Noise floor of Spectrum Analyzer

[kamranraja222]

Hello everyone,
I need to find out the noise floor of my spectrum analyzer(SA). I put a 50 ohm termination at the input port. It gave me the value of -130 dbm. Than I insert an attenuation of 20 dB/10 dB/30 dB one by one with 50 ohm termination. It always gave me the same value of -130 dBm(In my opinion it should raise the noise floor now). Please some one throw a light at this issue. Thanks

 

[jiripolivka]

With the input terminated in 50 Ohms, external attenuators do not generate more noise as they are at the same ambient temperature.
With a good spectrum analyzer you can reduce the noise floor by adjusting RF badwidth, video bandwidth and sweep rate, more than by input terminations.
If not satisfied, you should use a preamplifier for a selected frequency range. Spectrum analyzers use input harmonic mixers which have higher noise figure than a good preamplifier.

 

[FvM]

The noise floor would be raised if you correct the magnitude display for the external attenuation, to give correct level referred to the attenuator input.

 

[kamranraja222]

@jiripolivka: Alright but when i used my circuit which consists of active as well as passive devices, it also gave me same level of -130 dbm, which shouldn't. I also want to check the noise floor level of my design circuit.

@FvM: I have tried different reference level of SA to look more closely at noise floor but nothing I got. By changing RBW and span, the noise figure shift's accordingly but than it remain same for all calculations. This is strange to me.

Thanks

 

[pancho_hideboo]

Show me product name of your Spectrum Analyzer.
Show me screen shot.

 

[kamranraja222]

This is Agilent 8563e. I am not in the lab,cannot send you picture.

 

[KlausST]

Hi,

can you use a reference sine signal (-xxdB) and view the FFT plot. To get a signal-to-noise value.

Then it might be obvious.. especially with the magnitude correction.

Klaus

 

[FvM]

Then it might be obvious.. especially with the magnitude correction.

Yes. According to the previous reports, correction for the attenuators wasn't set.

 

[pancho_hideboo]

I put a 50 ohm termination at the input port.
It gave me the value of -130 dbm.

Than I insert an attenuation of 20 dB/10 dB/30 dB one by one with 50 ohm termination.

Does your attenuator is an external attenuator not build-in attenuator ?

It always gave me the same value of -130 dBm(In my opinion it should raise the noise floor now).

If your attenuator is an external attenuator, this result is reasonable.

Calibration point is a RF connector of your SA.
So system NF which seeing RF connector is constant, even if you place attenuator before RF connector.

 

[kamranraja222]

@FvM: I am still not getting desired response. Please explain me that how can i adjust magnitude display for the external attenuation??

 

[FvM]

With my E4407, you can set external preamplifier gain or external attenuation under the magnitude menu.

Just try to navigate through the menus. Agilent operation is almost self-explanatory, I think. In case of doubt, you may want to consult the user's manual.

 

[kamranraja222]

It is revealed that when we see the noise floor of attenuator(which is terminated by 50 ohm), Spectrum analyzer always show its own noise floor.In other words, the noise floor remain same regardless of the value of attenuator.

However when we apply an input RF signal which has some noise, the signal will attenuate(depending upon the value of attenuator) but its noise floor still remain at same level.

Please have some comments on these observations. Thanks