0
\$\begingroup\$

I am trying to do a one port VNA calibration for measuring filters at the GHz band. The VNA that I am using is Keysight E5063A. However, I am running into some conceptual as well as some practical problems. Any help will be greatly appreciated.

I do not have the calibration kit provided by Keysight. However, I am using a standard open, short, and load (50 Ohm) that I found in the lab I work in.

Details:

Open is the standard BNC Open. Short is the standard SMA short. Load is 50 Ohm mini circuits, operating range DC - 18 GHz, Model number: 30918. The band at which I am calibrating is 4 GHz to 6GHz.

Procedure:

I connect a standard ST18 SMA, 36 Inch cable to port 1 of the network analyzer. The purpose of this is to reduce mating cycles. Going into the calibration menu, select one port calibration. I connect BNC Open, and press open calibrate. I connect Short, and press short calibrate. I connect load, and press load calibrate, click done, and I save this calibration.

However, now when I see the results on the smith chart on the network analyzer, I do not see a dot on the edge of the smith chart for short and open. Instead, I see an arc. For the load, I see a dot. I do see open and shorts maintaining a 180 degrees phase separation throughout the 4 GHz to 6 GHz band.

Questions:

Why is the arc caused in the first place? Is it because of fringing capacitance (inductance) for open (short)?

Can I go ahead and make measurements of filters with this calibration?

I am worried that because of this, the complicated circuit network that I am trying to measure, may not reflect its right characteristics due to calibration errors, if there are any. In such a circumstance, what can I do, to know for sure that what I am measuring is exactly what it is at this band (and there are minimal systematic errors.)

\$\endgroup\$
  • \$\begingroup\$ A Smith chart plot might help us see and understand the magnitude of the errors you're asking about. \$\endgroup\$ – The Photon Mar 10 at 22:20
0
\$\begingroup\$

Why is the arc caused in the first place? Is it because of fringing capacitance (inductance) for open (short)?

It's because the cal standards you used don't match the expected behavior of a nearly ideal open, short, and load exactly.

Most likely either the open or the short has a small amount of distance between the SMA reference plane and the effective position of the reflection.

It may even be that the standard short (for example) has a small, but known and accounted for, delay between the reference plane and the actual reflection plane, but the short you cal'ed with doesn't, leading to a negative delay in your "calibrated" measurements.

Another issue could be that you said you're using an SMA cable, but told the VNA you were using BNC cal standards. The BNC cal standards might be significantly different from the SMA devices you used (for example the definition of where the reference plane is in a BNC connector is likely noticeably different from where the reference plane is in an SMA connector), and this is causing the calibration to be inaccurate.

Another issue could be that BNC connectors are usually only characterized for up to 1 GHz, while you were calibrating for 4-6 GHz. If the models for the BNC standards are only accurate up to 1 GHz and the machine extrapolated the performance out to 6 GHz, it could have wildly inaccurate model for how the standards (either BNC standards or the SMA devices you used) behave in the range you measured.

So I'd suggest:

  • If possible, use the proper SMA cal kit to calibrate an SMA connection.

  • If not possible, at least use the SMA cal kit model when cal'ing with SMA devices. But expect a few percent (of the Smith chart unit circle) error in the final measurements.

| improve this answer | |
\$\endgroup\$
0
\$\begingroup\$

An open/short/load calibration synthesises an accurate network analyser at the location of the reference plane. If you do not have the calibration standards that the VNA knows about, then you must enter the electrical characteristics (offset length for short, offset length and radiation factor for open) of the ones you are using. Then take the measurement using any cables you used during calibration.

| improve this answer | |
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.