What is exact definition of ADC SFDR?

Hi

What is exact definition of ADC SFDR?
What' is proper hardware or software setup to measure ADC SFDR?

Thanks

Re: SFDR What's this

Grig said:

Hi

What is exact definition of ADC SFDR?
What' is proper hardware or software setup to measure ADC SFDR?

Thanks

Click to expand...

SFDR is the difference in dB between the rms amplitude of the input signal and the peak spurious signal.

You should capture the outputs of ADC, then do FFT or DFT.

SFDR What's this

SFDR=P(input signal)-P(max spurious tone)

Re: SFDR What's this

Hi

OK, but single tone signal (on spectrum analyzer) hase some width (has phase noise) and has not flat shape.... Where is spur in such case? And what is spur?

Thanks

Re: SFDR What's this

In ADCs, Spurious-Free Dynamic Range (SFDR) is the ratio of the RMS amplitude of the carrier frequency (maximum signal component) to the RMS value of the next largest noise or harmonic distortion component. SFDR is usually measured in dBc (with respect to the carrier frequency amplitude) or in dBFS (with respect to the ADC's full-scale range).


A graphical illustration is at:
**broken link removed**

SFDR What's this

You can find the difinition in D.B.Chester...'s paper
"Analog to digital converter requiments and implementations for narrowband channelization applications"

Re: SFDR What's this

OK
I know this undestandable definition, BUT....
What is SFDR in the following example

Thanks

Re: SFDR What's this

Grig said:

OK
I know this undestandable definition, BUT....
What is SFDR in the following example

Thanks

Click to expand...

hmm...

you have high phase noise on source carrier _or/and_ sample clock on AD...

- or FFT-artefact depend of sample window and selected window-funktion as retangular, Haming, Hanning, riced cosine etc....

Remeber for exact frequency 'spike' from pure source you need exact even period of signal in sampling window, ie exact 10, 23, 100 periods on ex. 1024 sampel
not 10.01, 23.3 or 100.2 periods etc. - part of periods outside of even periods give smearing effekt of carriers 'spike' in FFT-spectrum analyze.

---


I cannot answer on SFDR here depend of dirty sample clock or/and signal input...


Try two-tone test (and ensure peak value of sum of tones not going outside full scale of AD-converter or intemodulate before AD-input) and look after IM-products.

IM-makes very quick if you have non linerarity in circurits and for example needs
60 dBc carrier to noise/IM floor in multicarrier system as Cable-TV distrubition.