Designing Test Fixture for S-parameter measurement

I am going to design a test fixture to measure the S-parameters and later de-embed in ADS. I have one question:

1. Does the electrical length of the lines in the fixture have to be less than 90 degrees?

Thanks in advance.

no, of course not. You should decide exactly what type of calibration method you will use with the network analyzer, and design your test fixture to accommodate it. You might need different length thrus, loads, short, opens, depending on what calibration method you expect to use.

You should strive to make your test fixture low loss, and with low parasitics

thanks for replying biff.

I usually use SOLT mechanical calibration kit.

I measure the S-parameters which contains both the response of DUT and the fixture. Later i de-embed by removing the fixture effects in ADS.

I follow this method: http://literature.agilent.com/litweb...980-2784EN.pdf

I have done this already. Now, I am trying to design a similar test fixture for another device. and some where I once read (in some forum, not in any book or paper) that the electrical length of the line in the fixture has to be less than 90 degrees, otherwise you end up with incorrect measurement. I could not find anything reference in the literature. That's why I wanted to know if someone knows from experience.