jwst-gc pipeline monitor

Stage ladder

Products per filter

* = the product name carries no _o<obs> token and this field has more than one observation, so the count cannot be attributed to 4147/o012.

Astrometry — m2 checkpoint

“tiles ok” is not a tolerance. measure_offset_grid runs with no max_off_mas, and astrometry_offsets sets off_ok=True whenever that is None — so N/N counts tiles whose offset histogram had a coherent peak, however large the offset. The column that carries the gate is worst tile, against 15 mas.

Provenance

Queue

No jobs in the queue for this field.

Findings

• fail
  F115W_o012: 10/48 exposures misaligned vs their visit consensus

  An exposure is misaligned when its offset exceeds 2.0 mas AND is significant against the peak error bars. At m2 this CORRECTS the offsets table and stops the run; the crf frames must be regenerated before cataloging continues.

  checkpoint_m2_F115W_o012_latest.json

  what is affected, and why

  Affected: 5/8 detectors — spread across detectors, so the frame as a whole moved: look at the offsets table and whether these exposures were regenerated from _cal after the table changed. fix_alignment SKIPS a frame that already has a RAOFFSET header, so correcting the table alone leaves stale frames in place; regenerate the working copy from _cal instead of re-applying on top.

  Misaligned by detector

  nrca23/6

  nrcb33/6

  nrca42/6

  nrca11/6

  nrca31/6

  nrcb10/6

  nrcb20/6

  nrcb40/6

  Diagnostic writeupD2 — internal astrometric repeatability · full diagnostic writeup (PDF)

  visit	detector	dRA (mas)	dDec (mas)	off (mas)	contrast	pairs	window ("")
  1	nrca1	0.4	-2.6	2.6	123.0	5368	10.0
  1	nrca2	-3.1	-1.7	3.5	178.0	7918	10.0
  1	nrca2	-2.6	-3.4	4.2	180.0	7832	10.0
  1	nrca2	1.6	1.9	2.5	320.0	17850	10.0
  1	nrca3	-1.7	-2.1	2.6	286.0	18024	10.0
  1	nrca4	-0.8	2.0	2.2	147.0	7427	10.0
  1	nrca4	-1.8	1.7	2.5	161.0	6632	10.0
  1	nrcb3	0.7	-2.1	2.2	401.0	62854	10.0
  1	nrcb3	0.2	3.0	3.0	248.0	27296	10.0
  1	nrcb3	-0.5	4.1	4.1	282.0	24192	10.0

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• fail
  F162M_o012: 24/48 exposures misaligned vs their visit consensus

  An exposure is misaligned when its offset exceeds 2.0 mas AND is significant against the peak error bars. At m2 this CORRECTS the offsets table and stops the run; the crf frames must be regenerated before cataloging continues.

  checkpoint_m2_F162M_o012_latest.json

  what is affected, and why

  Affected: 8/8 detectors — spread across detectors, so the frame as a whole moved: look at the offsets table and whether these exposures were regenerated from _cal after the table changed. fix_alignment SKIPS a frame that already has a RAOFFSET header, so correcting the table alone leaves stale frames in place; regenerate the working copy from _cal instead of re-applying on top.

  Misaligned by detector

  nrca25/6

  nrca34/6

  nrca44/6

  nrca13/6

  nrcb13/6

  nrcb23/6

  nrcb31/6

  nrcb41/6

  Diagnostic writeupD2 — internal astrometric repeatability · full diagnostic writeup (PDF)

  visit	detector	dRA (mas)	dDec (mas)	off (mas)	contrast	pairs	window ("")
  1	nrca1	-0.5	-2.5	2.6	643.7	207124	3.0
  1	nrca1	-0.6	-3.2	3.2	675.0	212378	3.0
  1	nrca1	0.2	-2.3	2.3	949.5	142449	3.0
  1	nrca2	-1.3	-2.5	2.8	1,191.7	236333	3.0
  1	nrca2	2.6	-4.1	4.9	1,017.2	334552	3.0
  1	nrca2	5.5	-3.7	6.6	944.8	407895	3.0
  1	nrca2	5.5	-3.2	6.3	1,213.2	407863	3.0
  1	nrca2	2.6	-2.5	3.6	1,078.6	340840	3.0
  1	nrca3	-2.2	4.1	4.6	998.2	450317	3.0
  1	nrca3	-1.9	2.6	3.3	945.7	423595	3.0
  1	nrca3	1.6	1.7	2.4	517.8	277793	3.0
  1	nrca3	1.4	1.6	2.1	486.8	265157	3.0
  1	nrca4	0.4	3.0	3.1	970.0	206606	3.0
  1	nrca4	0.4	2.8	2.8	891.0	262742	3.0
  1	nrca4	-0.8	2.8	2.9	632.7	207244	3.0
  1	nrca4	-0.8	2.6	2.7	486.0	182542	3.0
  1	nrcb1	0.0	5.7	5.7	439.9	763348	3.0
  1	nrcb1	-1.1	3.3	3.4	458.7	867947	3.0
  1	nrcb1	-1.1	3.6	3.7	550.6	898006	3.0
  1	nrcb2	1.0	5.7	5.8	962.5	423231	3.0
  1	nrcb2	-0.6	2.7	2.8	679.6	384145	3.0
  1	nrcb2	-1.2	3.2	3.4	749.0	441196	3.0
  1	nrcb3	0.2	-4.5	4.5	444.5	808431	3.0
  1	nrcb4	-0.8	-4.7	4.8	870.4	405944	3.0

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• fail
  F182M_o012: 7/48 exposures misaligned vs their visit consensus

  An exposure is misaligned when its offset exceeds 2.0 mas AND is significant against the peak error bars. At m2 this CORRECTS the offsets table and stops the run; the crf frames must be regenerated before cataloging continues.

  checkpoint_m2_F182M_o012_latest.json

  what is affected, and why

  Affected: 5/8 detectors — spread across detectors, so the frame as a whole moved: look at the offsets table and whether these exposures were regenerated from _cal after the table changed. fix_alignment SKIPS a frame that already has a RAOFFSET header, so correcting the table alone leaves stale frames in place; regenerate the working copy from _cal instead of re-applying on top.

  Misaligned by detector

  nrca12/6

  nrcb32/6

  nrca21/6

  nrcb11/6

  nrcb21/6

  nrca30/6

  nrca40/6

  nrcb40/6

  Diagnostic writeupD2 — internal astrometric repeatability · full diagnostic writeup (PDF)

  visit	detector	dRA (mas)	dDec (mas)	off (mas)	contrast	pairs	window ("")
  1	nrca1	-1.3	-1.7	2.2	730.3	699713	3.0
  1	nrca1	1.3	1.6	2.0	814.1	517232	3.0
  1	nrca2	-1.4	-2.0	2.4	704.9	792404	3.0
  1	nrcb1	-0.9	1.8	2.0	510.1	1454714	3.0
  1	nrcb2	1.4	-1.9	2.3	697.8	918722	3.0
  1	nrcb3	0.3	-2.0	2.0	352.6	1406580	3.0
  1	nrcb3	-0.3	2.5	2.6	491.3	1002743	3.0

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• fail
  F212N_o012: 10/48 exposures misaligned vs their visit consensus

  An exposure is misaligned when its offset exceeds 2.0 mas AND is significant against the peak error bars. At m2 this CORRECTS the offsets table and stops the run; the crf frames must be regenerated before cataloging continues.

  checkpoint_m2_F212N_o012_latest.json

  what is affected, and why

  Affected: 6/8 detectors — spread across detectors, so the frame as a whole moved: look at the offsets table and whether these exposures were regenerated from _cal after the table changed. fix_alignment SKIPS a frame that already has a RAOFFSET header, so correcting the table alone leaves stale frames in place; regenerate the working copy from _cal instead of re-applying on top.

  Misaligned by detector

  nrca23/6

  nrcb22/6

  nrcb32/6

  nrca11/6

  nrca41/6

  nrcb11/6

  nrca30/6

  nrcb40/6

  Diagnostic writeupD2 — internal astrometric repeatability · full diagnostic writeup (PDF)

  visit	detector	dRA (mas)	dDec (mas)	off (mas)	contrast	pairs	window ("")
  1	nrca1	-1.2	-2.1	2.4	1,202.3	225004	3.0
  1	nrca2	-0.9	-1.9	2.1	1,142.3	209792	3.0
  1	nrca2	0.2	2.2	2.2	821.5	296588	3.0
  1	nrca2	-0.3	2.7	2.7	936.0	287901	3.0
  1	nrca4	-0.2	-2.2	2.3	897.7	210906	3.0
  1	nrcb1	0.4	-2.4	2.5	678.4	492169	3.0
  1	nrcb2	0.1	-2.1	2.1	654.6	377350	3.0
  1	nrcb2	1.3	-3.2	3.4	1,053.4	340127	3.0
  1	nrcb3	0.3	2.1	2.1	778.6	401703	3.0
  1	nrcb3	-0.1	2.8	2.8	616.6	356846	3.0

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• fail
  F405N_o012: 2/12 exposures misaligned vs their visit consensus

  An exposure is misaligned when its offset exceeds 2.0 mas AND is significant against the peak error bars. At m2 this CORRECTS the offsets table and stops the run; the crf frames must be regenerated before cataloging continues.

  checkpoint_m2_F405N_o012_latest.json

  what is affected, and why

  Affected: 1/2 detectors — confined to nrcalong, which points at a detector-local defect (distortion reference, a bad filteroffset, or one chip's frames never re-aligned) rather than a whole-frame misalignment. fix_alignment SKIPS a frame that already has a RAOFFSET header, so correcting the table alone leaves stale frames in place; regenerate the working copy from _cal instead of re-applying on top.

  Misaligned by detector

  nrcalong2/6

  nrcblong0/6

  Diagnostic writeupD2 — internal astrometric repeatability · full diagnostic writeup (PDF)

  visit	detector	dRA (mas)	dDec (mas)	off (mas)	contrast	pairs	window ("")
  1	nrcalong	-6.9	-0.2	6.9	3,861.0	93496	3.0
  1	nrcalong	-6.9	-0.6	7.0	2,516.5	97230	3.0

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• fail
  F470N_o012: 9/12 exposures misaligned vs their visit consensus

  An exposure is misaligned when its offset exceeds 2.0 mas AND is significant against the peak error bars. At m2 this CORRECTS the offsets table and stops the run; the crf frames must be regenerated before cataloging continues.

  checkpoint_m2_F470N_o012_latest.json

  what is affected, and why

  Affected: 2/2 detectors — spread across detectors, so the frame as a whole moved: look at the offsets table and whether these exposures were regenerated from _cal after the table changed. fix_alignment SKIPS a frame that already has a RAOFFSET header, so correcting the table alone leaves stale frames in place; regenerate the working copy from _cal instead of re-applying on top.

  Misaligned by detector

  nrcalong5/6

  nrcblong4/6

  Diagnostic writeupD2 — internal astrometric repeatability · full diagnostic writeup (PDF)

  visit	detector	dRA (mas)	dDec (mas)	off (mas)	contrast	pairs	window ("")
  1	nrcalong	-0.8	-2.4	2.6	2,841.0	120915	3.0
  1	nrcalong	-0.7	-3.4	3.5	2,822.5	120677	3.0
  1	nrcalong	-3.3	-0.1	3.3	2,101.5	127122	3.0
  1	nrcalong	-2.4	-0.1	2.4	2,023.5	121060	3.0
  1	nrcalong	5.0	-0.1	5.0	1,974.5	121949	3.0
  1	nrcblong	-2.5	1.4	2.8	3,954.0	164428	3.0
  1	nrcblong	-3.6	2.9	4.6	4,103.0	154070	3.0
  1	nrcblong	3.7	0.1	3.7	2,323.5	154965	3.0
  1	nrcblong	1.2	2.0	2.3	3,845.0	165499	3.0

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  m7: 27 product(s) stamped with a -dirty tag

  The tree had uncommitted changes, so the tag does not identify the code that ran. Fine for a development run; not for a release.

  catalogs/*m7*.prov.json

  what is affected, and why

  Diagnostic writeupfull diagnostic writeup (PDF)

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  m8: 3 product(s) stamped with a -dirty tag

  The tree had uncommitted changes, so the tag does not identify the code that ran. Fine for a development run; not for a release.

  catalogs/*m8*.prov.json

  what is affected, and why

  Diagnostic writeupfull diagnostic writeup (PDF)

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F115W: 13/48 exposures misaligned vs their visit consensus (checkpoint predates the frames)

  An exposure is misaligned when its offset exceeds 2.0 mas AND is significant against the peak error bars. At m2 this CORRECTS the offsets table and stops the run; the crf frames must be regenerated before cataloging continues. NOTE: the reduced frames are NEWER than this checkpoint, so the correction it triggered has probably already been applied — re-run the m2 checkpoint to confirm rather than treating this as current.

  checkpoint_m2_F115W_latest.json

  what is affected, and why

  Affected: 7/8 detectors — spread across detectors, so the frame as a whole moved: look at the offsets table and whether these exposures were regenerated from _cal after the table changed. fix_alignment SKIPS a frame that already has a RAOFFSET header, so correcting the table alone leaves stale frames in place; regenerate the working copy from _cal instead of re-applying on top.

  Misaligned by detector

  nrca34/6

  nrca22/6

  nrca42/6

  nrcb32/6

  nrca11/6

  nrcb11/6

  nrcb21/6

  nrcb40/6

  Diagnostic writeupD2 — internal astrometric repeatability · full diagnostic writeup (PDF)

  visit	detector	dRA (mas)	dDec (mas)	off (mas)	contrast	pairs	window ("")
  1	nrca1	0.8	-2.0	2.2	146.0	4808	10.0
  1	nrca2	-3.0	-1.4	3.3	161.0	7405	10.0
  1	nrca2	-2.0	-2.4	3.1	154.0	7323	10.0
  1	nrca3	-2.6	-0.3	2.6	203.0	17288	10.0
  1	nrca3	-2.3	-1.8	2.9	311.0	16739	10.0
  1	nrca3	1.6	1.7	2.3	175.0	8029	10.0
  1	nrca3	1.1	2.2	2.5	162.0	8868	10.0
  1	nrca4	-1.2	2.6	2.9	181.0	6926	10.0
  1	nrca4	-2.0	2.6	3.3	194.0	6256	10.0
  1	nrcb1	-0.9	1.9	2.1	570.0	79130	10.0
  1	nrcb2	0.2	-2.2	2.2	194.0	14911	10.0
  1	nrcb3	0.7	2.2	2.3	333.0	25888	10.0
  1	nrcb3	-0.1	3.2	3.2	219.0	22700	10.0

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F162M: 20/48 exposures misaligned vs their visit consensus (checkpoint predates the frames)

  An exposure is misaligned when its offset exceeds 2.0 mas AND is significant against the peak error bars. At m2 this CORRECTS the offsets table and stops the run; the crf frames must be regenerated before cataloging continues. NOTE: the reduced frames are NEWER than this checkpoint, so the correction it triggered has probably already been applied — re-run the m2 checkpoint to confirm rather than treating this as current.

  checkpoint_m2_F162M_latest.json

  what is affected, and why

  Affected: 8/8 detectors — spread across detectors, so the frame as a whole moved: look at the offsets table and whether these exposures were regenerated from _cal after the table changed. fix_alignment SKIPS a frame that already has a RAOFFSET header, so correcting the table alone leaves stale frames in place; regenerate the working copy from _cal instead of re-applying on top.

  Misaligned by detector

  nrca14/6

  nrca44/6

  nrcb13/6

  nrcb23/6

  nrca22/6

  nrca32/6

  nrcb31/6

  nrcb41/6

  Diagnostic writeupD2 — internal astrometric repeatability · full diagnostic writeup (PDF)

  visit	detector	dRA (mas)	dDec (mas)	off (mas)	contrast	pairs	window ("")
  1	nrca1	-0.3	-2.5	2.5	567.3	212377	3.0
  1	nrca1	-0.2	-3.4	3.4	576.3	217869	3.0
  1	nrca1	0.6	-2.5	2.6	1,319.0	146884	3.0
  1	nrca1	0.7	-2.0	2.1	1,225.5	163027	3.0
  1	nrca2	-2.5	0.6	2.5	861.3	236078	3.0
  1	nrca2	-2.6	-0.8	2.7	805.0	242369	3.0
  1	nrca3	-2.1	3.4	4.0	1,004.0	459382	3.0
  1	nrca3	-1.8	2.0	2.7	955.5	432954	3.0
  1	nrca4	0.6	3.0	3.1	1,173.3	213530	3.0
  1	nrca4	0.8	2.8	2.9	1,133.2	272163	3.0
  1	nrca4	-1.4	3.7	3.9	1,025.0	214714	3.0
  1	nrca4	-1.3	3.5	3.7	811.0	189749	3.0
  1	nrcb1	0.1	6.5	6.5	414.8	777311	3.0
  1	nrcb1	-0.4	3.0	3.1	372.9	878785	3.0
  1	nrcb1	-0.6	3.3	3.4	463.9	904088	3.0
  1	nrcb2	0.6	4.9	5.0	918.0	433658	3.0
  1	nrcb2	-0.1	2.3	2.3	409.0	388858	3.0
  1	nrcb2	-0.8	2.8	2.9	560.2	449618	3.0
  1	nrcb3	0.2	-5.5	5.5	465.2	810721	3.0
  1	nrcb4	-0.9	-5.9	6.0	830.0	414029	3.0

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F182M: 7/48 exposures misaligned vs their visit consensus (checkpoint predates the frames)

  An exposure is misaligned when its offset exceeds 2.0 mas AND is significant against the peak error bars. At m2 this CORRECTS the offsets table and stops the run; the crf frames must be regenerated before cataloging continues. NOTE: the reduced frames are NEWER than this checkpoint, so the correction it triggered has probably already been applied — re-run the m2 checkpoint to confirm rather than treating this as current.

  checkpoint_m2_F182M_latest.json

  what is affected, and why

  Affected: 6/8 detectors — spread across detectors, so the frame as a whole moved: look at the offsets table and whether these exposures were regenerated from _cal after the table changed. fix_alignment SKIPS a frame that already has a RAOFFSET header, so correcting the table alone leaves stale frames in place; regenerate the working copy from _cal instead of re-applying on top.

  Misaligned by detector

  nrca12/6

  nrca21/6

  nrca31/6

  nrca41/6

  nrcb11/6

  nrcb41/6

  nrcb20/6

  nrcb30/6

  Diagnostic writeupD2 — internal astrometric repeatability · full diagnostic writeup (PDF)

  visit	detector	dRA (mas)	dDec (mas)	off (mas)	contrast	pairs	window ("")
  1	nrca1	-3.3	3.4	4.7	803.1	721540	3.0
  1	nrca1	1.8	-1.5	2.4	906.4	532685	3.0
  1	nrca2	-2.7	3.2	4.1	759.5	816888	3.0
  1	nrca3	-0.9	3.3	3.4	667.8	907831	3.0
  1	nrca4	-1.8	3.0	3.5	821.1	732631	3.0
  1	nrcb1	-1.4	2.9	3.2	609.4	1269002	3.0
  1	nrcb4	-1.5	3.2	3.6	690.8	921462	3.0

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F212N: 15/48 exposures misaligned vs their visit consensus (checkpoint predates the frames)

  An exposure is misaligned when its offset exceeds 2.0 mas AND is significant against the peak error bars. At m2 this CORRECTS the offsets table and stops the run; the crf frames must be regenerated before cataloging continues. NOTE: the reduced frames are NEWER than this checkpoint, so the correction it triggered has probably already been applied — re-run the m2 checkpoint to confirm rather than treating this as current.

  checkpoint_m2_F212N_latest.json

  what is affected, and why

  Affected: 8/8 detectors — spread across detectors, so the frame as a whole moved: look at the offsets table and whether these exposures were regenerated from _cal after the table changed. fix_alignment SKIPS a frame that already has a RAOFFSET header, so correcting the table alone leaves stale frames in place; regenerate the working copy from _cal instead of re-applying on top.

  Misaligned by detector

  nrca22/6

  nrca32/6

  nrca42/6

  nrcb12/6

  nrcb22/6

  nrcb32/6

  nrcb42/6

  nrca11/6

  Diagnostic writeupD2 — internal astrometric repeatability · full diagnostic writeup (PDF)

  visit	detector	dRA (mas)	dDec (mas)	off (mas)	contrast	pairs	window ("")
  1	nrca1	-1.5	4.3	4.5	1,298.0	222800	3.0
  1	nrca2	-1.1	4.9	5.0	1,221.7	206817	3.0
  1	nrca2	0.6	2.2	2.3	904.5	266053	3.0
  1	nrca3	-0.1	3.5	3.5	641.4	340141	3.0
  1	nrca3	-0.9	-4.0	4.1	1,202.3	259852	3.0
  1	nrca4	-0.8	3.2	3.3	1,153.7	207670	3.0
  1	nrca4	-0.7	-3.7	3.8	1,522.5	172699	3.0
  1	nrcb1	-0.2	2.9	2.9	577.9	489742	3.0
  1	nrcb1	-1.6	-3.6	3.9	892.2	591272	3.0
  1	nrcb2	0.9	2.2	2.3	986.0	338491	3.0
  1	nrcb2	-1.4	-3.7	3.9	1,000.6	376687	3.0
  1	nrcb3	-0.3	3.8	3.8	575.5	553436	3.0
  1	nrcb3	0.8	2.0	2.1	818.6	484785	3.0
  1	nrcb4	-0.3	4.9	4.9	827.8	322522	3.0
  1	nrcb4	0.9	1.9	2.0	1,031.0	320715	3.0

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F360M: 1/12 exposures misaligned vs their visit consensus (checkpoint predates the frames)

  An exposure is misaligned when its offset exceeds 2.0 mas AND is significant against the peak error bars. At m2 this CORRECTS the offsets table and stops the run; the crf frames must be regenerated before cataloging continues. NOTE: the reduced frames are NEWER than this checkpoint, so the correction it triggered has probably already been applied — re-run the m2 checkpoint to confirm rather than treating this as current.

  checkpoint_m2_F360M_latest.json

  what is affected, and why

  Affected: 1/2 detectors — confined to nrcalong, which points at a detector-local defect (distortion reference, a bad filteroffset, or one chip's frames never re-aligned) rather than a whole-frame misalignment. fix_alignment SKIPS a frame that already has a RAOFFSET header, so correcting the table alone leaves stale frames in place; regenerate the working copy from _cal instead of re-applying on top.

  Misaligned by detector

  nrcalong1/6

  nrcblong0/6

  Diagnostic writeupD2 — internal astrometric repeatability · full diagnostic writeup (PDF)

  visit	detector	dRA (mas)	dDec (mas)	off (mas)	contrast	pairs	window ("")
  1	nrcalong	0.6	2.4	2.5	2,225.5	100310	3.0

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F470N: 5/12 exposures misaligned vs their visit consensus (checkpoint predates the frames)

  An exposure is misaligned when its offset exceeds 2.0 mas AND is significant against the peak error bars. At m2 this CORRECTS the offsets table and stops the run; the crf frames must be regenerated before cataloging continues. NOTE: the reduced frames are NEWER than this checkpoint, so the correction it triggered has probably already been applied — re-run the m2 checkpoint to confirm rather than treating this as current.

  checkpoint_m2_F470N_latest.json

  what is affected, and why

  Affected: 2/2 detectors — spread across detectors, so the frame as a whole moved: look at the offsets table and whether these exposures were regenerated from _cal after the table changed. fix_alignment SKIPS a frame that already has a RAOFFSET header, so correcting the table alone leaves stale frames in place; regenerate the working copy from _cal instead of re-applying on top.

  Misaligned by detector

  nrcalong3/6

  nrcblong2/6

  Diagnostic writeupD2 — internal astrometric repeatability · full diagnostic writeup (PDF)

  visit	detector	dRA (mas)	dDec (mas)	off (mas)	contrast	pairs	window ("")
  1	nrcalong	1.5	-4.0	4.2	3,298.5	121366	3.0
  1	nrcalong	1.7	-3.4	3.8	3,358.0	122354	3.0
  1	nrcalong	-0.1	-2.9	2.9	2,013.0	126633	3.0
  1	nrcblong	1.5	2.9	3.2	3,951.0	161728	3.0
  1	nrcblong	2.2	2.9	3.6	4,103.0	160350	3.0

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F480M: 1/12 exposures misaligned vs their visit consensus (checkpoint predates the frames)

  An exposure is misaligned when its offset exceeds 2.0 mas AND is significant against the peak error bars. At m2 this CORRECTS the offsets table and stops the run; the crf frames must be regenerated before cataloging continues. NOTE: the reduced frames are NEWER than this checkpoint, so the correction it triggered has probably already been applied — re-run the m2 checkpoint to confirm rather than treating this as current.

  checkpoint_m2_F480M_latest.json

  what is affected, and why

  Affected: 1/2 detectors — confined to nrcblong, which points at a detector-local defect (distortion reference, a bad filteroffset, or one chip's frames never re-aligned) rather than a whole-frame misalignment. fix_alignment SKIPS a frame that already has a RAOFFSET header, so correcting the table alone leaves stale frames in place; regenerate the working copy from _cal instead of re-applying on top.

  Misaligned by detector

  nrcblong1/6

  nrcalong0/6

  Diagnostic writeupD2 — internal astrometric repeatability · full diagnostic writeup (PDF)

  visit	detector	dRA (mas)	dDec (mas)	off (mas)	contrast	pairs	window ("")
  1	nrcblong	-1.8	-0.8	2.0	2,538.0	163793	3.0

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F115W visit 1: reference tie recorded as could-not-verify (apply_ok=false)

  A correction is applied only when the tie is coherent AND the gross cross-check (100 mas) passes AND the per-tile map is clean. This one was not, so no correction was made — the frame keeps whatever it had.

  checkpoint_m2_F115W_latest.json

  what is affected, and why

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F162M visit 1: reference tie recorded as could-not-verify (apply_ok=false)

  A correction is applied only when the tie is coherent AND the gross cross-check (100 mas) passes AND the per-tile map is clean. This one was not, so no correction was made — the frame keeps whatever it had.

  checkpoint_m2_F162M_latest.json

  what is affected, and why

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F182M visit 1: reference tie recorded as could-not-verify (apply_ok=false)

  A correction is applied only when the tie is coherent AND the gross cross-check (100 mas) passes AND the per-tile map is clean. This one was not, so no correction was made — the frame keeps whatever it had.

  checkpoint_m2_F182M_latest.json

  what is affected, and why

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F182M_o012 visit 1: reference tie recorded as could-not-verify (apply_ok=false)

  A correction is applied only when the tie is coherent AND the gross cross-check (100 mas) passes AND the per-tile map is clean. This one was not, so no correction was made — the frame keeps whatever it had.

  checkpoint_m2_F182M_o012_latest.json

  what is affected, and why

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F405N visit 1: reference tie recorded as could-not-verify (apply_ok=false)

  A correction is applied only when the tie is coherent AND the gross cross-check (100 mas) passes AND the per-tile map is clean. This one was not, so no correction was made — the frame keeps whatever it had.

  checkpoint_m2_F405N_latest.json

  what is affected, and why

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F470N visit 1: reference tie recorded as could-not-verify (apply_ok=false)

  A correction is applied only when the tie is coherent AND the gross cross-check (100 mas) passes AND the per-tile map is clean. This one was not, so no correction was made — the frame keeps whatever it had.

  checkpoint_m2_F470N_latest.json

  what is affected, and why

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F115W visit 1: worst tile 55692.4 mas at cell (5,2) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F115W_latest.json

  what is affected, and why

  1/36 cells exceed 15 mas, and every one of them is on the mosaic EDGE — that is usually thin coverage (few exposures, so few pairs per cell) rather than a distortion error. Check the pair counts in the table before treating it as a defect. The bulk tie for this visit is 1.34 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (5,2)	55,692.4	-51,835.5	20,364.8	6.0	52394

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F115W_o012 visit 1: worst tile 55693.9 mas at cell (5,2) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F115W_o012_latest.json

  what is affected, and why

  3/36 cells exceed 15 mas, and every one of them is on the mosaic EDGE — that is usually thin coverage (few exposures, so few pairs per cell) rather than a distortion error. Check the pair counts in the table before treating it as a defect. The bulk tie for this visit is 0.50 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (5,2)	55,693.9	-51,837.3	20,364.3	6.0	50485
  (4,0)	15.5	-14.9	4.3	38.0	14459
  (3,0)	15.5	-14.7	4.9	28.0	10137

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F162M visit 1: worst tile 28.9 mas at cell (0,5) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F162M_latest.json

  what is affected, and why

  29/36 cells exceed 15 mas. They are not confined to the edge, so this is an interior residual: a distortion or per-detector alignment problem, not a coverage artefact. The bulk tie for this visit is 5.70 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (0,5)	28.9	-26.7	10.9	21.0	62940
  (5,5)	28.7	-5.4	-28.2	9.0	22806
  (1,0)	26.3	-12.9	-22.9	45.0	67202
  (1,3)	25.5	-20.4	-15.2	29.5	151220
  (0,0)	24.0	-16.9	-17.0	38.0	59202
  (1,5)	23.2	-17.7	15.0	20.3	202819
  (0,3)	23.0	-17.6	-14.8	19.5	131528
  (0,1)	21.4	-17.7	12.0	33.0	163723
  (5,3)	21.2	-15.1	-14.8	18.0	128513
  (3,3)	21.1	-12.6	-16.9	13.8	271012
  (1,1)	21.1	-16.8	-12.7	30.0	102603
  (2,3)	21.0	-11.6	-17.6	19.0	211811
  (3,1)	21.0	15.7	-13.9	21.0	103882
  (3,4)	20.9	16.1	-13.4	15.0	323939
  (4,3)	20.1	-15.3	-13.1	16.5	280537
  (5,1)	19.5	-11.0	-16.1	28.0	53528
  (2,0)	19.4	6.8	-18.2	15.0	114834
  (4,4)	19.4	7.9	-17.7	9.2	395145
  (3,0)	19.2	-12.0	-15.0	16.7	180280
  (1,2)	18.4	-11.2	14.6	15.7	219249
  (2,5)	17.7	-14.8	9.8	20.7	252587
  (3,2)	17.6	9.6	-14.7	19.5	152848
  (0,2)	17.5	8.4	15.4	18.5	160202
  (4,5)	17.4	7.4	15.8	15.0	29291
  (1,4)	17.2	-15.9	6.5	20.3	194201
  (4,1)	17.0	8.0	-15.0	24.5	109592
  (2,1)	16.6	-15.5	-5.7	30.5	125421
  (3,5)	15.9	-8.5	-13.4	10.0	23795
  (5,4)	15.0	-8.0	12.8	14.0	33845

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F162M_o012 visit 1: worst tile 38.3 mas at cell (4,4) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F162M_o012_latest.json

  what is affected, and why

  23/36 cells exceed 15 mas. They are not confined to the edge, so this is an interior residual: a distortion or per-detector alignment problem, not a coverage artefact. The bulk tie for this visit is 0.74 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (4,4)	38.3	24.7	-29.3	12.0	35781
  (3,4)	27.8	25.1	-12.0	17.5	321747
  (5,5)	26.2	8.0	-24.9	9.0	22505
  (1,0)	25.5	-14.6	-20.9	43.0	64964
  (5,0)	25.3	-24.0	8.0	12.0	107074
  (0,1)	23.7	-19.9	12.8	34.5	162786
  (0,0)	23.0	-18.8	-13.3	37.0	58007
  (1,1)	21.2	-17.6	-11.8	28.0	101592
  (3,3)	20.3	14.2	-14.5	15.0	269893
  (1,2)	20.0	-12.3	15.7	16.3	216795
  (3,1)	19.1	15.8	-10.7	18.5	102000
  (5,1)	18.6	-11.6	-14.5	25.0	52621
  (1,3)	17.6	-10.5	-14.2	22.5	148700
  (2,3)	17.6	13.7	-11.0	16.3	210609
  (3,2)	17.5	10.4	-14.1	19.5	149935
  (0,2)	16.9	-7.3	15.2	19.5	158758
  (2,1)	16.3	-15.9	-3.9	28.0	123288
  (0,4)	16.2	12.6	10.2	22.0	104207
  (3,5)	16.2	-10.3	-12.5	13.0	23699
  (4,3)	16.2	12.1	-10.8	13.8	279718
  (0,5)	16.0	-11.2	11.4	20.0	62922
  (2,2)	15.9	-14.8	5.6	15.7	224791
  (1,5)	15.6	7.7	13.5	19.0	18111

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F182M visit 1: worst tile 5733.5 mas at cell (5,0) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F182M_latest.json

  what is affected, and why

  28/36 cells exceed 15 mas. They are not confined to the edge, so this is an interior residual: a distortion or per-detector alignment problem, not a coverage artefact. The bulk tie for this visit is 11.68 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (5,0)	5,733.5	-2,162.6	-5,310.0	6.0	160788
  (4,3)	2,933.5	-2,926.1	207.8	5.0	35499
  (5,4)	2,757.6	1,589.4	2,253.4	7.0	42799
  (3,2)	2,569.1	1,005.0	2,364.3	6.0	21189
  (4,5)	2,548.5	-2,274.4	-1,149.6	6.0	36460
  (2,2)	2,108.2	-2,106.8	-77.2	5.0	28721
  (5,5)	2,082.3	1,788.4	-1,066.5	7.0	28515
  (1,2)	1,755.3	1,646.9	-607.3	6.0	29297
  (2,3)	1,554.8	-1,292.3	-864.6	5.0	27712
  (3,3)	1,465.3	-1,351.9	-565.3	7.0	34539
  (4,4)	1,453.2	-1,191.5	831.9	6.0	44010
  (3,5)	635.3	-611.2	-173.2	5.0	31334
  (3,4)	34.0	33.7	-4.2	7.0	39390
  (5,1)	25.9	-20.4	-15.9	10.0	106042
  (5,3)	24.7	-20.6	13.6	5.0	17403
  (1,4)	23.7	21.8	9.2	9.2	325084
  (4,1)	22.5	-12.3	-18.8	9.0	201763
  (2,0)	22.1	14.2	16.9	7.3	204739
  (0,0)	21.3	-13.1	-16.8	12.5	117893
  (0,3)	20.6	14.3	14.8	7.3	201078
  (1,3)	20.4	-13.3	-15.5	8.3	241869
  (0,4)	19.6	14.2	13.5	9.0	184371
  (0,2)	19.5	12.8	14.6	6.7	243406
  (0,5)	19.4	11.8	15.4	6.5	108616
  (1,0)	18.4	-15.3	-10.3	12.5	126895
  (1,1)	17.3	-12.8	-11.6	11.0	208083
  (5,2)	17.3	-13.6	10.6	14.0	40762
  (2,4)	15.8	15.0	4.7	7.0	35227

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F182M_o012 visit 1: worst tile 5736.7 mas at cell (5,0) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F182M_o012_latest.json

  what is affected, and why

  28/36 cells exceed 15 mas. They are not confined to the edge, so this is an interior residual: a distortion or per-detector alignment problem, not a coverage artefact. The bulk tie for this visit is 1.15 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (5,0)	5,736.7	-2,167.9	-5,311.3	6.0	159656
  (3,4)	2,977.0	-2,514.8	1,593.2	6.0	39283
  (2,5)	2,791.4	-2,309.9	1,567.2	6.0	31241
  (4,4)	2,667.0	-884.3	-2,516.1	6.0	44150
  (3,3)	2,584.7	-1,865.5	1,788.9	6.0	34387
  (4,5)	2,550.4	-2,273.7	-1,155.2	6.0	36403
  (2,2)	1,560.1	-165.7	1,551.3	5.0	28617
  (2,3)	1,556.7	-1,293.5	-866.2	5.0	27550
  (5,5)	1,416.9	-1,016.4	987.3	7.0	28305
  (3,5)	1,116.8	-153.5	1,106.2	5.0	31067
  (5,4)	1,112.3	-1,111.4	-43.4	6.0	42812
  (1,4)	30.3	29.7	-6.1	9.0	29499
  (1,2)	29.8	-24.3	-17.2	5.2	321872
  (5,1)	26.4	-20.9	-16.1	10.0	105644
  (5,3)	24.2	-21.6	10.8	6.0	17282
  (0,0)	23.5	-14.4	-18.5	13.5	115233
  (2,4)	21.8	20.5	-7.2	7.0	35236
  (1,3)	21.6	-13.7	-16.8	7.7	240640
  (4,1)	21.1	-12.8	-16.7	9.0	199713
  (0,4)	19.9	16.4	11.3	8.7	182010
  (2,0)	19.9	13.5	14.6	6.7	202839
  (0,2)	19.6	11.0	16.3	6.7	240437
  (3,2)	19.5	-14.5	-13.0	5.7	231773
  (1,0)	19.2	-15.4	-11.5	13.5	125042
  (5,2)	18.6	-15.1	10.9	14.0	39912
  (0,3)	18.4	13.3	12.7	7.0	199584
  (1,1)	17.9	-14.0	-11.1	11.7	205979
  (2,1)	15.8	-12.4	-9.7	10.0	233475

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F212N visit 1: worst tile 22.2 mas at cell (5,1) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F212N_latest.json

  what is affected, and why

  11/36 cells exceed 15 mas. They are not confined to the edge, so this is an interior residual: a distortion or per-detector alignment problem, not a coverage artefact. The bulk tie for this visit is 7.63 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (5,1)	22.2	-18.1	-12.8	37.0	52769
  (2,3)	19.9	19.4	-4.3	20.0	189087
  (2,4)	18.3	18.0	3.6	28.0	233395
  (1,3)	18.2	16.1	8.4	26.5	143668
  (3,0)	17.5	-16.8	5.0	33.0	157334
  (5,0)	16.6	-14.4	8.3	39.0	87139
  (4,1)	16.5	-12.3	-10.9	29.5	100500
  (5,2)	16.5	-15.7	-4.9	29.0	21245
  (0,0)	16.2	-14.8	-6.5	43.0	64683
  (0,1)	15.8	-11.8	10.5	39.5	154906
  (4,0)	15.4	-13.5	7.3	25.7	183994

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F212N_o012 visit 1: worst tile 21.1 mas at cell (5,1) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F212N_o012_latest.json

  what is affected, and why

  10/36 cells exceed 15 mas. They are not confined to the edge, so this is an interior residual: a distortion or per-detector alignment problem, not a coverage artefact. The bulk tie for this visit is 0.26 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (5,1)	21.1	-17.9	-11.2	37.0	52985
  (2,3)	20.2	19.8	-3.9	20.0	190830
  (2,4)	18.6	18.1	4.0	29.7	234314
  (1,3)	17.8	15.2	9.2	28.0	144056
  (0,1)	16.8	-12.2	11.6	39.5	155440
  (5,0)	16.8	-14.4	8.5	39.0	87511
  (3,0)	16.8	-15.8	5.5	31.0	157384
  (5,2)	16.3	-15.3	-5.7	27.0	21520
  (0,0)	15.9	-14.9	-5.3	41.0	65029
  (4,1)	15.2	-11.6	-9.8	27.5	101365

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F360M visit 1: worst tile 28003.1 mas at cell (0,5) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F360M_latest.json

  what is affected, and why

  33/36 cells exceed 15 mas. They are not confined to the edge, so this is an interior residual: a distortion or per-detector alignment problem, not a coverage artefact. The bulk tie for this visit is 8.01 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (0,5)	28,003.1	-11,089.7	25,713.6	8.0	84780
  (5,3)	25,127.4	-13,058.3	21,467.8	5.3	241441
  (5,5)	17,518.4	-17,506.7	-640.7	5.5	177765
  (3,3)	9,774.9	-5,753.5	7,902.2	9.0	73241
  (4,5)	9,727.4	-7,729.7	-5,905.3	5.0	34171
  (1,5)	8,993.1	-7,715.8	4,619.9	5.0	31551
  (0,3)	8,489.8	-8,488.1	-169.0	5.0	27098
  (1,3)	5,321.3	-5,312.2	311.6	6.0	54631
  (2,3)	4,764.3	-4,497.6	-1,571.9	7.0	66991
  (4,3)	3,739.9	1,443.9	-3,450.0	7.0	66635
  (3,4)	41.6	40.4	-10.1	20.0	58447
  (0,4)	40.5	18.3	36.1	7.0	29062
  (3,1)	33.7	22.7	-24.9	14.0	40143
  (2,5)	32.3	15.3	-28.5	10.0	33146
  (2,0)	31.0	-9.9	-29.4	11.0	32903
  (0,1)	28.3	-8.6	27.0	11.0	33236
  (4,1)	28.2	10.7	-26.1	16.0	39107
  (2,2)	26.5	18.2	-19.3	13.0	54643
  (4,4)	25.9	16.9	-19.6	12.0	64387
  (1,4)	25.8	21.3	-14.5	8.0	50956
  (5,4)	25.3	11.8	22.4	13.0	49036
  (5,0)	24.1	-6.0	-23.3	6.0	18399
  (2,1)	23.4	-22.1	-7.5	17.0	39514
  (1,1)	22.1	-15.1	-16.1	12.0	36387
  (5,1)	22.1	-20.1	-9.0	9.0	20072
  (1,2)	21.9	18.5	11.8	9.0	49332
  (4,0)	21.9	-13.1	-17.6	14.0	40705
  (1,0)	20.9	-12.9	-16.4	11.0	25614
  (0,0)	20.5	-12.2	-16.5	8.0	19793
  (3,5)	19.9	10.4	-16.9	10.0	32315
  (3,2)	18.8	10.1	-15.8	12.0	53089
  (2,4)	18.1	17.4	5.0	11.0	62165
  (3,0)	18.0	-13.7	-11.6	14.0	42096

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F360M_o012 visit 1: worst tile 26576.0 mas at cell (0,5) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F360M_o012_latest.json

  what is affected, and why

  33/36 cells exceed 15 mas. They are not confined to the edge, so this is an interior residual: a distortion or per-detector alignment problem, not a coverage artefact. The bulk tie for this visit is 1.83 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (0,5)	26,576.0	25,484.1	-7,539.7	5.5	138875
  (1,3)	9,612.8	-1,757.9	9,450.7	6.0	31157
  (5,3)	8,947.9	-8,618.2	2,406.4	5.0	27736
  (0,3)	8,185.5	6,695.4	4,708.9	5.0	21262
  (5,0)	8,111.2	3,277.1	7,419.7	6.0	16844
  (1,5)	6,857.5	6,412.6	-2,429.6	7.0	34856
  (2,3)	5,376.9	-3,916.3	3,684.2	6.0	39964
  (4,5)	3,915.4	-3,296.0	-2,113.5	6.0	39489
  (4,3)	1,963.4	-1,961.4	87.5	7.0	45181
  (1,2)	38.3	-25.2	28.8	11.0	31147
  (3,4)	33.5	33.1	-4.7	17.0	50195
  (2,0)	33.1	-11.8	-30.9	12.0	33466
  (2,1)	31.6	-27.8	-15.0	17.0	39650
  (5,1)	29.9	-28.0	10.5	9.0	18661
  (3,1)	29.2	20.0	-21.3	13.0	40656
  (5,4)	29.2	-5.8	28.6	12.0	51214
  (0,1)	29.0	-8.7	27.7	15.0	32247
  (0,4)	25.1	19.8	15.4	6.0	24690
  (4,0)	23.8	-14.2	-19.1	16.0	40700
  (0,0)	23.0	-17.9	-14.5	9.0	19703
  (3,5)	22.5	18.1	-13.3	11.0	35425
  (3,3)	21.5	18.0	11.9	15.0	47044
  (2,5)	20.5	1.0	-20.4	8.0	38034
  (2,2)	20.1	11.9	-16.2	12.0	36501
  (1,1)	19.7	-17.6	-8.9	13.0	36553
  (0,2)	18.9	-13.8	13.0	13.0	24180
  (1,0)	17.8	-17.8	0.5	8.0	26226
  (4,4)	17.7	14.9	-9.7	11.0	50799
  (1,4)	17.6	16.1	-7.2	7.0	43067
  (3,0)	16.2	-16.1	2.1	12.0	42370
  (4,1)	15.7	-12.0	-10.1	13.0	39441
  (2,4)	15.6	14.1	6.5	11.0	50484
  (5,5)	15.1	11.5	9.8	8.0	30751

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F405N visit 1: worst tile 27.9 mas at cell (5,1) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F405N_latest.json

  what is affected, and why

  17/36 cells exceed 15 mas. They are not confined to the edge, so this is an interior residual: a distortion or per-detector alignment problem, not a coverage artefact. The bulk tie for this visit is 8.53 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (5,1)	27.9	-21.8	-17.5	30.0	17863
  (3,4)	22.3	19.8	-10.4	73.0	73314
  (0,2)	22.2	7.8	20.7	30.0	45476
  (2,3)	22.1	21.4	-5.2	51.0	72871
  (3,0)	20.3	-19.4	-5.8	59.0	46445
  (0,1)	19.9	-13.2	14.9	67.0	48805
  (2,4)	19.0	18.5	4.2	67.0	68595
  (1,3)	18.6	16.3	9.0	42.0	60552
  (0,0)	18.4	-17.7	5.1	26.0	16813
  (0,4)	17.6	14.0	10.6	42.0	37120
  (2,0)	17.2	-15.9	-6.6	45.0	36803
  (0,3)	17.0	16.7	3.1	33.0	46273
  (1,4)	16.6	15.2	6.7	55.0	64630
  (4,1)	16.5	-9.3	-13.6	52.0	45777
  (1,2)	16.3	-8.5	13.8	49.0	71980
  (5,4)	15.8	-9.0	13.0	67.0	76872
  (1,5)	15.3	13.8	6.7	52.0	48403

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F405N_o012 visit 1: worst tile 27.9 mas at cell (5,1) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F405N_o012_latest.json

  what is affected, and why

  18/36 cells exceed 15 mas. They are not confined to the edge, so this is an interior residual: a distortion or per-detector alignment problem, not a coverage artefact. The bulk tie for this visit is 1.29 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (5,1)	27.9	-21.6	-17.5	28.0	20952
  (0,0)	24.3	-23.4	-6.6	32.0	23352
  (3,4)	22.3	19.7	-10.3	73.0	72449
  (3,0)	20.8	-20.0	-5.7	62.0	50754
  (2,3)	20.5	19.4	-6.5	51.0	53731
  (0,1)	18.5	-12.1	14.0	75.0	33138
  (2,4)	18.1	17.8	3.1	65.0	68865
  (4,1)	17.3	-11.4	-13.0	58.0	36010
  (0,4)	17.2	13.9	10.2	41.0	35691
  (1,3)	17.0	14.0	9.7	37.0	41406
  (5,0)	16.8	-16.4	-3.9	34.0	25366
  (1,0)	16.1	-15.3	-4.9	32.0	27423
  (4,0)	15.9	-15.1	-5.0	66.0	55264
  (0,2)	15.6	8.0	13.4	38.0	28455
  (2,0)	15.5	-14.0	-6.6	45.0	38800
  (5,4)	15.3	-9.3	12.2	66.0	76543
  (1,2)	15.3	-8.5	12.7	53.0	39111
  (1,5)	15.2	13.8	6.4	52.0	48511

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F470N visit 1: worst tile 25.9 mas at cell (5,1) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F470N_latest.json

  what is affected, and why

  13/36 cells exceed 15 mas. They are not confined to the edge, so this is an interior residual: a distortion or per-detector alignment problem, not a coverage artefact. The bulk tie for this visit is 7.53 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (5,1)	25.9	-22.0	-13.8	31.0	21481
  (0,0)	21.0	-20.3	5.1	39.0	25932
  (3,4)	20.1	17.5	-9.9	85.0	73321
  (3,0)	19.4	-18.5	5.9	73.0	54336
  (5,0)	19.2	-18.6	4.7	37.0	27804
  (0,1)	18.9	-8.9	16.7	69.0	47912
  (4,0)	18.3	-17.0	6.8	84.0	57657
  (2,3)	17.8	16.5	-6.6	65.0	59713
  (2,4)	16.2	16.1	1.9	78.0	73146
  (5,2)	15.9	-13.3	-8.7	22.0	11191
  (3,3)	15.9	12.2	-10.1	67.0	68817
  (0,2)	15.8	10.4	11.9	48.0	39959
  (2,0)	15.7	-12.9	9.0	59.0	42625

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F470N_o012 visit 1: worst tile 26.5 mas at cell (5,1) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F470N_o012_latest.json

  what is affected, and why

  14/36 cells exceed 15 mas. They are not confined to the edge, so this is an interior residual: a distortion or per-detector alignment problem, not a coverage artefact. The bulk tie for this visit is 0.72 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (5,1)	26.5	-22.7	-13.7	31.0	21634
  (0,0)	22.3	-21.8	4.8	39.0	25996
  (3,0)	20.7	-19.8	6.2	76.0	53971
  (5,0)	20.2	-19.8	4.1	36.0	27726
  (0,1)	19.8	-10.5	16.8	75.0	47158
  (4,0)	18.6	-17.5	6.4	81.0	57648
  (3,4)	18.5	16.0	-9.4	82.0	72774
  (2,3)	18.2	17.3	-5.8	66.0	59560
  (2,0)	16.7	-14.1	8.8	58.0	42364
  (5,2)	16.5	-13.9	-8.9	22.0	11050
  (4,1)	15.9	-12.0	-10.5	61.0	38002
  (1,0)	15.9	-14.6	6.2	39.0	30983
  (5,5)	15.5	9.1	-12.5	60.0	54054
  (2,4)	15.3	14.0	-6.1	75.0	73168

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F480M visit 1: worst tile 28530.1 mas at cell (0,5) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F480M_latest.json

  what is affected, and why

  24/36 cells exceed 15 mas. They are not confined to the edge, so this is an interior residual: a distortion or per-detector alignment problem, not a coverage artefact. The bulk tie for this visit is 9.80 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (0,5)	28,530.1	27,907.4	5,928.2	5.0	102545
  (4,5)	3,911.9	-3,294.9	-2,108.7	8.0	48881
  (5,3)	74.5	70.1	25.3	8.0	44132
  (0,3)	45.8	42.4	-17.3	9.0	31199
  (0,4)	35.5	30.1	18.8	14.0	35163
  (3,4)	34.9	33.8	-8.9	28.0	75612
  (1,3)	34.5	30.1	-16.9	14.0	68165
  (3,5)	30.7	26.4	-15.8	15.0	41317
  (4,4)	28.2	24.4	-14.1	18.0	85694
  (0,0)	26.8	-16.8	-20.9	18.0	28983
  (1,4)	25.2	24.9	-4.2	17.0	73955
  (2,3)	24.1	23.4	6.0	8.0	90356
  (3,1)	22.7	11.8	-19.4	35.0	50747
  (3,2)	22.4	12.4	-18.6	38.0	73553
  (2,5)	22.3	20.6	8.7	16.0	45356
  (5,1)	21.5	-17.5	-12.5	16.0	26629
  (2,2)	20.7	15.5	-13.7	40.0	82517
  (0,2)	20.6	16.9	11.7	28.0	51081
  (4,1)	19.7	9.7	-17.2	29.0	47149
  (3,3)	19.2	16.4	10.0	10.0	98986
  (3,0)	18.3	-11.7	-14.1	31.0	59165
  (1,1)	18.0	-8.1	-16.1	24.0	47598
  (5,4)	16.6	13.8	9.3	19.0	75683
  (5,2)	16.2	0.4	-16.2	12.0	22602

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• warn
  F480M_o012 visit 1: worst tile 31.0 mas at cell (3,5) (> 15 mas)

  36/36 tiles are reported "ok", but that counts tiles whose offset histogram had a coherent PEAK — not tiles within tolerance. The m7 cross-band gate is no significant 2" cell above 15 mas, and this cell exceeds it, so the bulk tie being ~0 does not mean the field is flat.

  checkpoint_m2_F480M_o012_latest.json

  what is affected, and why

  25/36 cells exceed 15 mas. They are not confined to the edge, so this is an interior residual: a distortion or per-detector alignment problem, not a coverage artefact. The bulk tie for this visit is 0.16 mas — a small bulk value cannot cancel a local one.

  Diagnostic writeupD3 — absolute tie to the reference frame · full diagnostic writeup (PDF)

  cell	off (mas)	dRA	dDec	contrast	pairs
  (3,5)	31.0	27.2	-14.8	15.0	46606
  (3,4)	30.9	30.6	4.3	23.0	67000
  (1,3)	30.6	22.5	20.8	14.0	41696
  (0,3)	30.3	30.2	3.0	14.0	28116
  (4,5)	26.9	26.7	3.4	12.0	54071
  (5,1)	26.5	-23.9	-11.6	16.0	25102
  (0,0)	23.8	-16.4	-17.2	21.0	29116
  (0,4)	23.6	20.3	12.2	13.0	32342
  (3,3)	22.6	17.4	14.5	24.0	60991
  (0,1)	20.5	-12.1	16.6	37.0	46311
  (3,0)	20.2	-18.1	-8.9	33.0	59430
  (1,2)	19.7	-10.0	17.0	29.0	49151
  (0,5)	19.7	-13.4	14.4	6.0	18560
  (1,0)	19.5	-19.4	-2.6	20.0	33563
  (0,2)	19.0	10.6	15.8	28.0	39417
  (1,1)	18.6	-14.6	-11.5	24.0	47889
  (2,5)	18.3	14.0	11.8	13.0	52381
  (1,4)	18.3	18.2	-2.2	14.0	57658
  (4,4)	18.0	14.6	10.4	19.0	68954
  (3,2)	17.0	6.8	-15.6	30.0	47753
  (4,0)	16.2	-11.9	11.0	34.0	61308
  (2,4)	16.0	15.5	4.0	21.0	66766
  (2,3)	15.7	14.5	5.9	15.0	54341
  (5,4)	15.7	8.1	13.4	23.0	70400
  (4,1)	15.4	-7.2	-13.6	31.0	48097

  Existing figures

  • astrometry_offsets_2026-07-10.png audit_plots
• ok
  F360M_o012: all 12 exposures within 2.0 mas of consensus

  checkpoint_m2_F360M_o012_latest.json
• ok
  F405N: all 12 exposures within 2.0 mas of consensus

  checkpoint_m2_F405N_latest.json
• ok
  F480M_o012: all 12 exposures within 2.0 mas of consensus

  checkpoint_m2_F480M_o012_latest.json
• ok
  m7 present (3 file(s))

  catalogs/
• ok
  m8 present (3 file(s))

  catalogs/
• ok
  8 recent log(s), no error signatures

  logs/
• ok
  m7: single tag 2026-08-30_PR579_7a591cb-dirty (27 products)

  catalogs/*m7*.prov.json
• ok
  m8: single tag 2026-08-30_PR579_7a591cb-dirty (3 products)

  catalogs/*m8*.prov.json