hello, I have read the datasheet of AD9163, it works well on phase noise performance. but i need two channel output directly, so i would like to choose AD9173.
the phase noise of AD9173 seems to degrade a lot than AD9163, because the datasheet only showed the PLL DAC clock source, while the AD9163 is directly driven by OCXO multiplication. If the AD9173 is driven without PLL ,would it work as nice as AD9163?
The short answer is yes. Phase noise is typically (for any DAC) a function of supply noise on the clock receiver, and existing clock phase noise, in addition to the additive phase noise from the DAC. Typically, the limitation is in the clock's phase noise and/or supply noise (1/f). We do have data in the datasheet showing PN for when providing an external clock (as opposed to the onboard DAC PLL), with the cleanest lab source one would find (either Keysight's PSG, R&S's SMA100, etc).
Can you supply the Residual Phase Noise curves at various operating frequencies, say 200MHz-1GHz? I need to compare the DAC's (non PLL mode) noise against the AD9914 DDS.
I do not have the residual PN at the moment. But if you are after the absolute best PN, the AD916x is the way to go. The AD917x is good but it is a dual DAC with a mroe complex clocking path. Which typically influences phase noise.
You can infer by looking at the AD916x data (non residual) and compare against the AD917x. They were both measured using similar LDOs (mainly clock path) and similar clock sources. So the relative difference stands.