You are absolutely right, and I used to know this and completely forgot! Now that I set my master_clock_rate to 200MHz, sample_rate to 120MHz, and DDC from 200MHz to 100MHz, everything is working great! I appreciate the time for the answers and explanations! --------- Original Message --------- Subject: Re: [USRP-users] X310's sample rate From: "Brian Padalino" <bpadal...@gmail.com> Date: 12/4/18 10:40 am To: "Jason Matusiak" <ja...@gardettoengineering.com> Cc: "USRP-users@lists.ettus.com" <usrp-users@lists.ettus.com>
On Tue, Dec 4, 2018 at 10:35 AM Jason Matusiak via USRP-users <usrp-users@lists.ettus.com> wrote: I have a X310 with a pair of CBX-130s installed and am running RFNoC. The flowraph looks like this: Radio (running at 200MHz) -> DDC (200MHz down to 50MHz) -> splitter -> off to some other blocks. What is weird to me is that I wasn't thinking and I set the sample_rate to to be the master_clock_rate (200MHz in this case), but the CBX-120 only has a sample-rate of 120MHz. This process seems to work and my flowgraph is running as I expect it should. There are no warning or errors from UHD when I run this saying that I had an incorrect sample_rate. The 120 is the analog filter bandwidth of the radio card. The samplerate is very limited on the X310 to be either 200MHz or 184.32MHz and must be set during device initialization, I believe. Now, if I leave the master_clock_rate alone, but set the sample_rate to be 120MHz, the flowgraph barfs immediately and I get timeouts non-stop. So I am guessing that there is a concept between the master_clock_rate/sample_rate, or some combination of them and the DDC that I am missing. Need to use integer division to get to 120MHz, and that isn't going to happen from 200MHz or 184.32MHz. Take master_clock_rate and divide by positive integers, and that's how you get valid sample rates. Moreover, if you want to ensure good passband behavior, it should be even, and if you want better passband behavior, it should be divisible by 4. In the even cases, halfband FIR filters are used instead of strictly a CIC filter, which has some passband rolloff. Hopefully this helps. Brian
_______________________________________________ USRP-users mailing list USRP-users@lists.ettus.com http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
