Programming a RF transmission and reflection test for an antenna

[Brian@Channel]

I'm creating a system for characterizing the frequency response of an antenna pair. I want to sweep the TX antenna over a fixed fz range and measure the reflected power (due to the impedance mismatch) and the power received by the RX antenna. I am generating the sweep signal using a PXI-5610 and 5421(AWG).  Let's just consider the frequency response measured by the antenna pair in order to keep things simple. So I have a device under test (DUT) connected between my RFSG and my RFSA. I use the Frequency_sweep.vi provided in the examples to produce a sweep of 100 points over a range of 400 MHz to 1000 MHz. The questions are:

1. Do I implement the measurement within the loop at each fz point or do I set a continuous sweep and then trigger the 5660 to measure at the same time?

2. Do I use a technique for reading the spectral data such as the example - ni5660_Getting_Started_Spectrum.vi ?

The end result is a graph of power versus frequency.

Can you recommend any other example programs to get me up and running? Thanks!

[Andy_Hinde]

Hello Brian,
Let me see if I can answer your questions:

1. Do I implement the measurement within the loop at each fz point or do I set a continuous sweep and then trigger the 5660 to measure at the same time?

Assuming each loop iteration corresponds to a new measurement at a new frequency, your loop should tune the generator and analyzer to the new frequency of interest, acquire a power spectrum of the acquired data, and then perform a peak search on the acquired power spectrum (assuming you are generating a CW sine wave signal).

2. Do I use a technique for reading the spectral data such as the example - ni5660_Getting_Started_Spectrum.vi ?
Yes - the ni5660 VIs make it easy to configure and acquire both spectral data and IQ data. In this case, we are acquiring power spectra and performaing peak search measurements upon the spectrum to measure the tone's peak power value.

I am attaching an example which combines the PXI-5660 and PXI-5670 hardware to accomplish this task.

Regards,
Andy Hinde
National Instruments

[Andy_Hinde]

My attachment didn't quite make it. Here's another try....

Andy

[Brian@Channel]

Hi Andy,

Thank you very much for the example code. I tested it with an attenuator connected between my RFSG and RFSA and it produces the power vs fz spectral curve that I am after. When doing this type of frequency response test, is this technique the best way? ie do we produce a new fz point and take a new measurement at each loop iteration? The other technique would be to sweep at some rate (say 10ms) and trigger the RFSA to measure at the same time. I intend to sweep continuously to monitor the DUT during an experiment. The impedance of the DUT changes and we wish to log the frequency response over time. I will also be measuring the reflected power from the load and will use a PXI-2590 multiplexer to feed the two signals into the RFSA. Any advice with respect to this is greatly appreciated.

[Andy_Hinde]

Generating tones at frequency points and measuring the peak power values is the best way I know of doinf this measurement. I am not sure what way you are referring to by 'sweeping' the PXI-5660, as it is not a swpet-tuned narrowband spectrum analyzer. The PXI-5660 is a wideband RF Vector Signal Analyzer that can also calculate zoom FFTs, so you are not 'sweeping' a narrowband Resolution Bandwidth (RBW) filter. Rather, you are calculating a FFT/power spectrum from acquired time-domain data.

Hope this helps!

Andy

[em_uwave_ee]

Just a thought about the frequency response measurement and calibration.
This system has the same requirements as a magnitude only reflectometer or network analyzer.  In order to take out the system variations a set of "calibration data" should be collected and used to "subtract out" the effect of cables, connectors etc.   To get a good |s21| measurement insert a well characterized cable as a "through" standard.   This magnitude can be added back in after the two antennas are inserted.  The standard does not have to be loss-less.  In fact if an attenuator is used at each end of the cable the return loss of the cable would be improved. The total attenuation of the "through" standard should be about the same as the raw |s21| measured with the antennas present.
To get a good |s11| you need to make two measurments.  One with a "short" standard the and another with the port left "open".   The RMS value of the two gives an approximate "return" loss of the system.  This system return loss (db) can be removed from the antenna measurement and give you a much better view of what your antena is doing.
regards.
.chuck

[Brian@Channel]

Chuck,

That's good advice! I do have a short and an open to perform the s11 measurements. I have a couple of questions...

To perform the s21 measurement, do I create a short, through cable or use a BNC to BNC adapter? I have been doing this with an attenuator (HP 8495A) in place of my antennae. Do I include the transmission reflection test set? This causes a typical attenuation of 12dBm. Again for the open and short, do I connect them at the end of my cable run, where I normally connect to my antenna?

Do you know of any LabVIEW examples or other reference material that explains how these calibration measurements can be implemented? Thanks very much.

[em_uwave_ee]

Hi,
Glad to help.
The direct answer to your question is Yes you do want to include the reflectometer (coupler and power meters or what ever).  the place you use the short and open are at the plane where the antenna would be attached.  I think you have some cable to go from your test set to the antenna.  Just unplug the antenna and put in the short and open. 
Using the attenuator in place of the antennas is perfect.   Just unplug the transmitt antenna and put it into the attenuator.  Then unplug the Receive antenna and plug the cable into the other end of the attenuator. You know the |s21| to add to your antenna gain computation.  It may be that you only care about the relative or gain flattness.  Then the magnitude of the offset is unimportant.  It is important that you do the calibration with the attenuation because the receiver's / power meter will not be exactly flat with respect to incident power level.
I dont have any LabView stuff but I'll check for a more indepth application note.  btw are you at channel microwave in camarillo?
.chuck

[Brian@Channel]

Thanks again. I really appreciate the opportunity to bounce ideas off someone else.

I located at Channel Systems in Canada (aka "way up north"), about due north of the North Dakota/Minnesota border.

[_Andrey_]

Hi,

I found this 5660 Acquire IQ and Spectrum example (http://zone.ni.com/devzone/cda/epd/p/id/5140) and I was wondering if it is possible to make it work somehow with v7.1?

- Andrey