beautypg.com

Using electronic frequency locking (efc) – Atec Agilent-8563E User Manual

Page 89

background image

Performance Tests

ESG Family Signal Generators

21. Measuring Phase Noise and Residual FM (ESG-AP and ESG-DP Series Signal Generators)

4-40

Calibration Guide

Using Electronic Frequency Locking (EFC)

Figure 4-36

shows the result of using the electronic frequency control (EFC) method of

phase-locking the DUT to the E5500 system (see the E5500 manual for set up). This
method produces the best result, but is not available on all sources (such as the
HP/Agilent 8644B). Note that the close-in phase noise is now much improved and the
60 Hz related spikes are eliminated. Although the two large spikes still exist, they are the
result of the HP/Agilent 8662/3 and are not characteristics of the DUT.

Figure 4-36

To get an accurate idea of the actual phase noise of the DUT, it may be necessary to
combine the two plots, always using the lowest result of both (see

Figure 4-37

). Keep in

mind that the actual phase noise results are probably better than what is displayed.

The results shown on the E5500 system may be worse than the actual DUT. If a tested
device has the same phase noise characteristics as the system’s down-converting source,
the displayed result will be 3 dB worse than either device. To eliminate this error, measure
three different sources. This results in three equations and three unknowns, and the
software can then sort out the actual results for each device. This 3-source substitution
method generates a correction table that can be used for future measurements, effectively
eliminating the combined error (of up to 3 dB) for subsequent measurements. For best
results, use this method. Refer to the E5500 documentation for instructions on how to
implement a 3-source substitution. When the plots in this section were taken, this method
was not used. If it had been used, portions of the plots could have been up to 3 dB better.

Another way of stating the above: for any phase noise measurement where a 3-source
substitution is not performed, the actual phase noise of one of the two sources involved
(DUT or down-converting source) is at least 3 dB better than shown.