Measuring DSB signal sidebands on the FFT of the HP54520A
I want to measure the sidebands levels (~2KHz apart) of a DSB signal on HF, on the FFT of a HP54520A.
Is that possible based on it's specs, and what kind of "resolution" should I expect in general? (slow "sweep rates" are not a problem, and I am aware that FFT dynamic range is low 55-65dB)
I attach you the manual of the HP54520A and the relevant pages for your answer might be the 177 and the 261.
Please help
Depends on number of FFT points and signal frequency, unfortunately not mentioned.
it is mentioned in the page 261. The max value says more than 32000
You still did not tell the signal frequency.
The max points are 32768 and the max frequency I need to measure is 30MHz.
In tests that I did, the "resolution" depended on the frequency. At about 1.8MHz I could get something lime 1.5KHz for every dot in the grid on the screen (which is alreaty quite bad), but at 3.5MHz I could only get 3KHz. Higher frequencies were even worst. I calculated the KHz/grid-dot manually, by observing the max and min frequency spans on the display, then subtract these, then divide the result by the number of squares in the grid (10), and then divided the result by the num of dots in each square.
If you don't have the right tool to measure the DSB levels, you can do a simple down conversion in the 0Hz-10kHz spectrum and use a simple sound card spectrum analyzer, as the one in the link below which have 130dB dynamic range and pretty good FFT selectivity (depending a bit by your computer soundcard).
http://www.techmind.org/audio/index.html#specanaly
Thank you. how about the dynamic range of this one? http://www.sillanumsoft.org/
I would be interested to know what is the special thing about the link you posted, related to this feature-rich visual analyzer.
Also is my observation correct, that the resolution depends on the frequency on the FFT, for the same number of points?
Of course it does. "Frequency" in this regard is the oscilloscope sampling rate, which must at least fulfill the Nyquist criterion. In so far it's unlikely that you achieve the intended resolution with 30 MHz signal frequency.