VNA setup for S-parameter measurement

[Zebxcore]

Hello guys,

We have an HP/Agilent 8753D, with what I am assuming is incomplete options and fixtures, that I want to use for 2-port testing of equipment as well as S-parameter characterization. So far, I've been able to test 2-port systems to a certain degree (to which I don't know if it's right really) by coupling the ports as shown in the picture.

Ports are, from left to right: RF out (50Ohm), R, A, and B. All 50Ohm.

On coupler: 1) output? 2) R + B , 3) RF Out

However, when it comes to S11 of an antenna, for example, all I see is a sweep of what I assume is the resistance out? I would couple the antenna to the output of the coupler, being port 1.

I've looked up the VNA's manual and videos everywhere online and it always comes with an extra component on the bottom to which you plug all four ports in and makes the VNA a true 2-port device. I don't have that luxury here though. I have never used a 2-port VNA like this before so I am kind of in a loss as to how to proceed.

I am wondering if anyone has any experience with this VNA that could help me figure out how I can plug it to get the measurements needed without the fixture.

Thanks!

 

[Zebxcore]

Clearing my ignorance some more, I found that the bottom block that I am missing is a solid-sate switch referenced as 85046A/B. Also, this VNA would be option 011, which still doesn't answer my question as to how I can use the ports since the manual shows the usage of the VNA with the test fixture on.

Any light on the issue is appreciated!

 

[D.A.(Tony)Stewart]

Your photo looks incorrect for setup. You better show a block diagram

 

[Zebxcore]

Here is a rough block diagram of what the setup looks now. To me, it seems like RF out should be combined with R while B is coupled to their sum. Still, the VNA can't seem to operate near one with the test fixture.

Looking at the manual of option 011, they show a lot of testing procedures, mostly mixer style, that could be useful later on. However, I couldn't find any description of S-parameter, or even just a simple frequency response, without the need for the test fixture.

 

[D.A.(Tony)Stewart]

This is my thinking.

You need to have a Short-Open-50R calibration process at least when DUT is out.
I suggest a 4 port 2-way Directional Coupler from the Generator out.
Forward Splitter out goes to R (reference)
Reverse Splitter (combiner) goes to A to measure reflected signals
Forward out goes to DUT to B.
Attenuation for B may be needed if the DUT has high gain.

https://www.vadiodes.com/VDI/pdf/2010MarNews.pdf

 

[Zebxcore]

Tony,

I didn't have a 4 port directional coupler, but I tried the setup with the 3 port coupler I have in the pictures and it worked! To think that I was just beating my head against the desk trying to think of how many more components I would need to replicate the S-parameter test fixture but just a coupler and a combiner do the trick.

Calibration now seems required since before we were just doing simple response tests. Now I will be able to do full 2-port tests.

Thanks a lot for the help!

 

[D.A.(Tony)Stewart]

Tony,

I didn't have a 4 port directional coupler, but I tried the setup with the 3 port coupler I have in the pictures and it worked! To think that I was just beating my head against the desk trying to think of how many more components I would need to replicate the S-parameter test fixture but just a coupler and a combiner do the trick.

Calibration now seems required since before we were just doing simple response tests. Now I will be able to do full 2-port tests.

Thanks a lot for the help!

Yer welcome.
If you have 2 good 3 dB splitters, that might also work well back-back. -< >-

 

[FvM]

An Agilent/Keysight document shows the below schematic for the 8753 S-parameter setup. It uses in fact 3-port couplers and derives incident wave with a simple power divider.

My 8714 with built-in S-parameter fixture is using a directional coupler for incident wave, I guess resulting in higher accuracy or at least lower calibration effort.