I have several questions about the detectors and the display points in the PNA-X SA option.First,if number of points is larger than FFT points within the span,how dose the system display the extra points?Where do these points come from?Second,in detector type option,there are seven types available,but I'm not sure what's the Peak Sample and Peak Average.I read the help file ,seems like the system should determine whether there are signals in the bucket or only noise in the bucket.It's very odd because determine a signal as signal or noise based on power seems unreasonable.The last question is whether the detector can do frequency and amplitude correction.For example,if a signal's frequency fall between two FFT bins,dose the detector have the ability to correct it to the signal's original frequency and amplitude?Dose the system use curve fitting to do that?Thanks a lot.

We know the shape of the RBW filter (typically Gaussian). The FFT grid ensures we have sufficiently sampled the signal to know the time domain waveform, thus resampling can gives us the amplitude value at any arbitrary frequency. This is like a sophisticated interpolation. So we can display correctly any number of points, larger or smaller than the FFT grid, with complete accuracy (if you believe in Nyquist). As such we can also determine precisely the peak value even if it falls between two FFT bins; thus we can have the markers read out a peak to about 1% of the RBW, regardless of the display point spacing.

Peak Sample and Peak Average use statistical methods to determine if a sparse signal lands insight a display bucket. If so it returns the peak to the display, if not it returns either the sample, or the average, off the FFT bins inside the display bucket. It is most useful when having a low VBW w.r.t. the RBW (1/10 or 1/100).