### Q

What is the RELATIVE level of the DDS ""spurious"" at DDS-Clock ( AD9954 at400 MHz.) / 8 ,and DDS- Clock / 16 ?

( Ofcourse with no filtering but direct at R-DAC and single ended oputput )

Any measuremnts done ?

And are these spurious levels affected if output is done with broadband balun ?

### A

The DDS functions like a high-resolution frequency divider with the referenceclock as the input and the DAC as the output. The spectral characteristics of

the reference clock directly impact the spectral. characteristics of the DDS.

Any phase noise or spur on the reference clock will also appear at the DAC

output, but will be reduced in magnitude by the frequency division process. The

improvement is directly related to the ratio of input to output frequencies,

expressed as:

dB = –20 log(N)

where N is the ratio of the reference clock frequency to the output frequency.

Figure 2 shows phase noise improvement resulting from division of the reference

clock frequency. Residual phase noise should be just the phase noise added by

the DDS, but the contribution from the reference clock is not completely

eliminated. Here the reference clock is 300 MHz, with output frequencies of

5MHz (blue) and 80 MHz (red). The difference between the phase noise of 5 MHz

and 80 MHz at lower frequency offsets is 24 dB, as predicted by the equation

above. This indicates that the inherent REFCLK path (internal / external) is a

significant contributor to the overall phase noise.