AnsweredAssumed Answered

8720ES TRL Cal, the good, the bad, the ugly

Question asked by mikejvir- on Sep 18, 2014
Hello All,

I am working on de-embedding a evaluation board using a 8720ES.  My frequency of interest is 0.3 to 10 GHz and I am mostly concerned in the K-factor of my amplifier (out of the operating band) over temperature.  I designed a three line TRL board.  One line goes from 0.3 to 2.7 GHz, the second goes from 2.2 to 17.8 GHz, and the third is my through.  The board material is FR4 (not my design choice).

The Good.
Thanks to the forum I was re-informed about the need for 4 receivers to do a TRL.  IF you only have 3 then  you need to pad the two ports.  I use 12 dB pads and my ripply was drastically reduced.

The Bad,
I have a discontinuity in the Return Loss and Gain at 2.7 GHz.    It is minor, but visible to the eye.  Do I assume this is a issue with tow line standards being different?   Is this  a difference in the line impedance?  Or something else.

The Ugly  Part 1
Above 9 GHz  the data is wacky (See file "Wacky TRL Screen.gif").  Though out my band of interest I am interested in determining what is wrong.

The Ugly  Part 2
Data at cold, -50 C, is really bad (see "TRL Issue.pdf").  It is seen mostly in the return losses and makes the K-factor go wild.  Luckily the laws of physics work and the data looks okay again after I bring it back to room temperature.   Is my problem FR4 or something else.

Thanks for your time, review, and comments.

Mike Virostko