Hi PSA users:
I am using a PSA E444X spectrum analyzer. I can use TRAC:DATA? TRACE 1 to query the trace data. This returns the y data, i.e. the amplitude data. I also need the x data, i.e. the frequency points which accompany each amplitude point. I tried to generate my own vector of x data based on center frequency and span, but this is where the difficulties arose. Although I set start frequency = 50 kHz, and stop frequency = 1050 kHz, in fact I noticed on the spectrum analyzer that the start frequency actually got set to 37 kHz. I query the start and stop frequency over GPIB, and the PSA responds with 50 and 1050 kHz, respectively, but this is not what is shown on the screen. When I move the marker to the far left of the screen, it starts at 37 kHz.
Since in general, I am programming remotely, I don't have the opportunity to notice that the actual start frequency is not the start frequency which I requested in my command. Then, when I plot the results of the TRAC:DATA? TRACE 1 command, I assume the wrong start frequency, and all of my frequency data is off, sometimes by as much as 10% in frequency. Does anyone have a solution to this problem?
Thanks.
Hello jbostak,The PSA had an interesting quirk for spans <2MHz when in dual loop mode (not using FFTs or not in single loop phase noise optimization) where the span may be different than what is selected. However, the marker is always accurate.
As a workaround for these low spans, you can try these commands:
[:SENSE]:SWE:POIN to find the number of returned trace points
:CALCulate:MARKer[1]|2|3|4:X <Param> Set X marker position
:CALCulate:MARKer[1]|2|3|4:X:POSition? Query X marker position
If you move the marker to an X position which is not on-screen, it will default to the farthest side of the screen in that direction. You can use this to put the marker in the first (the real START frequency) and the last (the true STOP frequency) locations and the query to record those frequencies with your data. Then, using the number of trace points, if you assume the points are distributed evenly (they aren't perfect, but close enough for most applications) you can get a good X value for all your amplitude data.
I got this information from the User's and Programmer's Reference for the PSA starting on page 153.
Hope this helps.