I felt a little tingle in my whiskers that prompted me to call up S*man's data via MDWS . . .
Taking a broad look at the historic Hlog plots we can see that a difference is visible either side of this morning's resynchronisation event at 0822 hours. I attach copies of the relevant plots from 0800 hours and 0900 hours, below.
However after a more critical examination, we can see that there is not that much of a difference -- but a difference, indeed, is present. In the 0800 hours Hlog plot the high frequency end of the curve is just before tone 4000 with a value of -62 dB. In the 0900 hours plot the high frequency end of the curve remains just before tone 4000 but with a value of -59.8 dB.
Show me both of those plots in isolation and I would declare that "I see no indicated problem of the metallic pathway" for either. What
is making that approximate 2.2 dB change appear to be more significant is that
MDWS is using a different scale for the Y-axis of each plot.Show me both of those plots and tell me that they are of the same circuit but from different times -- which is this case -- I would declare that "the latter plot, that of 0900 hours, is the better of the two". Huh?
Very crudely put, the closer the curve to the X-axis, the "better the condition" of the circuit from which the data has been obtained.
We must remember that although we regularly see Hlog plots as quadrant one curves (be it in published papers, in JDSU or Exfo documentation, from DSLstats or HG612_stats via MDWS or those curves that I plot when examining a circuit), Hlog plots are actually quadrant four curves.
The Y-axis runs up/down the left hand side and
the X-axis runs across the top, not the bottom.