Attached are the matching ratio plots for several runs before the HV change combined ( mostly day345) and after the HV change (mostly day 347).
I didn't check which runs are "good" (all modules active) or "bad" (missing modules, Get4s dropping out), before I combined the statistics...

For example in several runs of day347 sectors 15, 17 and 23 had troubles with dropping out Get4s, and sector 24 had one module missing. That explains there lower matching rations.
Overall the matching ratio in the other sectors increased compared to before the HV change.

I combined the "good" Tsignhua sectors (16, 21, 22 ) and USTC sectors ( 13, 14, 18, 19, 20 ) in a second plot.
The two curves are now basically identical and level out at a matching ratio of ~85%.

I'll redo the analysis selecting runs in which also the other sectors didn't have problems.