Question about TRL standards for calibration

[chunweimin]

lambda over 4 trl calibration

Hi,

I am trying to use microstrip lines structures to build up TRL standards for calibration from 1GHz to 40GHz. (1) Is there anything I should be careful about the length of the "Thru" standard for wideband measurement? (2) I am wondering that for the "Reflect" standard, is it necessary to compensate the open end effect which is from the fringing fields of the lines? If so, since this additional length is frequency-dependent, then how should I manage that for this wideband measurement. (3) How should I appropriately pick the lengths of the "Line" standard for different ranges of frequencies? Thank you.:|

Cheers,

Chunwei

 

[Mazz]

TRL Calibration

If I well remember (I made it few years ago):
Thru is zero length (although it is non zero) so no special requirements
Reflect is better to do with a short, for the reason you mention using an open.
Line should be lambda/4 in the middle of your frequency range, and you have to use (theorically) a different line for each octave you need to include into calibration.
So you will need 1-2GHz, 2-4 GHz, 4-8 Ghz 8-16 GHz, 16-32 GHz.... TRL is inherently not a wideband cal, but is very accurate.
A common practice is to add a match load for low freq range.
I hope it can help.
Mazz