AnsweredAssumed Answered

Any disadvantage in having long offset delays on open/shorts ?

Question asked by drkirkby on Sep 28, 2013
Latest reply on Oct 15, 2013 by drkirkby
Is there any disadvantage in a cal kit having a long offset delay? 

I hit a problem this week, in that the open offset male (85052-60008) in my 85052B was damaged. As a *short-term fix*, before the new male offset open arrives, I decided to male standards by combining a male-male 3.5 mm adapter (85052-60014) along with female opens and shorts. The combination of a male-male adapter and a female standard, obviously givens one a male connector. 

I assumed the offset delays of the open was 29.2 ps and the short=31.8 ps. Those are the values reported at

although I gather from Brad in another forum post, there are some small rounding errors in those values. Anyway, as a short-term fix, I added those delays to the delay of the adapter (94.75 ps), to come up with a new cal kit, with offset delays of: 

short + adapter =94.75+31.8=126.55 ps 
open + adapter =94.75+29.2=123.95 ps

Those values are far in excess of any commercial cal kit I'm aware of. 

I then calibrated the VNA with "adapter+offset short" and "adapter + offset open", using those values. I made no correction for loss of the adapter. 

Much to my surprise, this seemed to work. Even at 20 GHz, the measured phases of the open and short were close to 180 degrees apart. At 20.00 GHz, the open had a phase of 155 deg, and the short -22 deg, so the difference was 177 degrees, which is not too far from the optimal 180 degrees, and well within the permissible range.  

This got me wondering, is there any particular disadvantage in using offset delays much longer than necessary? I realize the loss will be higher, but one could in principle correct for that, with a revised value of offset loss. 

I've not checked any of this on an airline, and for the actual measurements I only needed to work to 1.3 GHz. But the combination of male-male adapter, female standards, and a bit of basic arithmetic, I seemed to be able to get a calibration that looked ok. 

I expect a long adapter would have a less stable delay than a short one, but I would have expected the effects of that to be small compared to the changes that result from even the slightest cable movement. 


Edited by: drkirkby on Sep 28, 2013 6:30 PM