AnsweredAssumed Answered

question on TRL standard design for differential signals

Question asked by snaffle on Jan 20, 2013
Latest reply on Jan 23, 2013 by Dr_joel
I am trying to determine if I should do a better job of de-embedding. The DUT is a low insertion loss switch that has -3dB point at 2GHz and I need to measure the switch insertion loss from 1MHz to 4.5GHz. 

The DUT is soldered to a PCB and the pins are brought out with microstrips to SMA connectors. I have been performing a normalization where I measure the S21 of a test microstrip, store this measurement in memory and then display my measurement traces as data divided by memory to remove the microstrip effect. Now I have an opportunity to redesign the PCB fixture boards and I want to build in a TRL calibration kit. But I have never done this. 

I have read the application note http://cp.literature.agilent.com/litweb/pdf/5968-5329E.pdf
It seems that I should lay out thru, reflect, line and match standards. I have some questions after reading the app note:

- How to design the match standard?

- The switch transmits differential signals so I do measurements in balanced mode. I display traces with the VNA in balanced mode. This converts 4-port measurements to balanced 2-port measurements. Do I need to create differential calibration standards and perform the TRL calibration while the VNA is in balanced mode? Or can I perform TRL calibration on all 4 ports separately?  

Outcomes