I am trying to understand the time domain gating shape used in the PNA. This is because I want to evaluate measurement uncertainties in my data due to the gating, especially the band edge effects.
In the attached file, the red line is the TD gate in dB ("normal" gate) used by the PNA (gate start 40ns stop 50 ns). The PNA gate is obtained by sending the PNA a frequency response of 1 at center frequency, and zero everywhere else (1GHz-15GHz). The plot is normalized so the passband of the gate is at 0 dB. It has two strange sidelobe near 35ns and 55 ns.
According to the PNA manual, for the normal gate, it has a sidelobe level of -68 dB, and cutoff time of 2.8/freq span. I designed a Kaiser window filter which meets the same requirement. It is the blue line. It can be seen that the PNA time gate is not a Kaiser Windowed filter as someone suggested in an earlier thread in this forum. The shape suggests it is not an optimal FIR filter either.
For the minimum gate, per the manual, it is supposed to have a sidelobe of -48 dB. The actual gate in the PNA falls short of that spec at -44 dB (see attached).
Please inform what the designs of these gates are? I don't see a reason to keep it proprietary, as it is needed to understand the measurement uncertainties of gated data.