aaron_do
|
Hi mixed_signal,
since you say its a 1-port cal, I guess you aren't experiencing the problem I was referring to. I just re-read your post and see that you mentioned it was a 1-port cal...
The first several times I did my VNA cal, I would do a short, open, and load on each port, and then I would use the thru connector (some people spell it this way...not sure why) for the last step in the cal since both ports had a male connector at the end. I originally thought that the thru connector was specific to the cal kit, and that its data was stored in the VNA. Turns out I was wrong, and you need to calibrate out (de-embed might have been the wrong word) the thru connector as a final step. If you don't, as I remember, you will get some ripple in the frequency response.
Seems like rfidea was saying that your cal kit isn't perfect, but that the VNA knows the imperfection. I no longer have access to the VNA I was using, but I remember I could get a very good calibration as at the time we were looking at loss on the order of just points of a dB. Perhaps your VNA itself needs to be calibrated, or the cal kit is no longer valid (you have to be careful with those cal kits).
Aaron
|