A future release of Paragon-neo/CAT will improve the behaviour described below. Any version up to and including R10 (CAT v30) is affected.
The SyncE to PTP (or 1PPS) transient test requires that a specific SyncE wander pattern and ESMC transitions are provided as stimulus to the DUT as defined in G.8273:
The DUT output is measured against the appropriate mask as defined in Figure B.1 for Class A/B clocks and Figure C.1 for Class C devices, reproduced below for reference:
Paragon-neo can both apply the input stimulus and make the measurement with very high accuracy. However, the timeline used for the measurement is not aligned with the generation timebase. This means there is a variable difference between t=0 in the SyncE generation timebase and t=0 in the measurement, which can lead to it looking like the DUT responded to the stimulus change before that change has actually happened.
Below are example results.
Wander stimulus starts at t=100.033s
PTP response starts at t=98.901s:
This can cause the result to look like a fail:
The variability causing this failure is attributable to the measurement process and not the device under test.
Failures can therefore be ignored ONLY if they meet the following conditions:
the failure is only seen sometimes. Repeatable failures are likely to be DUT-related and should be investigated and resolved as such.
the failure is because a transient is seen with the wrong timing i.e. the fail could be resolved by moving the mask in the x-axis (elapsed time). Any failure because the time error amplitude is greater than allowed by the mask when they are aligned in the x-axis is a true fail and should be resolved in the DUT.