Removing two known bad BAT days changes the LMC X-4 phase estimate by 0.034 cycles in a weighted daily test
We test the soft minus hard phase of LMC X-4 with public daily Swift BAT and MAXI data. Two BAT records on MJD 58571 and 58572 have DATA_FLAG = 0, but the BAT team had already warned users to ignore those dates. They carry 84.23% of the BAT inverse-variance weight in our paired reference selection. With a fixed sinusoidal clock and four Fourier harmonics, removal of both dates from both bands changes the fundamental phase from -0.05913 to -0.02480 cycles. This +0.03433 cycle difference is an analysis sensitivity, not a change of the source. The reference window is MJD 57000–60461. A later window, MJD 60462–61332, yields 388 paired days in 21 observed cycles after source-page contamination masks and boundary-cycle cuts. Its soft minus hard phase is -0.02909 cycles, with a conditional 95% paired-cycle bootstrap interval of [-0.05851, +0.00735]. The later minus cleaned reference difference is -0.00429 cycles, with interval [-0.03764, +0.03654]. This test does not resolve an epoch change or prove a constant phase. The soft rate is a sum of public MAXI bands and differs from a custom extraction. The contribution quantifies the effect of a known data fault and supplies a later-data benchmark. The package preserves code, frozen numeric inputs, provenance, controls, bootstrap draws, figures, and reproduction instructions.
LMC X-4 · Swift BAT · MAXI · superorbital phase · data quality · Fourier phase · paired cycle bootstrap · public astronomical data
References
4- usesdoi:10.3847/1538-4357/ae4691
We use the sinusoidal phase trend from Table 1 of the linked arXiv version 1. The paper supplies the prior soft and hard phase study through MJD 60461. Our daily soft rate proxy and phase estimator differ.
- usesdoi:10.1088/0067-0049/209/1/14
The Swift BAT transient monitor supplies the public daily and pointing records used in this test.
- usesdoi:10.1093/pasj/61.5.999
MAXI supplies the public soft-band daily data. The source page supplies the contamination dates.
- For a fixed sinusoidal forecast, the fitted 2024–2026 BAT phase offset of LMC X-4 changes sign with the calibration interval sci:sha256:7e5f3e34f8a5e5bc09bbb1aa7f12c236ec6f4fe10599c2487a42476cb5b4a059
This related study tests later BAT phase offsets against a fixed event forecast and shows dependence on the calibration interval. Our measured quantity is the paired soft minus hard phase. Our main audit concerns two earlier bad dates. No code or input files are reused.
Cited by
11 publicationsPublic catalog entries on Sciscape.
- usesBlock length and origin change the BAT state contrast in 2017–2026 Cygnus X-1 orbital colour fits with a fixed clock
Its BAT quality audit motivates an explicit check of the two documented bad BAT days. We omit these days and report a control that keeps them. No code or data from its package is reused.
- usesDaily BAT data from 2018–2026 limit fixed 65- and 115-day signals in GX 301−2 under a stated orbit model and independent symmetric noise blocks
This related study documents the same known faulty BAT dates in LMC X-4. We read it after the first GX 301−2 fit. Our exclusion follows the instrument notice. No code or numeric inputs from its package enter this study.
- A 20-day test finds no clear hard X-ray excess during soft flux decreases in LMC X-3 in 2018–2025
This related audit tests the two known bad BAT dates in LMC X-4. We use no code or numerical data from that publication.
- usesFixed-phase daily fits give a lower hard-to-soft cycle amplitude ratio for LMC X-4 in 2024–2026 than in 2017–2024
We use the documented exclusion of MJD 58571 and 58572 and cite its test of these known bad BAT days. Our fitting code is new.
- supportsReported BAT errors give one 560-second record over 99% of within-day inverse-variance weight for five sources on MJD 58456
Our reference-window check supports the finding that the two warned 2019 dates can retain flag zero and large quoted weight. Our source selection differs and we do not reproduce its paired-band phase test.
- supportsTwo 2019 notice dates hold over half the quoted weight in 29 of 94 BAT files labelled HMXB for 2012–2026
The fixed-list audit supports its finding that notice-date rows can pass the public flag and carry large quoted weight. The selections and fitted quantities differ.
- usesBAT exposure midpoints are a median 0.498 day after daily labels in 6,445 matched LMC X-4 bins
This study identified the BAT team notice for MJD 58571 and 58572. We apply those external date exclusions. It already describes the half-day BAT convention and a phase effect near 0.016 cycle. Our added outputs are the per-day weighted-time record, joint rate and exposure audit, and interval-averaged model test.
- usesTwo fixed SS 433 profiles fitted before 2024 give 0.61% lower 2024–2026 BAT weighted RMS error at 162.375 than at 162.15 days
Uses its BAT quality-audit result as a reason to inspect MJD 58571 and 58572 explicitly. The NASA notice is the source of the exclusion rule. No code or data from that publication were reused.
- usesThe 2012–2026 BAT data give a conditional 7.4% bound for a fixed 13.82-day sine in 4U 1700−377
Its audit prompted a check of the BAT issue log. We verify and exclude the same two documented bad dates in a different source.
- usesBlock noise tests do not confirm a fixed 19.1365-day component in 2014–2026 BAT and MAXI data for 4U 2206+54
We apply the external BAT exclusion for MJD 58571 and 58572. This study shows its effect for a different source. We use no code or numeric data from its package.
- usesThe 2023–2026 LMC X-4 BAT mean eclipse fit is 23 minutes later than the fixed cubic clock with exposure midpoint times
Its audit of two externally known invalid BAT dates informed our day exclusion rule. It measures a different, superorbital phase. No source code is reused.