AnsweredAssumed Answered

Highly Reflective Passive Device

Question asked by chard on Jul 16, 2009
Latest reply on Aug 10, 2009 by chard
I'm using an E5071B with an Agilent 85033E calibration kit. I've done a 1-port calibration in the range 300kHz-8.5Ghz--the entire range of the VNA. 

The DUT is passive but highly reflective--it's a metallization pad with via holes to ground. This pad is fed by a short, 400mil length feed line and then connectorized with an SMA-to-Microstrip transition. My VNA calibration plane is at the SMA.

The issue I'm seeing is that the measured reflection coefficient (RC) rises slightly above unity (1.02) over specific narrow ranges of RC phase. These ranges are encountered at higher frequencies---in this case, it just happens at the upper frequency 8.5Ghz.  But I tried it with a longer length of feed line and can shift the frequency down where the |RC|>1.

I was expecting a spiraling locus of points inward toward the center of the chart due to increasing line loss versus frequency.

I'm guessing this is some sort of VNA directivity issue.   My question is whether this phenomena is expected for this combination VNA and Cal Kit. If so, what enhancements can be done to improve the accuracy?