I am currently working on defining a .CKT file with calibration coefficients for a 75-Ohm MCX SOLT cal kit. One of the things I am looking at while calibrating is the open-circuit return loss ripple at the calibrated port. When I set the Z0 impedance of the open and short standards in the calibration kit to 75 Ohms, there is roughly 2dB of ripple when I calibrate. However, when I set the Z0 of the opens and shorts to 50 Ohms, the ripple is less than 0.2dB, which is much closer to what I would expect for a good calibration.
I am working on a N5225A PNA and using a 12GHz 11852B (option H12) min loss pad to transform the impedance from 50 ohms to 75 Ohms, followed by a 75-Ohm type N to MCX female adapter and calibrated using a set of custom MCX male standards.Testing is being done from 10MHz to 6GHz.
Changing the system Z0 in the network analyzer has no effect, and this open circuit ripple is repeatable across multiple calibrations. We believe that the standards are working correctly mechanically, as measuring a MCX male-to-female adapter and termination produces results that roughly agree with measurements taken using a 12GHz 85036B 75-Ohm Type N kit with TDR gating. Both Z0 settings seem to produce comparable measurement results, and the issue only seems to be with the open-circuit ripple. The values of the correction coefficients do not seem to be at fault, either, as inputting zeroes for all the coefficients and leaving only the delays produces the same result.
Example of male to female adapter 50 vs. 75 Ohm cal standard Z0:
I am curious to know what might be causing this issue, and how I can fix my .ckt file to work properly with 75 Ohms as the Z0 of the standards. I have worked on other 50-Ohm calibration kits and had no issues with this. The issue seems to perhaps be with something in how system impedances are being handled by the correction in network analyzer.