Thank you, Bob!
I've seen the FFT spur when increase the peak to peak voltage.
However, when measuring in the time domain, the peak to peak variation of the data include large parts of white noise, how can I keep them away from calculating?
In addition, How can I test the PSRR in high frequency, say 10kHz. I found the digital filter of ADS1220 has a big attenuation on these high frequency components, which make it difficult to see the spur in FFT.