How can I test the overall noise in a circuit and what would be the experimental setup?

[Msmatrixin]

How can i test overall noise in a circuit and what would be the experimental setup?

If I design my own analogue circuit, then how should I measure the overall noise in that circuit. If I say, that overall noise in a circuit is the summation of all noises - thermal, shot, flicker, etc.- then would it be correct?
What is the approach basically?

 

[KlausST]

Re: How can i test overall noise in a circuit and what would be the experimental setu

Hi,

Calculation:
If the noise sources are independent, then you need to add the squared values, then take the squareroot.

Measurement:
Take a lot of ADC samples. Remove the DC offset of all samples, ten calculate RMS.

Klaus

 

[Msmatrixin]

Re: How can i test overall noise in a circuit and what would be the experimental setu

ok.
Going into basics a bit, if I have to get the input noise voltages and currents of an analogue circuit(not talking about opamps or adc but any analog circuit), how should i measure it?
Are there any calculations? or a test set-up to measure Input noise voltages and currents wherein, I give signal and supplies to the analogue circuit in the input and then check for the noise contribution in Output?(want to go by an expert's opinion)

 

[KlausST]

Re: How can i test overall noise in a circuit and what would be the experimental setu

Hi,

I´m a bit confused. I think I already answered your questions, but then you ask the same questions again.

Let´s imagine your "analog circuit" as black box. With one input and one output.
Noise sources are:
* noise input to your box (generated outside the black box)
* noise generated inside the black box.

Maybe read through this aplication note: https://www.analog.com/media/en/reference-design-documentation/design-notes/dn015f.pdf
Although it´s name says "in Opamp circuits" most of the information can be related to any analog circuit.

Some questions:
* do you want to know overall noise (which I understand as output noise), or
* do you want to know input noise values (black box inside generated only)?
* do you want the noise value as "noise voltage RMS" or do you want the "noise as spectral distribution" in a chart?

I´m not sure why you mix "calculation" and "measurement". This are different things. Either you want to calculate it (see application note) or you want to measure it (see post#2).

If you want to discuss several problems/solutions, then please decide wich one to discuss first, then focus on it.

Klaus

 

[Msmatrixin]

Thank you sir for the answer.

Pardon my lack of experience this field. I was unable to distinguish b/w measurement and calculations. The matter is, I just wanted to 'calculate' the noise values before-hand, and then may-be cross verify using a digital scope. I basically was confused that, when I say this is the overall noise in the circuit then how should I be mentioning it. Going fwd i wanted to reduce the overall noise from the circuit and that would happen if i go component by component in the circuit.

Some questions:
* do you want to know overall noise (which I understand as output noise), or
* do you want to know input noise values (black box inside generated only)?
i want to know both because i am working on both things
* do you want the noise value as "noise voltage RMS" or do you want the "noise as spectral distribution" in a chart?
i believe most of the datasheets mention it as noise spectral density, so noise as spectral distribution will be preferred.

Again if i am making no sense then please pardon.

 

[KlausST]

Hi,

To measure the spectral distribution one method is:
* measure a lot of samples. Preferred count = 2^x, sampling rate: higher than 2x the highest frequency of interest. Range and resolution according your application. (The voltage resolution of most digital scopes won't give good results). Maybe use a PC soundcard.
* then run a fourier analysis on the sample data. FFT. Even possible with Excel.

* or use a scope with FFT function.

Klaus