For measuring dynamic parameters we use sine wave.. How ever for static performance, we use a slowly varying ramp..
Obviously the dynamic parameters should get worse when we increase the frequency. I don't think static performance (measured in terms of DNL and INL) will vary with increase in sine frequency. That is why they call it static. It does not vary with frequency but the error is due to static error in the system..
Right as frequency increase, the circuit operates in more non-linear region and non linear effects increases. This translates into increased error in dynamic performance..
besides, SNDR=Ps/(Pn+Pd), Ps is the signal power, Pn is noise power, Pd is
distortion power, when simulation, i find that when fin increases to a large value,
such as near to fs/2, Pd becomes very large relative to Pn, like Pd=10Pn, why?
why Pd so large? and what is the possible reasons from circuit design?
which factors from circuit design can cause Pd ten times larger than Pn?
Am not sure whether you introduced noise in your input. If not, then you cannot expect it to be comparable with Pd right? The noise you assume might be the internal noise in the devices and may not be the noise which might be present in the signal! And distortion naturally increases due to increased influence of non-linear effects when frequency increases..