Rob, Question for you - is your usecase one where you can simply tolerate that 180 degree ambiguity? Or do you have a routine in your code to characterize and compensate for it on each run?
Sam Reiter On Tue, Mar 3, 2020 at 11:44 AM Marcus D. Leech via USRP-users < [email protected]> wrote: > On 03/03/2020 12:31 PM, Rob Kossler via USRP-users wrote: > > Hi Len, > I have run some testing of this capability. A few remarks: > > - In my test setup, I had a signal generator running into a 1:4 > splitter with outputs connected to the four N310 Rx2 inputs. I used > another signal generator as the LO source running into a 1:2 splitter with > outputs connected to the N310 Rx LO inputs with amplitude near 3 dBm > <https://kb.ettus.com/N300/N310#Front_Panel> at N310 LO inputs. > - When using external LO, the RF range is limited to a max of 4 GHz > (rather than 6 GHz which is the limit with internal LO) > - I had problems with UHD 3.14 because there was some bug related to > resetting the phase of the DDC. I had success with 3.15. But even with > success, there is still a 180 degree ambiguity between channels 1/2 and > channels 3/4. I am not certain regarding the circumstances that can cause > the ambiguity to flip, but I think it only occurs at application startup > such that once your application is running, you can expect consistent > results with no ambiguity flips. > > This is probably due to the 2XLO phase-splitter inside the AD9371 > chips--it's based on a flip-flop, whose initial state is unpredictable. Any > 2XLO phase-splitter has this property--including on the WBX daughtercard > used on N210/X310 etc. The mixer has a 2XLO phase-splitter, > so also shows this 180deg phase ambiguity. > > > > - When you launch your application and create a usrp device instance, > the UHD software will automatically try to run some IQ calibrations > (RX_QEC_INIT) expecting that the RF is at 2.5 GHz, which implies that your > external LO must be at 5 GHz at startup. After startup is complete (and > this calibration is complete) you can change the freq of your LO and retune > the RF of the various channels keeping in mind that the external LO must > always be twice the desired RF. If you want to avoid the startup > calibration (with LO at 5 GHz), you can specify explicitly which > calibrations > <https://files.ettus.com/manual/page_usrp_n3xx.html#n3xx_mg_calibrations> > you want using the device args and choose not to run RX_QEC_INIT. But, I'm > not sure if skipping this calibration could have negative effects on signal > quality. > > Let me know if you have questions. > Rob > > > On Thu, Feb 27, 2020 at 12:22 AM Len via USRP-users < > [email protected]> wrote: > >> Folks, >> I'm looking for information from any subscriber(s) to this forum that >> have successfully used the USRP N310 to provide 4 Rx channels which are >> both coherent and synchronized. I've heard this could be challenging and I >> know people have been able to do this successfully so I was wondering where >> I might find some detailed information/instructions on how best to >> accomplish this. I was told this forum would be a good place to start. So >> any block diagrams illustrating the Master LO and Master Clock hookups >> needed along with any available calibration procedure(s) describing how to >> actually check/verify coherence and synchronization between the 4 channels >> would be great. >> >> >> Thanks, >> Len >> _______________________________________________ >> USRP-users mailing list >> [email protected] >> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com >> > > > _______________________________________________ > USRP-users mailing > [email protected]http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com > > > _______________________________________________ > USRP-users mailing list > [email protected] > http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com >
_______________________________________________ USRP-users mailing list [email protected] http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
