You can do a more approximate validation on the cheap by measuring certain devices for which the response is more or less known: They are expensive though and require some skill to use.
High precision airlines (precision machined coax lines with air dielectric) are often part of such verification kits since they can be machined very accurately and their response can be very well predicted from textbook coax formulas. More discussion of VNA calibration verification is in this Maury Microwave appnote, and probably Keysight and other VNA manufacturers have relevant app notes too. One such standard may be a precision offset short. To validate a calibration properly you need a set of independent verification standards which don't have the same reflection/transmission characteristics as your cal standards.
It can catch some problems but it's not a complete validation. Remeasuring the standards is more of a test of repeatability of the VNA hardware and the repeatability of connecting a cal standard at a reference plane. The VNA solved for a set of error terms based on the exact cal standards, so even if there is a problem the re-measurement of one or more cal standards may look fine while measuring anything else is way off. Thank you.Ī point to add on to what others have said is that directly remeasuring your calkit standards is not the best way to validate a calibration. (So open response measures open and load pretty cleanly, but short spreads out on the left side of the chart.)Īny thoughts on what I could try? This equipment has worked before, but I don't know if anything may have happened to it since it last worked well. We get normal looking results if we do an Open or Short Response Calibration (only one standard used), but the measurements are fuzzier the further they go from the standard used. We tried going through cables, not through cables (directly on VNA port), calibrating on port 2 and measuring S22 in case port 1 was bad, but always get the same kind of result. The load measurement is beautifully centered, but the open measurement is an arc along the bottom of the outer circle, and a short measurement is an arc along the top of the circle open and short aren't nice little points at the right/left sides. We go through the steps for a 1-port mechanical calibration with short, open, and load. We're using an Anritsu Male N connector standard, sweeping 600-2700MHz, and even doing 4 averages to try to smooth things out a bit. Our Field Fox VNA isn't getting good calibration results. My department is currently a bit stumped.
0 kommentar(er)
