, this
distance cannot be removed by a using port extensions.  These results
are quite puzzling, because the dimensions of the TRL calibration standards
seem to conform precisely to Agilent's recommendations.