Bounds of three quoted errors plus half a time sample allowed peak spans below 1 ms with a free DM shift in 98 selected CHIME burst fits

We tested representative component peak coordinates in 98 selected bursts from CHIME/FRB Catalog 2 release 20251231. Each burst had at least three components. We minimized their corrected time range over one cold-plasma dispersion-measure (DM) shift and independent coordinate intervals. We used bounds of three quoted time and spectral parameter errors, conservative outer frequency intervals, and an added half time sample on each side. With any real DM shift, all 98 minimum spans were below 1 ms; the largest was 0.987762 ms. With fixed catalogue coordinates, 85 spans were above 1 ms. In a test added after the primary result, a DM shift limited to three quoted DM errors left 88 spans above 1 ms under the same coordinate bounds. Their median was 12.458712 ms. Spectral bounds caused most of the reduction when the DM shift was free. Linear programming and an independent enumeration method agreed. The package contains the method, code, data extract, provenance, figures, selection table, and coordinate witnesses. These intervals are sensitivity sets, not confidence regions. Alignment in these outer intervals does not establish a valid joint fit or show that dispersion explains the full burst shape.

fast radio bursts · CHIME · dispersion measure · component timing · interval bounds · spectral parameters · catalogue audit · reproducible methods