15m long RG-213 with a notch at 450 MHz

[cdev]

One might even make the argument that toroids, being on the outside, should not effect the impedance at all, unless the shielding was dysfunctional. (in some way)

I actually have some clip on toroids that I leave on my nanovna2's test leads all the time, to decouple the leads from the shields. I dont see any reason to not do that, and several reasons to do it.

I mostly do it because of high RF noise in my area.

Sometimes I squeeze coax between my fingers (not too hard) i order to "see" if I can locate where a fault is by the changed point. But that doesn't always work.

[Paul B]

hi just a thought try changing the BNC connectors for TNC this should be an improvement and eliminate
any  bad joint  then repeat experiment without any BNC connectors in the CCT

[biastee]

I don't have any TNC suitable for RG-213; i.e. OD10 mm. I have fitted the same model of BNC connectors to other coax without any issue.

[Paul B]

The fault you have is strange but If you do not have TNC you could try N  The reason to change from BNC is 2 fold first from experience BNC connectors are not good at 1 Ghz when the joint is flexed  / moved  this was found out on the in and out of a inter digital filter changing just the outer-cases to TNC fixed the problem
the second is that there are a lot of very cheap BNC connectors and i have herd form others that the insulation can be suspect from memery of the picture assume you have not used screw on connectors Paul 

[biastee]

I don't want to go to the trouble of changing to another connector type, especially one which I don't have a cal-kit for, when the available evidences do NOT point to the BNC connector:
1. I have other cables using the same BNC connector model without issue, including the one which I posted in reply #8
2. Telegartner has been selling this BNC connector model for decades and any problem would have surfaced by now


Thanks to all for the responses!

[Joel_Dunsmore]

I don't want to go to the trouble of changing to another connector type, especially one which I don't have a cal-kit for, when the available evidences do NOT point to the BNC connector:
1. I have other cables using the same BNC connector model without issue, including the one which I posted in reply #8
2. Telegartner has been selling this BNC connector model for decades and any problem would have surfaced by now

Unless it is a one-off problem with your particular connector, but I doubt it. The Return-Loss plot should show a large bump where the Transmission has a dip if it was reflective and at the connector end. Could still be a bad spot causing reflection in the middle of the cable, hidden by the loss of the cable from the S11 or S22 plot, but more likely something lossy in the cable or dielectric itself.