AnsweredAssumed Answered

Problem with S11 S22 on long cable

Question asked by begleysm on Jan 13, 2009
Latest reply on Jan 13, 2009 by jvall

I am fairly new to the use of Network Analyzers.

My goal is to characterize several 125' Heliax cables with male precision-N connectors.  I have attached a screenshot.

In using the Network Analyzer I have found that the sweep time needed to be increased (from the initial setting of 60 msecs to about 180 msecs) for a good S12 & S21 measurement.  I have decided that these measurements are good because they fall within the spec given for the cable on the cable's datasheet and they roughly match the (uncalibrated) sanity check I did by connecting a signal generator to a spectrum analyzer with the cable and driving a CW signal through at 0dBm from 2 to 12 GHz with a 100 MHz step.

The S11 and S22 measurements still seem to be off.  As you can see in the attached screenshot the S11 and S22 have fairly high levels of reflection and, you may notice the dip at around 9GHz in S22.  I took the cable and flipped it around so that, end 2 is now end 1 and visa-versa.  This should mean that the S22 (green) and S11 (yellow plots) should essentially "change color."  In fact they stayed very much the same.  The S22 plot still had the dip at 9GHz while the S11 plot remained noisy, but somewhat flat.  This is contradictory to what I believe should have happened since I, physically, swapped the ends of the cable that the network analyzer saw.

I am looking for general help with this issue.  One consideration that I have not been able to find any information on relates to the length of the cable.  Is there a problem, or special procedures that must be undertaken, when testing "long" cables?  Could there be reflection issues with long cables that do not have time to appear on short cables.  I know that the long cables caused many problems with the S12 and S21 parameters until I increased the sweep time 3x.  This was due, I believe, to the RF energy not having sufficient time to make it down the cable before the next test was executed.

Thanks in advance,