Agilent Technologies N4413A, User Manual

Page: 107 / 234

Performing Time Domain Measurements
Practical Considerations
Chapter 4 97
This has implications when selecting the measurement parameters for a DUT, namely that
the time window must be larger than the delay through the device to have valid time-domain
data. Therefore, the absolute maximum frequency step for the measurement must be Fstep
max = 1/Tgd. For example, if the DUT is a SAW filter with a group delay of 2000 nS, the
maximum step size is 500 kHz.
Also interesting to note is the frequency domain (S-Parameters (S11)) plot at the bottom of the
figure above. Aliasing has caused the data that has been transformed into time domain and
back (upper trace) to differ significantly from the original data.
Adjusting the start- and stop-points in the time domain can improve the agreement between
the two frequency domain responses (before and after), eliminate the induced aliasing, and
thereby validate the calculated time domain response.
4. Response Resolution
The response resolution describes how close in time two responses can be distinguished. This
depends on the width of the impulse response, which is inversely related to the measurement
bandwidth. The relationship between the three is approximately R=T=1.25/BW; where R is
the response resolution in picoseconds, T is the effective impulse width in picoseconds, and
BW is the frequency span in GHz.
5. Range Resolution
As described previously in “Analyzing Time-Domain Signatures” on page 84 , the TDR
signature provides specific circuit detail. Range resolution (TD span/# of points, or
Stop-Start/# of points) will define how accurately the signature of a response can be identified.
In general, a wider measurement bandwidth will provide finer spatial resolution.
To improve range resolution, zoom in on the section of interest and adjust the start- and
stop-points to be as narrow as possible without compromising the agreement in the frequency
domain.