AnsweredAssumed Answered

Unwrapped Phase Results w/ Different Frequency Span Settings

Question asked by TedAyres Employee on Jan 9, 2010
Latest reply on Oct 29, 2013 by hfler
I am trying to help two PNA users understand why their measurements of unwrapped phase are different, when the number of points are the same, but sweep spans are different.  They want to know which measurement is correct.  In the process of helping them, I want to better understand PNA's phase measurements, too. 

In the first user's case, "absolute" phase measurements (in units of degrees) is desired.  This person is using an E8362B to measure phase through A)  12-inch-long coaxial cable, and B)  digital phase shifter.  The hope is to correctly measure the cable first, then move on to the digital phase shifter, in which it is desired to balance phase through two separate RF paths.  When unwrapped phase measurements are made on the cable, something other than the expected 5400 degrees is seen when a marker is placed at 15 GHz. 

In the second user's scenario, an unwrapped phase measurement is also made on E8364B and E8361A.  The issue is that with two different spans, two different measurements of unwrapped phase occur.  When the sweep span is set from 100 MHz to 18 GHz, he sees 5500 deg when the marker is placed at 2 GHz.  When the start frequency is changed to 1 GHz, (i.e. span 1GHz:18GHz), he sees 3300 deg on the marker set at 2 GHz.  This user states that standard (i.e. wrapped) phase format measurement doesn't change, regardless of span settings.  

I made my own measurements to replicate the issues observed by the PNA users.   

Attached are two screen images that show my measurements of unwrapped phase on a 965 mm long SMA (m:f) coaxial cable. 

[attachment=1]1601 pts GD Unwr Phase 10M-26_5G Span.png[/attachment]
[attachment=0]1601 pts GD Unwr Phase 2G Span.png[/attachment]

In both of my tests, the following conditions apply:

- E8364B running A.07.50.57 firmware
- 85134F test port cables
- N4691B-M0F ECal module
- NOP = 1601 (used b/c default NOP = 201 yields negative GD results)
- IFBW = 100 Hz
- Averages = 1
- Linear Frequency Sweep Type
- Stepped Sweep ON
- First test start/stop frequencies are 10MHz/26.5 GHz
- Second test start/stop frequencies are 14GHz/16GHz
- Single Channel used for measurement and calibration
- Tr1 S21 Unwrapped Phase Format
- Tr2 S21 Group Delay Format

Both of my tests show very similar Group Delay measurements of about 4.1 ns.  But, the unwrapped phase measurements (22,194 deg over wider span, 1,312 deg over narrower span) are very different. 

I would like to ask a few questions to help me (and others) understand what is happening here. 

Q1.  Why do the unwrapped phase measurements differ when the sweep frequency spans differ?

Q2.  Should the standard phase format measurements be the same, regardless of start/stop frequency settings?

Q3.  What is the correct way to measure the so-called "absolute" phase through a linear device like a coaxial cable? 

Any guidance from others who encountered this issue before would be welcome.