AnsweredAssumed Answered

Custom calibration board on HP8753ES

Question asked by wiwyam on Nov 9, 2018
Latest reply on Nov 12, 2018 by Dr_joel

Like • Show 0 Likes0
Comment • 3

Hi,

We built a custom calibration kit to calibrate out our RF switch and cable. Below is a picture of it and the VNA response to "open" post calibration.

My problem is that there seems to be a large error in my measurements. For example I am measuring the impedance of a dev-kit which is supposed to be 50ohms, but my VNA shows 20 + 10j; really far off.

I think the issue is my "calkit" settings, for this custom calibration board I must define the calkit. I don't need this to be extremely accurate, in fact all I want is for the VNA to NOT correct for any calkit parasitics. I'm sure the calkit parasitics are way less than the parasitics I'm trying to match out - my matching network will be an order of magnitude larger than any calkit parasitics.

I'm struggling to get my "open" impedance to have no imaginary component when connected to the VNA, can you offer any advice?

Looking at my calibration board, can you offer rough offset delay,loss and impedance values?

Thanks a lot, Will

Attachments

IMG_0554.JPG420.2 KB

No one else has this question

Outcomes

wiwyam Nov 9, 2018 12:56 PM

I found this blog which explains how to use custom cal kits, filling in a lot of blanks left in the user manual

Setup an HP or Agilent 8753 VNA to use a calibration kit

But my post-calibration inspection of a 50ohm impedance is still shifted to the top left of the smith chart. Which calibration coefficients are incorrect to result in a shift like this?

Actions

Dr_joel

Nov 12, 2018 3:14 PM

What you show on the display looks like an open circuit response. Remember opens and shorts in the 8753 cal kit are almost always "offset opens" and "offset shorts".

If you look at chapter 9 in my book I walk through the details of creating a PCB calibration kit and how to load the values.

Show a plot of the 50 ohm state and I'll have a better idea. At 2.4 GHz, you have to be careful that the test cable doesn't flex so much it changes phase enough to make the calibration invalid.

Actions

Dr_joel

Nov 12, 2018 3:27 PM

One more thing: You might be confused by the Z Smith Chart. This is showing that your open is a series connection of 98 femptoFarads and 9 ohms. If you switch to an inverted Smith Chart, you will see it has a very large shunt resistance and the proper portion of the shunt capacitance, will be something like 10 uSiemens (1/ohms, so 100k ohm) and 100 fF (or .1 pF shunt). The reason the open is not exactly at zero is the definition of the open has a built-in offset delay (differs depending upon connector type).

Actions

3 Replies

wiwyam Nov 9, 2018 12:56 PM
Mark Correct
Correct Answer
I found this blog which explains how to use custom cal kits, filling in a lot of blanks left in the user manual
Setup an HP or Agilent 8753 VNA to use a calibration kit
But my post-calibration inspection of a 50ohm impedance is still shifted to the top left of the smith chart. Which calibration coefficients are incorrect to result in a shift like this?
Like • Show 0 Likes0
Actions
Dr_joel Nov 12, 2018 3:14 PM
Mark Correct
Correct Answer
What you show on the display looks like an open circuit response. Remember opens and shorts in the 8753 cal kit are almost always "offset opens" and "offset shorts".

If you look at chapter 9 in my book I walk through the details of creating a PCB calibration kit and how to load the values.

Show a plot of the 50 ohm state and I'll have a better idea. At 2.4 GHz, you have to be careful that the test cable doesn't flex so much it changes phase enough to make the calibration invalid.
Like • Show 0 Likes0
Actions
Dr_joel Nov 12, 2018 3:27 PM
Mark Correct
Correct Answer
One more thing: You might be confused by the Z Smith Chart. This is showing that your open is a series connection of 98 femptoFarads and 9 ohms. If you switch to an inverted Smith Chart, you will see it has a very large shunt resistance and the proper portion of the shunt capacitance, will be something like 10 uSiemens (1/ohms, so 100k ohm) and 100 fF (or .1 pF shunt). The reason the open is not exactly at zero is the definition of the open has a built-in offset delay (differs depending upon connector type).
Like • Show 0 Likes0
Actions

Custom calibration board on HP8753ES

Attachments

Outcomes

Related Content

Recommended Content