It could be caused by something as simple as a synthesiser/tracking error when used on the lowest range. I don't think this can be blamed on a 'casual' calibration or 'lower return loss' as the performance suddenly degrades (badly) at a distinct part of the frequency range. It looks to me that something isn't right about the lower range because of the noisy and ripply response and also because of the sporadic trace blips in this area. When a VNA changes range on the synthesiser the VCO has to jump to a very different frequency and settle correctly before the VNA can gather any s-parameter measurements. Also, at 29:11 to 29:12 in the video a couple of brief but suspicious trace blips occur in this 1850 MHz region.
Above 1850 MHz the trace noise shows an instant improvement and looks to be as expected for -30dBm and 5kHz BW. Below this range change the performance seems to degrade a lot as it has a lot of ripple and noise. See the image below where I think an internal range change occurs at about 1850 MHz. I know the source power has been reduced to -30dBm but a typical lab VNA should perform much better than this in terms of trace noise and ripple for -30dBm source and 5 kHz BW. It is on this range where the performance seems to be degraded during the tests of the 4 port active device. I think a downconverter is used for the lowest part of the signal generation range and this probably covers something like 0-1.8 GHz. I watched the whole video this evening and spotted that the top (RF) side of the board used for tracing the signal is actually the 'flipped image' rather than the bottom although it doesn't really matter.
0 kommentar(er)
