The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> Noise
https://designers-guide.org/forum/YaBB.pl?num=1261627724

Message started by raja.cedt on Dec 23^rd, 2009, 8:08pm

Title: Noise
Post by raja.cedt on Dec 23^rd, 2009, 8:08pm

hi,
can any one please explain out off noise spectrual density and rms noise, which one should i concider because in many circuits we can reduce rms noise without toching spectrual density.

Thanks,
Rajasekhar.

Title: Re: Noise
Post by Mayank on Dec 23^rd, 2009, 8:48pm

Hi raja,
Can you pls explain what you meant by
Quote:

in many circuits we can reduce rms noise without toching spectrual density.

RMS Noise includes noise from all harmonics across the PSD. So it gives you overall Noise that would be present in your circuit. If you are employing some kind of Noise-Shaping OR Some Filter to filter out Noise, then the Spectral Distribution matters. Maybe someone else can explain better.

--
regards,
Mayank.

Title: Re: Noise
Post by raja.cedt on Dec 23^rd, 2009, 9:47pm

hi mayank,
what i am saying is you can reduce integrated noise by decreasing BW (by means of cap), at the same time specrual density can be reduced by decreasing R, but don't see any change in integrated noise. i just want to know where we need to concider integrated noise and where density.

Thanks,
Rajasekhar.

Title: Re: Noise
Post by currant on Dec 24^th, 2009, 2:43am

raja.cedt wrote on Dec 23^rd, 2009, 9:47pm:

where we need to concider integrated noise and where density.

In my opinion, this is, from one side, depend on application and signal detection method. If our detector - voltage comparator, then we are interested in small rms value, but if our detector works in spectral domain (after FFT), then - in low noise floor.
But, from another side when we use ADC, problem of folding exists and rms value of noise is very important.
In almost all systems main parameter is SNR (not SFDR), so rms noise value more important parameter.

Best Regards

Title: Re: Noise
Post by raja.cedt on Dec 24^th, 2009, 2:47am

hi currant,
as agree with you, but i heard there are some case's in which noise floor is imp. In you your ans you are saying that when we are working with FFT's (whats the meaning)

Thanks,
rajasekhar.

Title: Re: Noise
Post by currant on Dec 24^th, 2009, 4:39am

I've seen several systems, in which detection was produced in spectral domain. That is, we need collect N points of signal, take FFT of its and compare, if any spectral component is higher then threshold.
But Rms noise voltage very closely interrelated with spectral density, especially for thermal noise.
I think, theoreticaly spectral density is important then noise has some shaped power spectrum and usefull signal is occupy small part of analysis band. when the noise power spectral density in signal band is important. This is my mind only.
Btw noise power spectral density important in Local oscilator signal due possible reciprocal mixing with blocking (interference) signal.

Title: Re: Noise
Post by thechopper on Dec 24^th, 2009, 6:39am

Generally speaking, noise floor is important when considering the very input of your system (or the first amplifier in the signal change). That happens because this is the point in the circuit where mostly the overall noise levels are set for the whole system.

Following gain stages will somehow limit the BW and therefore integrated noise starts to make sense to be considered there.

In the end, both integrated and psd are different expression of the same parameter: noise power. In my opinion noise psd is more general since integrated noise is derived from the psd applying the system limitations (i.e. BW)

Best
Tosei