Noise floor of spectrum analyzer?

[GDF]

I found the noise floor of my spectrum analyzer is -120dBm/Hz.

Is this a normal value? What's the typical value?

Thanks,

 

[ulkucu_hareket]

I found the noise floor of my spectrum analyzer is -120dBm/Hz.

Is this a normal value? What's the typical value?

Thanks,

near normal value..dont fear

 

[Ghost Tweaker]

Hi,

depends a lot on the model you use... I would say -150dBm/Hz is standard and -165dBm/Hz can be achieved with a preamp placed at the input of your spectrum analyzer.
One thing that could be degrading your noise floor is the attenuation, check that it is set to 0dB...

Regards

 

[GDF]

By the way, may I use this spectrum analyzer to measure noise floor of a system
such as active filter?

Thanks,

 

[biff44]

Ghost is right. Most spectrum analyzers have a broadband mixer right at the front end, so the system noise figure is poor. You can reduce the input attenuation to zero dB, BUT you have to make sure you do not exceed the safe input power limits or you will be out $5000 fixing the front end (most spectrum analyzers require you to manually program the removal of ALL of the input attenuation). A LNA preamp is the standard way to improve the system noise floor, but you have to be carefull that the amp does not saturate (or your readings in dBc will not be accurate), and that the amp does not generate unwanted spurious products (you can tell by momentarily programming the SA to have a 10 dB attenuator, then a 0 dB attenuator switched in, and see if the products go up/down by 10 dB. If the products change by more than 10 dB,then they are not real (generated in the amp or mixer front end instead of the device under test)).

 

[VSWR]

A rule of thumb is that the noise floor of the spectrum analyzer shall at least be 10 dB lower than the level to be measured. What is the expected noise level of your active filter to be measured?

By the way, may I use this spectrum analyzer to measure noise floor of a system
such as active filter?

Thanks,

 

[sergio mariotti]

Generally SA are not low noise instruments. Most of them, syntetised or not, very expensive or expensive, have an equivalent input noise figure of 30...40 dB when input attenuator is set to 0 dB.
Take in mind that noise figure constant even if you ghange resolution bandwidth.

I don't think SA may be used to measure noise floor withoud a preamplifier

 

[GDF]

Generally SA are not low noise instruments. Most of them, syntetised or not, very expensive or expensive, have an equivalent input noise figure of 30...40 dB when input attenuator is set to 0 dB.
Take in mind that noise figure constant even if you ghange resolution bandwidth.

I don't think SA may be used to measure noise floor withoud a preamplifier

I insert an unity gain buffer based on OPamp between my DUT and spectrum
analyzer. My interesting frequency is from 200Hz to 100KHz, and the Spectrum
Analyzer is hp 8601E. My question is can I trust the measurement result?
Because this is an RF spectrum analyzer, but my interesting frequency is
below hundred kilo hertz.


By the way, I heard we can't feed DC to our spectrum analyzer before, but I dont' know the reason until now. Can someone tell me why?

 

[sergio mariotti]

I insert an unity gain buffer based on OPamp between my DUT and spectrum
analyzer. My interesting frequency is from 200Hz to 100KHz, and the Spectrum
Analyzer is hp 8601E. My question is can I trust the measurement result?
Because this is an RF spectrum analyzer, but my interesting frequency is
below hundred kilo hertz.

In the front panel of SA you should read the frequency range, usually the lower freq. is 9 KHz

By the way, I heard we can't feed DC to our spectrum analyzer before, but I dont' know the reason until now. Can someone tell me why?

The input stages are composed by an attenuator followed by an expensive mixer. The mixer is DC coupled. a DC voltage ≠ 0 will destroy the mixer.
Look at the yellow label near the connector