For the purpose of injecting PCAL lines, there are two channels that can be used: H1:CAL-PCALY_RX_PD_OUT_DQ (for reflected light) and H1:CAL-PCALY_TX_PD_OUT_DQ (for transmitted light). At various points during O2, we switched off between these two channels in C00 and C01 data at Livingston to account for things like PCAL clipping in the RX channel. The exact dates of each change are listed below, as documented on the filter configurations wiki page.
Version C00
Used RX from 116440990 (2016 Sep 14, 16:43:07 UTC) -- 1182618646 (2017 Jun 27, 17:10:28 UTC)
Used TX from 1182618646 (2017 Jun 27, 17:10:28 UTC) -- now
Version C01
Used RX from 1163173888 (2016 Nov 14, 15:51:11 UTC) -- 1181941760 (2017 Jun 19, 21:09:02 UTC)
Used TX from 1181941760 (2017 Jun 19, 21:09:02 UTC) -- now
The corresponding record over at Livingston is in aLOG 35948.
Attached plot show the ratio of TX PD to RX PD during O2. The size of deviations from one in that plot, when RX PD is used, indicates the systematic uncertainty introduced by the PCal clipping. The circled regions show the times when RX PD was used and the ratio was not close to one. The systematic uncertainties introduced by the clipping during those time is <3 %. This systematic would be reduced in the C02 frames. Each data point in that plot is running average of an hour of data. We plot data only during the observation times and with hoft ok state.