I rolled a new "matching transformer" since I have more noise these days and I thought my old one may failed. The new one uses a ferrite toroid from a busted PC motherboard. It seems that the noise wasn't coming from there, something else is happening. At least the new transformer seems to work better on higher frequencies (preliminary result of course!) So here are some more charts from Cape Naval. And you'll see what I mean with "blurred" charts.
So this is how a "blurred" one looks like.
Note the "UNCLASSIFIED" notice. (Last word in the text box)
Also a blurred one.
Today they were almost exact regarding start time. But still, if the signal is not strong at your location, it's difficult to catch the start, since the phasing sequence is brief (where is Kyodo, with the nice long tuning session?)