On 5/29/19 10:39 PM, Jason Roehm via USRP-users wrote:
On 5/29/19 3:58 PM, Marcus D. Leech via USRP-users wrote:
I'm a little bit surprised this worked AT ALL--you have 3 independent
multi-usrp objects all pointing at the same hardware.
Time alignment is something that is done WITHIN A SINGLE multi_usrp object.
The right way to do this is to use a single multi_usrp object, with a
multi-channel stream--that way the time-aligner within multi_usrp
will take care of time alignment.
I would tend to agree with you. I only arrived at the current setup by specific
recommendation from an Ettus employee (see the thread I linked). Under
v3.10.3.0, trying to use a single multi_usrp object with a TwinRX and LFRX
simultaneously did not work; the library would complain about mismatching
master clock rates.
For what it's worth, I tried the same single multi_usrp approach again with
v3.14.0.0 to see if anything had changed. I didn't get any complaint about
mismatching clock rates, but I did see the same behavior that I complained
about in my earlier message (no data from the LFRX, and a LATE_COMMAND error).
Bumping this thread since I haven't heard anything regarding this
regression.
Jason
_______________________________________________
USRP-users mailing list
USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
