Part Number: DAC39RFS10EF
Other Parts Discussed in Thread: DAC39RFS10, , LMX1204, LMK04208, LMK04828, LMX2594, LMX2820
Hi All
We're studying reference designs for the DAC39RFS10/EF. In our targeted application we only have a single converter system: Kintex 7 FPGA --> DAC39RFS10EF. To my understanding a subclass0 JESD204 architecture would be sufficient.
We're intending to run max 6.6Gbps (FPGA GTX limit) on 8 lanes to achieve a sampling rate of about between 2 and 4 GSPS (while 3.878 GSPS seems to be the max out of the EF DAC at 64b/66b).
In the reference designs I see there are powerful and expensive PLLs used such as the LMX1204 and LMK04828. To me they seem to be an overkill for our application. Can we use also the LMK04208? It does not have dedicated SYSREF outputs and has a slightly worse jitter.
Other question: shall we best clock the DAC with the 1:1 sampling rate, so at 3.2'ish GHz or is it reasonable to use the DAC's internal PLL? Maybe the latter has worse performance? Actually I see that both LMK04828 nor LMK04208 go beyond 3.07 GHz. Would that cap our sampling rate to 3 GSPS?