Currently, we do a THRU-only calibration and then execute a sweep with correction on to measure S21 over a given span. We were wondering if it would be possible to somehow do the calibration correction afterwards, and if so, how to calculate the interpolated calibration coefficients.

1) The center and span of the measurement sweep will differ from the center and span of the calibration sweep, and the measurement sweep is done with fewer points. How is the calibration interpolation done in cases like these? Is it simply the case that one linearly extrapolates between known values of Ett?

For example, if we know the Real and Imaginary values of Ett at 10.1MHz are 0.96 and 0.11 and the Real and Imaginary values of Ett at 10.2MHz are 0.97 and 0.10, would we simply calculate the Real and Imaginary values of Ett at 10.13MHz to be 0.963 and 0.107? Or are the interpolated calibration coefficients calculated in some other manner? Additionally, does a change in the IFBW value between the calibration sweep and the measurement sweep have any effect (I wouldn’t think so, but I just wanted to be sure)?

2) And one more general question: are there any reasons why the calibration correction should not be done in this manner?

Thanks, and I appreciate any input any of you would have about this. If there’s anything I’m missing or should be aware of, please let me know.

Edited by: Cheenu on Jan 7, 2014 3:36 PM

1) The center and span of the measurement sweep will differ from the center and span of the calibration sweep, and the measurement sweep is done with fewer points. How is the calibration interpolation done in cases like these? Is it simply the case that one linearly extrapolates between known values of Ett?

For example, if we know the Real and Imaginary values of Ett at 10.1MHz are 0.96 and 0.11 and the Real and Imaginary values of Ett at 10.2MHz are 0.97 and 0.10, would we simply calculate the Real and Imaginary values of Ett at 10.13MHz to be 0.963 and 0.107? Or are the interpolated calibration coefficients calculated in some other manner? Additionally, does a change in the IFBW value between the calibration sweep and the measurement sweep have any effect (I wouldn’t think so, but I just wanted to be sure)?

2) And one more general question: are there any reasons why the calibration correction should not be done in this manner?

Thanks, and I appreciate any input any of you would have about this. If there’s anything I’m missing or should be aware of, please let me know.

Edited by: Cheenu on Jan 7, 2014 3:36 PM

So go to your widest span, set the point count to 1601, do a thru response calibration. Change to your narrow span with whatever point count you like. Caution, if you reset by first changing center and then changing span, you might set a condition that has the start or stop out of the calibrated frequency range, and so cal will turn off.