Tektronix Technical Forums are maintained by community involvement. Feel free to post questions or respond to questions by other members. Should you require a time-sensitive answer, please contact your local Tektronix support center here.

RSA5106B Trace Averaging

Post Reply
rcervantes
Posts: 3
Joined: June 19th, 2017, 6:15 pm
Country: United States

RSA5106B Trace Averaging

Post by rcervantes » June 20th, 2017, 9:37 am

Hi,
What type of trace averaging does the RSA5106B perform?

In my case, the averaged trace is always lower than the instantaneous trace. I expect that the averaged trace would reside in the middle of the fluctuations of the instantaneous trace.

To test the RSA, we injected 1,120 MHz tone directly into the RF Input using a Agilent N5183B MXG Signal Generator and measured the spectrum. We saved both the instantaneous trace and another trace that averaged for about 60 seconds. We find that the averaged trace is almost always 10 dBm lower than the instantaneous trace, regardless of the input power. This trace decrease is true for both the noise background and the signal. How could this possibly be the case?

I've attached two files. One shows the spectrum after injecting a -40 dBm tone. The other shows the results of the same measurements with different injected powers. The maximum of the spectrum is displayed. I also wanted to attach the data files, but dpt files are not accepted as valid file extensions.
Attachments
tone_injections_max_comparison.pdf
We performed the same injection measurement for various tone powers, from -80 dBm to -40 dBm. The averaged trace seems to be lower than the instantaneous trace by about 10 dBm for all of the input powers.
(17.99 KiB) Downloaded 882 times
trace_spectrum_n40dBm.pdf
Injecting 1,120 MHz, -40 dBm tone directly into RSA. This is the resulting spectrum. The averaged trace is 10 dBm lower than the instantaneous trace.
(37.38 KiB) Downloaded 871 times

w2aew
Tektronix Applications
Tektronix Applications
Posts: 83
Joined: December 2nd, 2008, 2:27 pm
Country: United States
Contact:

Re: RSA5106B Trace Averaging

Post by w2aew » June 20th, 2017, 10:22 am

Here are a couple of comments regarding your trace averaging observations. Note that you have to take a look a the trace detector types used for each of the traces. By default, Trace1 defaults to using a +Peak Detector. Trace3 is the default averaging trace, and it uses an Average Detector which will read lower than a +Peak trace. Also, remember that the scale is logarithmic but the averaging is done in with the linear measured results - so the trace won't be in the "middle" of live trace variation. There is an option to perform an "average of logs", which is mathematically incorrect, but is included for legacy applications that used this method years ago.

If you could save a screen image (File > Save As..) or even a Setup file, that would help to us to see what settings you used.

rcervantes
Posts: 3
Joined: June 19th, 2017, 6:15 pm
Country: United States

Re: RSA5106B Trace Averaging

Post by rcervantes » June 21st, 2017, 12:53 pm

Thank you.

It appears that Trace 1 is +Peak, but I was told by others that Trace 2 displayed the maximum. Can you tell from the attached setup files what Trace 1 and Trace 3 correspond to?

Also, I just realized that the tone generator power is the amplitude. I don't know if this amplitude corresponds to an RMS amplitude or a peak amplitude, but it could explain the difference between the two traces. Unfortunately, I haven't found anything in the Agilent N5183B MXG Signal Generator documentation that tells me what type of amplitude it displays.

"Also, remember that the scale is logarithmic but the averaging is done in with the linear measured results - so the trace won't be in the "middle" of live trace variation."
Sorry. I was sloppy in my wording. By "middle," I meant in between.
Attachments
20170609_114456_Trace3_rf_tone_3_data.txt
Datafile for Trace 3. I had to change the .txt file from .dpt to .txt because .dpt wasn't a valid extension to upload.
(33.24 KiB) Downloaded 872 times
20170609_114512_Trace1_rf_tone_1.txt
Setup file for a Trace 1 measurement. I had to change the .txt file from .Setup to .txt because .Setup wasn't a valid extension to upload.
(551.87 KiB) Downloaded 874 times
20170609_114456_Trace3_rf_tone_3.txt
Setup file for a Trace 3 measurement. I had to change the .txt file from .Setup to .txt because .Setup wasn't a valid extension to upload.
(551.87 KiB) Downloaded 878 times

w2aew
Tektronix Applications
Tektronix Applications
Posts: 83
Joined: December 2nd, 2008, 2:27 pm
Country: United States
Contact:

Re: RSA5106B Trace Averaging

Post by w2aew » June 21st, 2017, 1:05 pm

According to your setup files, Trace1 uses the +Peak detector with "Normal" display function (updates live). Trace2 is also +Peak detector, but with a Max Hold display function. Trace3 uses the Avg detector and the Average display function. You can see this for yourself by hitting Setup > Settings and looking at the settings on the Traces tab.

rcervantes
Posts: 3
Joined: June 19th, 2017, 6:15 pm
Country: United States

Re: RSA5106B Trace Averaging

Post by rcervantes » June 21st, 2017, 1:59 pm

Excuse my ignorance. I am new to Real-time Spectrum Analyzers.

Does a +Peak detector with a "Normal" display function mean that the trace displays the maximum in an acquisition time? I am using acquisition time as the amount of time it takes to make one spectrum window (~1/85MHz). Sorry if my terminology is incorrect.

Does a +Peak detector with Max Hold mean that it keeps the highest value while data accumulates for an indefinite time?

What exactly is the Average display function? Does it take into account the entire measurement run or just one acquisition length.

w2aew
Tektronix Applications
Tektronix Applications
Posts: 83
Joined: December 2nd, 2008, 2:27 pm
Country: United States
Contact:

Re: RSA5106B Trace Averaging

Post by w2aew » June 21st, 2017, 2:15 pm

There is a nice blog posting that describes the spectrum trace processing that occurs in the RSA. It can be found here:
http://www.tek.com/blog/spectrum-trace-processing-rsa

In a nutshell - the Detector affects how the acquired samples are processed into spectrum trace points. The trace Function refers to what happens to the traces with multiple acquisitions.

The realtime DPX display is unique in that it can capture/process hundreds of thousands of spectrum results per second. This is much faster than the display update rate. So, multiple spectrum traces are computed and entered into a database. Then, about 30x per second, these results are converted into the color graded bitmap display, where the colors indicate the "density" of the accumulated traces in any given area of the display.

Post Reply

Return to “RSA5000/6000 Series”

Who is online

Users browsing this forum: No registered users and 1 guest