Satoshi and Kiwamu,
Related log: 17188
We have swapped an AA unit in the ALS0 rack with a spare, today.
Channel 25 in the old unit showed an anomalously high noise level as reported in 17188. This channel by chance had been used for measuring the phase error for the X PLL. Although we didn't think this is related to the glitch events that we have been trying to hunt, we replaced the unit in order to sort out existing issues.
Resulting noise:
See the screenshot below. The discrepancy between the two channels (15063, Sep. 2020) are now much smaller. We confirmed that the PLL X guardian can still bring it to the locked state.
A caveat:
Even though the new unit has been thoroughly tested by Shimode san by measuring the transfer functions of all the channels, channel 31 showed a high noise peak at around 1 kHz dominating the rms. The screenshot below shows the spectrum of channel 31 without connecting the signals. The other neighboring channels are also plotted for comparison. Fortunately, this channel is not assigned to any signals. Therefore we will keep using this unit.
