Tests Results

1.Spectrum of the Output signals:
The following figures show the spectra of the 105MHz clock used for testing and the two outputs of the DDSs. The 105MHz input to the system is fed from HP ( Hewiett Packard ) synthesized signal generator and all spectra are measured using the HP spectrum analyser 8591E. The phase noise of the outputs of DDS is ~ -70dBc at 1KHz offset. The power of the sidebands due to "intermodulation" of the two outputs are -53 dBc and -39 dBc respectively for 32.25MHz and 32MHz signals. All other spurious signals over a span of 20MHz about the fundamental are > 40dBc for both outputs.
2.Clock Jitter :
Differential jitter of the rising edge of the output clocks are measured using a 400MHz Tektronix oscilloscope 2465B. The jitter is measured about 9.6µsec after triggering with the same clock and the time resolution used is 500 ps/div. The measured jitter is about 500 ps, which is comparable with the jitter of the 105MHz clock measured using the same test setup. Thus we feel that the jitter measurements are limited by the measuring instrument and should be better than 350ps ( = 500 / √2 ).
3.Synchronisation Signal
The synchronisation signals generated by 32MHz and 32.25MHz are aligned with in an accuracy of 9.5nsec ( 1 / 105 x 10^-6 sec ). The specific accuracy in nsec is the relative delay between the two synchronizing signals generated by the DDS output clocks. For the correlator application this accuracy is sufficient. The measurement is done with the 400MHz oscilloscope 2465B.
4.Programming the DDS :
The two DDSs can be programmed to a specified frequency using the software as described above. The highest frequency that can be set is limited to 105 / 3 = 35MHz. The programmed frequencies at the two outputs are measured with HP frequency conter 5386A.

5.Spectra :

GOBACK TO CLOCK SUBSYSTEM
GOTO DIGITAL BACKEND