AnsweredAssumed Answered

Open circuit cable shows S11 being +7 dB. Any ideas what could be wrong?

Question asked by drkirkby on Nov 10, 2012
Latest reply on Nov 23, 2012 by drkirkby
I'm getting problems trying to calibrate my VNA. After doing a full 2-port cal, if I leave port 1 open, I read an S11 of about + 7dB. The setup is this: 

* HP 8720D 20 GHz VNA with 3.5 mm connectors
* Agilent  85131F 3.5 mm  test port cables. 
* Agilent  N -> 3.5 mm adapters on the end of the test port cables. One connector is male N, the other female N. 
* HP 85032B calibration kit (this is N, rated to 6 GHz). 

After doing a full 2 port cal, putting a termination on port 1 shows an S11 of about -15 dB. Removing the load, the S11 is peaking at +7 dB at 4 GHz (see diagram). 

If I:

* Remove the  N -> 3.5 mm adapters and use an Agilent 85033E 3,5 mm cal kit, there is no problem. All results seem sensible in 3.5 mm. 

* Do a full two-port call in 3.5 mm, then add N -> 3.5 mm adapters with N loads, the return loss is about 30 dB (from memory). This would suggest the N adapters and N loads are not totally screwed up. 

* Measure DC resistance of 909F N loads, and find they are 50 Ohms. 

But doing a two port cal with my N cal kit is giving stupid results. 

Could an incorrectly defined user calibration kit cause this? The 85032B calibration kit is not in the firmware of my 8720D. Since I don't have a Windows computer with a GP-IB board in it, I can't use the VNA Cal Kit Manager software at http://na.tm.agilent.com/vnahelp/products.html to upload a cal kit to the VNA. Therefore I have entered the Standard Class Assignments and Standard Definitions manually using the buttons on the VNA. 

I would add my cal kit does have the extender for the female N open - I think most people manage to lose these, but my kit does have it. 

I was thinking of running a test not using the user defined cal, but the N calibration kit in firmware. But that is for an 85032F, not the older 85032B. I don't know how much difference there is between those kits and so what effect there would be in using the wrong kit. Would it cause huge errors like +7 dB gain on an open circuit?  

Any suggestions how to debug this? 

Dave

Edited by: drkirkby on Nov 10, 2012 11:28 AM  

Attachments

Outcomes