AnsweredAssumed Answered

8753ES and 8720ES Programming Inquiry re: Interpolation

Question asked by TedAyres Employee on Feb 29, 2012
I am submitting the below questions on behalf of someone who programs 8753ES and 8720ES frequently.  Assume the firmware involved is A.07.74. 

The user explains the situation asks the following questions: 

I am writing ATE software to test microwave filters. I am right now trying to 'error check' the input specs to make sure the ATE software doesn't allow dumb things.  I need to know which stimulus parameters can be changed in a sweep record without invalidating a shared calibration, and for both the situation where error correction interpolation is ON, and when it's OFF.

The specific issue I am trying to get a handle on is: 

(1) what differences in stimulus settings between the sweep and the cal will invalidate the cal;

and (2) what stimulus settings can change without invalidating the cal, and where appropriate, what are the limits/rules?

Here's what I've gleaned so far:

(I believe this to be true for the 8753D, 8753E, 8720D, 8720E, 8720ET, 8720ES)

1. Interpolation works for sweep types LINFREQ, POWS, CWTIME
          does NOT work for sweep types LOGFREQ, LISTFREQ

2. With interpolation OFF:

•       changing freq range, Npoints, or sweep type INVALIDATES the calibration

•       changing sweep time, IFBW, or ouput power DEGRADES the calibration

•       Questions:

                     Q1.   The book(s) say that changing the SweepType INTERP OFF)  invalidates the cal. Does that
                                 mean when that when we do a calibration we have to select a SweepType first?

                     Q2. The book(s) [i.e. manuals] say that changing the SweepTime invalidates the cal.  If the calibration is
                                 performed with SweepTime=auto, what does that mean? Must I interrogate the calibration
                                 to see what sweep time was actually chosen in 'auto' mode in order to know if my
                                 measurement sweeps will have a different sweep time, and therefore a degraded error correction?

    With interpolation ON:

•        changing freq range (within cal span), IFBW, output power, Npoints is OK

•        changing from LINFREQ to CWTIME is OK

3. Interpolated correction works if phase shift is < 180 degress per 5 measurement points

4. Error correction is not accurate if there are less than 67 points per 1GHz of freq (in a paragraph about interpolated correction)

                      Q3. Is the "67 pts/1GHz" limit true only with INTERP ON, or is this also true with INTERP OFF?

                      Q4. Is the "phase shift < 180 degrees per 5 measurement points" limit true only with INTERP ON, or is this
                                also true with INTERP OFF?

                    5. And finally:

                                           Q5. Is any of the stuff I've gleaned incorrect? Anything else I should know?

If there is anyone in the VNA user community who can answer these questions based upon their knowledge and prior experience, there is likely to be more than one 875x or 872x user who will appreciate it.