On Tue, Mar 08, 2022 at 05:46:47PM -0800, Richard Cochran wrote: > The problem with uncorrected PHY time stamps is that they affect the > boundary point between the node and the network. A static error there > will create a path asymmetry that can neither be measured nor > corrected by the PTP, and a variable error degrades the time signal. And FWIW, the imperfections in the cable *also* introduce path asymmetry and thus uncorrected offsets. The twisted pairs are never exactly the same length in both directions, for example. However the PHY delays can be in the microseconds, but cable delay deltas in the nanoseconds, so correcting PHY delay is much more important. Thanks, Richard
