In my use case, I have a single WR Switch (from Seven Solutions) connected
to 8 N3xx devices (WR slaves) via single-mode fiber (~ 10m). I have a few
questions:
- *Accuracy*: How does the accuracy of WR compare to running 10 MHz/PPS
over coax of similar length (~10m) from a reference such as an OctoClock? I
don't need specific quantification, but I am just looking for general
comments.
- *WR switch mode*: There are 3 choices: GrandMaster, Boundary Clock, or
Free-running Master. (In my use case, there are occasions where I will
have a 10MHz/PPS reference available and other occasions where I won't.)
- It appears that if I choose GrandMaster, I *must* have a 10 MHz/PPS
reference input. The WR Switch does not seem to behave well if these
signals are not present. Is this correct?
- Conversely, if I have a 10MHz/PPS reference available and want to
use it, I *must* choose GrandMaster mode. Is this correct?
- I have found documentation that indicates that it is optional to
have a 10MHz/PPS reference. But, it is not clear which mode I
should choose
for the case where I do not have such a reference. I have also found
documentation that indicates that the PPS output is *not* available in
Boundary Clock mode, but that it *is* available in Free-Running Master
mode. Does this mean that in my use case with devices that
require PPS, I
must use the Free-Running Master mode for times when I don't have an
external reference but then switch to GrandMaster mode for times
when I do
have an external reference? (This would imply that I would never use the
default Boundary Clock mode.)
- *Optical Transceivers*: There are only a few optical transceivers
listed on the WR website compatibility list and in the dot-config file of
the WR Switch device. All of them are BiDi transceivers.
- In my use case with short distances, is there any disadvantage to
using 2-fiber transceivers (common wavelength of 1310nm) at both
master and
slave ports? I recognize that single fiber BiDi solutions have the
advantage that the Tx and Rx path lengths are identical, but I
would expect
that a duplex fiber has *very* similar path lengths and it has the
advantage of common wavelength.
Rob
_______________________________________________
USRP-users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
