Corrected safety information (!99) · Merge requests · detchar / ligo-channel-lists

Reed Essick requested to merge corrected-safety-information into master Nov 01, 2019

There appear to be several conflicts between the information stored in this repository and the conclusions of the safety studies in LIGO-T1900697 and LIGO-G1901988. Based on discussions on-list, there was a general consensus to adopt the safe/unsafe declarations specified in LIGO-T1900697. It seems like !96 (merged) was supposed to accomplish this, but this is not reflected in the current lists. Rationale for some of the decisions made in !96 (merged) appear to be documented in these two notebooks (here and here). However, only one of them is linked within !96 (merged), making it difficult to reconstruct the logic applied within that merge reqeust.

This begs the question of whether the review procedures currently in place are sufficient to ensure merge requests are accurate and complete.

Specifically, this sanity-check between the final safe/unsafe lists from LIGO-T1900697 to the current master found several issues. These were discussed on the DetChar list and in great depth during the Nov 4 DetChar call (agenda+minutes). For the sake of clarity, I document the decision reached on that call here.

There additionally were disagreements between the channel lists that have been corrected. For example, {H,L}1:OMC-LSC_DITHER_OUT_DQ was labeled safe in the deep channel lists but unsafe in the lldetchar channel lists. All such discrepancies should now be resolved with consistent designations in all lists matching those in LIGO-T1900697.

I have also added several channels to the lldetchar lists that are present in the frames but were not previously included. These are called out in more detail in the comments.

Channels supposed to be marked unsafe that were marked safe in !96 (merged)

All channels declared unsafe in the original pointy-poisson analysis will be labeled unsafe in all channel lists.

I list these channels below for reference, shown with the erroneous safe labels that are corrected within this merge request.

O3/H1-O3-deep.ini
	H1:CAL-PCALX_RX_PD_OUT_DQ 16384 safe clean
	H1:CAL-PCALX_TX_PD_OUT_DQ 16384 safe clean
O3/H1-O3-lldetchar.ini
	H1:ASC-AS_B_RF45_Q_PIT_OUT_DQ 2048 safe clean
	H1:ASC-AS_B_RF45_Q_YAW_OUT_DQ 2048 safe clean
O3/H1-O3-standard.ini
	H1:CAL-PCALX_RX_PD_OUT_DQ 16384 safe clean
	H1:CAL-PCALX_TX_PD_OUT_DQ 16384 safe clean

O3/L1-O3-deep.ini
	L1:CAL-PCALX_RX_PD_OUT_DQ 16384 safe clean
	L1:CAL-PCALX_TX_PD_OUT_DQ 16384 safe clean
O3/L1-O3-lldetchar.ini
	L1:ASC-AS_B_RF45_Q_PIT_OUT_DQ 2048 safe clean
	L1:ASC-AS_B_RF45_Q_YAW_OUT_DQ 2048 safe clean
O3/L1-O3-standard.ini
	L1:CAL-PCALX_RX_PD_OUT_DQ 16384 safe glitchy
	L1:CAL-PCALX_TX_PD_OUT_DQ 16384 safe glitchy

{H1,L1}:ASC-AS_B_RF45_Q_{PIT,YAW}_OUT_DQ are discussed in !96 (merged), with the claim that they were marked unsafe throughout. They seem to have been missed in the lldetchar lists.

The CAL-PCALX channels are not discussed within !96 (merged). It is worth noting that the injections were done with x-arm PCAL (PCALX) at each site and therefore are expected to be strongly statistically correlated with the injections. While this may complicate the safety inference, it certainly does not imply that they are safe. Given that there was no discussion of this on-list when both the hVeto and pointy-poisson groups compared results, the fact that these were not marked unsafe also seems to be an error.

These issues were corrected by direct comparison to the safe/unsafe channel lists stored in LIGO-T1900697 via

for IFO H L
do
    for chan in $(awk '{print $1}' ${IFO}1-raw-unsafe.txt)
    do 
        for list in O3/${IFO}1-O3-*.ini
        do
            if [ -n "$(grep $chan $list | grep ' safe ')" ]
            then
                line=$(grep $chan $list)
                new=$(echo $line | sed -e s,safe,unsafe,)
                sed -i -e "s,$line,$new," $list
            fi
        done
    done
done

Channels marked unsafe that are supposed to be marked safe

Again, there seems to have been no discussion of these channels on-list or within !96 (merged). As far as I can tell, there also seems to have been no discussion about these on-list or in any other venue I can find with the exception of {H1,L1}:OMC-LSC_DITHER_OUT_DQ, which is discussed in LIGO-G1901988 and was mentioned on-list. Both hVeto and pointy-poisson found this channel to be uncorrelated with the injections and declared it safe.

Based on the discussion on the Nov 4 DetChar call, all channels historically marked unsafe that were found to be safe with the current injection set will continue to be labeled unsafe for the time being. I note again that neither hVeto nor pointy-poisson found any of these historically unsafe channels, or even similar channels, to be statistically correlated with the Sept 2019 injections and that the original rationale for declaring these channels unsafe is not readily apparent. The decision to continue labeling these channels unsafe was made for the sake of expidiency given the need to promptly begin offline candidate validation. However, that decision, and the precident it sets, need to be revisited in the immediate future. This corresponds to 39 channels at LHO and 45 channels at LLO.

I list these channels below for reference with the historically motivated unsafe labels that are not altered by this merge request. Please note, LIGO-T1900697 has been updated to reflect the decision to label these channels unsafe based on undocumented historical precident and should still be considered the definitive safety list for O3a.

O3/H1-O3-deep.ini
	H1:CAL-DELTAL_CTRL_DBL_DQ 4096 unsafe clean
	H1:CAL-DELTAL_CTRL_UIM_DBL_DQ 4096 unsafe clean
	H1:LSC-X_TR_A_LF_OUT_DQ 2048 unsafe clean
	H1:OMC-ASC_P1_I_OUT_DQ 2048 unsafe clean
	H1:OMC-ASC_P2_I_OUT_DQ 2048 unsafe clean
	H1:OMC-ASC_Y1_I_OUT_DQ 2048 unsafe clean
	H1:OMC-ASC_Y2_I_OUT_DQ 2048 unsafe clean
	H1:OMC-LSC_SERVO_OUT_DQ 2048 unsafe clean
	H1:PEM-EX_MAG_EBAY_SUSRACK_Z_DQ 8192 unsafe clean
	H1:PEM-EY_MAG_EBAY_SUSRACK_QUAD_SUM_DQ 4096 unsafe clean
	H1:PEM-EY_MAG_EBAY_SUSRACK_X_DQ 8192 unsafe clean
	H1:PEM-EY_MAG_EBAY_SUSRACK_Y_DQ 8192 unsafe clean
	H1:PEM-EY_MAG_EBAY_SUSRACK_Z_DQ 8192 unsafe clean
	H1:PSL-OSC_PD_INT_DC_OUT_DQ 16384 unsafe clean
O3/H1-O3-lldetchar.ini
	H1:OMC-LSC_DITHER_OUT_DQ 16384 unsafe clean
	H1:SUS-ETMY_L1_NOISEMON_LL_OUT_DQ 1024 unsafe clean
	H1:SUS-ETMY_L1_NOISEMON_LR_OUT_DQ 1024 unsafe clean
	H1:SUS-ETMY_L1_NOISEMON_UL_OUT_DQ 1024 unsafe clean
	H1:SUS-ETMY_L1_NOISEMON_UR_OUT_DQ 1024 unsafe clean
	H1:SUS-ETMY_L2_NOISEMON_LL_OUT_DQ 2048 unsafe clean
	H1:SUS-ETMY_L2_NOISEMON_LR_OUT_DQ 2048 unsafe clean
	H1:SUS-ETMY_L2_NOISEMON_UL_OUT_DQ 2048 unsafe clean
	H1:SUS-ETMY_L2_NOISEMON_UR_OUT_DQ 2048 unsafe clean
	H1:SUS-ITMX_L1_NOISEMON_LL_OUT_DQ 1024 unsafe clean
	H1:SUS-ITMX_L1_NOISEMON_LR_OUT_DQ 1024 unsafe clean
	H1:SUS-ITMX_L1_NOISEMON_UL_OUT_DQ 1024 unsafe clean
	H1:SUS-ITMX_L1_NOISEMON_UR_OUT_DQ 1024 unsafe clean
	H1:SUS-ITMX_L2_NOISEMON_LL_OUT_DQ 2048 unsafe clean
	H1:SUS-ITMX_L2_NOISEMON_LR_OUT_DQ 2048 unsafe clean
	H1:SUS-ITMX_L2_NOISEMON_UL_OUT_DQ 2048 unsafe clean
	H1:SUS-ITMX_L2_NOISEMON_UR_OUT_DQ 2048 unsafe clean
	H1:SUS-ITMY_L1_NOISEMON_LL_OUT_DQ 1024 unsafe clean
	H1:SUS-ITMY_L1_NOISEMON_LR_OUT_DQ 1024 unsafe clean
	H1:SUS-ITMY_L1_NOISEMON_UL_OUT_DQ 1024 unsafe clean
	H1:SUS-ITMY_L1_NOISEMON_UR_OUT_DQ 1024 unsafe clean
	H1:SUS-ITMY_L2_NOISEMON_LL_OUT_DQ 2048 unsafe clean
	H1:SUS-ITMY_L2_NOISEMON_LR_OUT_DQ 2048 unsafe clean
	H1:SUS-ITMY_L2_NOISEMON_UL_OUT_DQ 2048 unsafe clean
	H1:SUS-ITMY_L2_NOISEMON_UR_OUT_DQ 2048 unsafe clean
O3/H1-O3-pem.ini
	H1:PEM-EX_MAG_EBAY_SUSRACK_Z_DQ 8192 unsafe clean
	H1:PEM-EY_MAG_EBAY_SUSRACK_QUAD_SUM_DQ 4096 unsafe clean
	H1:PEM-EY_MAG_EBAY_SUSRACK_X_DQ 8192 unsafe clean
	H1:PEM-EY_MAG_EBAY_SUSRACK_Y_DQ 8192 unsafe clean
	H1:PEM-EY_MAG_EBAY_SUSRACK_Z_DQ 8192 unsafe clean
O3/H1-O3-standard.ini
	H1:LSC-X_TR_A_LF_OUT_DQ 2048 unsafe clean
	H1:OMC-ASC_P1_I_OUT_DQ 2048 unsafe clean
	H1:OMC-ASC_P2_I_OUT_DQ 2048 unsafe clean
	H1:OMC-ASC_Y1_I_OUT_DQ 2048 unsafe clean
	H1:OMC-ASC_Y2_I_OUT_DQ 2048 unsafe clean
	H1:OMC-LSC_SERVO_OUT_DQ 2048 unsafe clean
	H1:PEM-EX_MAG_EBAY_SUSRACK_Z_DQ 8192 unsafe clean
	H1:PEM-EY_MAG_EBAY_SUSRACK_QUAD_SUM_DQ 4096 unsafe clean
	H1:PEM-EY_MAG_EBAY_SUSRACK_X_DQ 8192 unsafe clean
	H1:PEM-EY_MAG_EBAY_SUSRACK_Y_DQ 8192 unsafe clean
	H1:PEM-EY_MAG_EBAY_SUSRACK_Z_DQ 8192 unsafe clean
	H1:PSL-OSC_PD_INT_DC_OUT_DQ 16384 unsafe clean

O3/L1-O3-deep.ini
	L1:ASC-AS_A_DC_SUM_OUT_DQ 2048 unsafe clean
	L1:ASC-AS_B_DC_SUM_OUT_DQ 2048 unsafe clean
	L1:CAL-DELTAL_CTRL_DBL_DQ 4096 unsafe clean
	L1:CAL-DELTAL_CTRL_UIM_DBL_DQ 4096 unsafe clean
	L1:LSC-X_TR_A_LF_OUT_DQ 2048 unsafe clean
	L1:OMC-ASC_P1_I_OUT_DQ 2048 unsafe clean
	L1:OMC-ASC_P1_Q_OUT_DQ 2048 unsafe clean
	L1:OMC-ASC_P2_I_OUT_DQ 2048 unsafe clean
	L1:OMC-ASC_P2_Q_OUT_DQ 2048 unsafe clean
	L1:OMC-ASC_Y1_I_OUT_DQ 2048 unsafe clean
	L1:OMC-ASC_Y1_Q_OUT_DQ 2048 unsafe clean
	L1:OMC-ASC_Y2_I_OUT_DQ 2048 unsafe clean
	L1:OMC-ASC_Y2_Q_OUT_DQ 2048 unsafe clean
	L1:OMC-LSC_SERVO_OUT_DQ 2048 unsafe clean
	L1:PEM-EX_VMON_ETMX_ESDPOWER18_DQ 4096 unsafe clean
	L1:PEM-EY_MAG_EBAY_SUSRACK_QUAD_SUM_DQ 4096 unsafe clean
	L1:PEM-EY_MAG_EBAY_SUSRACK_QUAD_SUM_DQ 4096 unsafe clean
	L1:PEM-EY_MAG_EBAY_SUSRACK_X_DQ 8192 unsafe clean
	L1:PEM-EY_MAG_EBAY_SUSRACK_Y_DQ 8192 unsafe clean
	L1:PEM-EY_MAG_EBAY_SUSRACK_Z_DQ 8192 unsafe clean
	L1:PSL-OSC_PD_INT_DC_OUT_DQ 16384 unsafe clean
O3/L1-O3-lldetchar.ini
	L1:OMC-LSC_DITHER_OUT_DQ 16384 unsafe clean
	L1:SUS-ETMY_L1_NOISEMON_LL_OUT_DQ 1024 unsafe clean
	L1:SUS-ETMY_L1_NOISEMON_LR_OUT_DQ 1024 unsafe clean
	L1:SUS-ETMY_L1_NOISEMON_UL_OUT_DQ 1024 unsafe clean
	L1:SUS-ETMY_L1_NOISEMON_UR_OUT_DQ 1024 unsafe clean
	L1:SUS-ETMY_L2_NOISEMON_LL_OUT_DQ 2048 unsafe clean
	L1:SUS-ETMY_L2_NOISEMON_LR_OUT_DQ 2048 unsafe clean
	L1:SUS-ETMY_L2_NOISEMON_UL_OUT_DQ 2048 unsafe clean
	L1:SUS-ETMY_L2_NOISEMON_UR_OUT_DQ 2048 unsafe clean
	L1:SUS-ITMX_L1_NOISEMON_LL_OUT_DQ 1024 unsafe clean
	L1:SUS-ITMX_L1_NOISEMON_LR_OUT_DQ 1024 unsafe clean
	L1:SUS-ITMX_L1_NOISEMON_UL_OUT_DQ 1024 unsafe clean
	L1:SUS-ITMX_L1_NOISEMON_UR_OUT_DQ 1024 unsafe clean
	L1:SUS-ITMX_L2_NOISEMON_LL_OUT_DQ 2048 unsafe clean
	L1:SUS-ITMX_L2_NOISEMON_LR_OUT_DQ 2048 unsafe clean
	L1:SUS-ITMX_L2_NOISEMON_UL_OUT_DQ 2048 unsafe clean
	L1:SUS-ITMX_L2_NOISEMON_UR_OUT_DQ 2048 unsafe clean
	L1:SUS-ITMY_L1_NOISEMON_LL_OUT_DQ 1024 unsafe clean
	L1:SUS-ITMY_L1_NOISEMON_LR_OUT_DQ 1024 unsafe clean
	L1:SUS-ITMY_L1_NOISEMON_UL_OUT_DQ 1024 unsafe clean
	L1:SUS-ITMY_L1_NOISEMON_UR_OUT_DQ 1024 unsafe clean
	L1:SUS-ITMY_L2_NOISEMON_LL_OUT_DQ 2048 unsafe clean
	L1:SUS-ITMY_L2_NOISEMON_LR_OUT_DQ 2048 unsafe clean
	L1:SUS-ITMY_L2_NOISEMON_UL_OUT_DQ 2048 unsafe clean
	L1:SUS-ITMY_L2_NOISEMON_UR_OUT_DQ 2048 unsafe clean
O3/L1-O3-pem.ini
	L1:PEM-EX_VMON_ETMX_ESDPOWER18_DQ 4096 unsafe clean
	L1:PEM-EY_MAG_EBAY_SUSRACK_QUAD_SUM_DQ 4096 unsafe clean
	L1:PEM-EY_MAG_EBAY_SUSRACK_QUAD_SUM_DQ 4096 unsafe clean
	L1:PEM-EY_MAG_EBAY_SUSRACK_X_DQ 8192 unsafe clean
	L1:PEM-EY_MAG_EBAY_SUSRACK_Y_DQ 8192 unsafe clean
	L1:PEM-EY_MAG_EBAY_SUSRACK_Z_DQ 8192 unsafe clean
O3/L1-O3-standard.ini
	L1:ASC-AS_A_DC_SUM_OUT_DQ 2048 unsafe clean
	L1:ASC-AS_B_DC_SUM_OUT_DQ 2048 unsafe clean
	L1:OMC-ASC_P1_I_OUT_DQ 2048 unsafe clean
	L1:OMC-ASC_P2_I_OUT_DQ 2048 unsafe clean
	L1:OMC-ASC_Y1_I_OUT_DQ 2048 unsafe clean
	L1:OMC-ASC_Y2_I_OUT_DQ 2048 unsafe clean
	L1:OMC-LSC_SERVO_OUT_DQ 2048 unsafe clean
	L1:PEM-EX_VMON_ETMX_ESDPOWER18_DQ 4096 unsafe clean
	L1:PEM-EY_MAG_EBAY_SUSRACK_QUAD_SUM_DQ 4096 unsafe clean
	L1:PEM-EY_MAG_EBAY_SUSRACK_QUAD_SUM_DQ 4096 unsafe clean
	L1:PEM-EY_MAG_EBAY_SUSRACK_X_DQ 8192 unsafe clean
	L1:PEM-EY_MAG_EBAY_SUSRACK_Y_DQ 8192 unsafe clean
	L1:PEM-EY_MAG_EBAY_SUSRACK_Z_DQ 8192 unsafe clean

related merge requests

cc/ @derek.davis @thomas-massinger @jessica.mciver @laura-nuttall @geoffrey.mo @erik-katsavounidis @joseph-areeda @joshua-smith @keith-riles

Edited Nov 05, 2019 by Reed Essick

Admin message

Corrected safety information

Merge request reports