# PNA-X. How can I best display phase angles close to +/- 180 deg?

Question asked by drkirkby on Jul 27, 2017
Latest reply on Jul 27, 2017 by drkirkby

VNA is an N5247A. Firmware  A.10.49.08

What the best way to display phase angles, with good resolution (say 1 deg/division), where the phase angles are wrapping around at close to +/- 180 deg?

By way of example, I attach a screen shot, where I'm trying to determine the delay of an offset short. (This is actually from a recalled .s1p file, but I doubt that makes any difference to my question.)

• The window in the top left shows the phase variation with frequency of an offset short. No port extensions are used.
• The window in the bottom left shows the phase variation with frequency. A port extension of 57.9 ps is applied. The phase is close to +/- 180 degree, so one might conclude the delay of the offset short is close to 57.9 ps. But one can't see just how far the phase angle varies from +/- 180 degrees, as the scale is so large (50 degrees/division).
• On the window at the top right, the reference is +180 degrees, but with a resolution of 1 deg/division. We can see that the phase drops to as low as about 179 degrees.
• On the window at the bottom right, the reference is -180 degrees, again with a resolution of 1 degree/division. The phase rises as high as -179 degrees.

Is there a way to display the phase, such that angles that are close to -180, get 360 degrees added to them, so allowing phase angles close to +/- 180 degrees to be shown at a reasonable scale, allowing small departures from +/- 180 degrees to be seen?

What I really want to do is find what port extension will cause the minimum departure of the phase from the ideal 180 degrees.

Strangely, on my 8720D,. setting the reference as +180 degrees, but 1 degree/division, phase angles that are close to -180 are actually shown close to +180 degrees. I've never quite understood why that happens, but it is quite convenient if one wants to look at phase angles which are very close to +/-180 degrees.

Phase wrapping is often an annoying problem. It's even more tricky if it happens in 2 or more dimensions, but at least this is a 1D problem.

Dave