wr_mon shows incorrect link length value
It does not affect timing, it's only an extra information but may be confusing for users.
Full report by Anders Wallin:
For f1, the short fiber, I used a 2m LC-LC patchcord.
Following the calib-document I get rho_1 = 19.785 ns
Assuming n=1.467 this rho_1 corresponds to L = 0.5* rho_1 * c / n =
2.024 m (within tolerance, after all length metrology is a messy
business :)
However after inserting TXM, RXM, TXS, RXS, and alpha values according
to the calibration procedure wr_mon on slave shows:
Link Length (m): 91
This is obviously calculated directly as L = D_ms * c / n from the
value that wr_mon shows as “Master Slave delay”.
However it seems that “Master Slave delay” shown by wr_mon is TXM + rho_MS + RXS, and for a short fiber this is much larger than rho_MS.
When calculate this myself from the values shown on screen I get RTT –
TXM – RXM – TXS – RXS = 19.798 ns (close enough to my rho_1 during
calib)
And from this again I would get the patch-cord length 2m.