Measuring the bandwidth of a sinus on a spectrum anlyzer

Hello everybody,
i'm new in the forum.
I have a fundamental doubt and I can not get over it.
If I measure a sinus on a spectrum analyzer, how I can measure it's bandwidth?
Let's suppose a have a Gaussian shape with a certain 3db bandwidth; my problem is that if I change the frequency window, resolution and sampling time/points, my result changes dramatically. The strange thing is that, if I measure with the large possible resolution and in the smaller frequency window possible, the signal seems just having a smaller bandwidth.... I really don't understand.

Any suggestion?

Thank You

Hi,

with gaussian shape you mean "windowing"... for FFT analysis for example.

My recommendation: Read about windowing and the effects of it. There are various window functions each with it′s benefits and drawbacks.
You get the best FFT results when you adjust the sampling frequency and FFT size to the desired signal. Then you don′t need a window function.
If it fits or not you see on the noise floor or the harmonics of the fundamental frequency.

Test if you get better results with a hanning window.


Klaus

Thank You Klaus,
Bygaussian shape I mean that I have a sinus generated by a signal generator. It goes through my RF system, and then I measure it on a spectrum analyzer. Since it is a sine, I see a peak, that I suppose to be gaussian. I don't know if I was clear in the first post. Anyway I will read about windowing as suggested.

Hi,

Ok I misunderstood.

The output shape of the spectrum analyzer is gaussian shape (or at least estimated to be).

Of a pure sine the bandwith is zero. How to measure zero?
It is a problem of:
* resolution
* sample clock jitter
* S/H aperture jitter
* and other factors.

The most problem is the resolution. For sure it depends on sample rate and window size.
With varying sample rate and window size the frequency grid = resolution varies.
Can you show us some diagrams of the "different" bandwidths you measure. Maybe zoomed in to the frequency of interest.

Klaus

If the time base speed is set too high, i.e. the receiver has a high Df/Dt sweep, the sine wave is only in the spectrum analysers pass band for a very short time. The result is that the apparent amplitude falls and the filters in the analyser "ring", so the bandwidth of the signal appears larger. To measure the amplitude just slow the sweep down until the signal does not get bigger.
Frank