Electronics > RF, Microwave, Ham Radio

understanding VNA impedance plot? (inappropriate impedance test errors)

(1/13) > >>

coppercone2:
So I was experimenting with that VNA fixture project I was working on for the E5100A and I get a very wavy output


--- End quote ---


That is for a 0603 7.5kOhm resistor.

I did the through cal on a BNC-BNC connector, and then I hooked up the fixture.

I also soldered a 1/8 watt through hole 1Kohm resistor into a piece of brass tube with BNC connectors at both ends (so its just a series resistor between the two center pins of the connectors).. and it looks similar (I thought it had something to do with the pogo pin springs at first, but even without the fixture with a resistor in a tube it does similar. I am using a transmission reflection test set from HP. 

I thought maybe the problem is I need to do a proper calibration on it with a cal kit, that instrument has never seen a cal kit, I just have depended on the 'thru' method of calibration.

So it seems the resistance measurement is in the ball park but its very wavy. I don't think its the test set because I previously made this out of a minicircuits directional coupler and a 3 way power splitter and the results were similar looking.

When I tried to measure a tantalum SMD capacitor in series like that, to see ESR, I also got some wavy stuff (started at like 2 ohms and ended at 6 ohms).

Bud:
Beside that you must perform user calibration, VNAs are 50 Ohm devices and are not suitable for accurate high impedance measurements, they have large uncertainty outside of being close to 50 Ohm DUT  value.

coppercone2:
Hmm I know, but I was looking at that document that I wrote a thread about and it seemed that the series measurement method may be accurate on higher resistances.

Is the waveyness just a manifestation of the error? I tried several diff BNC-BNC connectors from diff manufacturers for the calibration and they all seem to agree with each other.

I guess I did not figure the error would be sinusoidal, instead that it would 'trail off' in one direction or the other. The graph in that VNA series/shunt/shuntthru methods of measurement kinda gave me the feeling that the magnitude of error would just increase in one direction, not be sinusoidal.

Now when I think about it, it does make sense, since the graph was error vs impedance not error vs freqeuncy, each impedance would need a separate graph to show error... I thought for some reason that the resistance/error graph would apply as error/frequency.

https://base.imgix.net/files/base/ebm/mwrf/image/2019/06/mwrf_com_sites_mwrf.com_files_21J_Fig3.png?auto=format&fit=max&w=1440

If anyone wants to post some inappropriate VNA impedance measurements that would be welcome... btw if you look at that on the dB scale, its around +-4dB

Bud:
Waviness would very much be because of impedance mismatch (causing reflections at both DUT ports) caused by the high impedance DUTs you are trying to measure. I'd say the practical limit for a VNA would be 1:10 SWR, i.e. 5...500 Ohm DUT impedance range.

Edit: i can't see the picture in your original post

coppercone2:
https://+i.imgur.com/ZY4HPOq.jpeg

Can you see that if you copy paste it and get rid of the + sign?

I will try to put like a 200 ohm resistor in the fixture and see what happens.

Navigation

[0] Message Index

[#] Next page

There was an error while thanking
Thanking...
Go to full version