GPS glitching

There's a bug in the GPS system.. April 6 is the change (a week-count rollover bug), and depending on the local receiver, the time may roll back right away, or maybe in a few weeks--it depends on the firmware. AFAIK it won't matter for consumer devices; they don't really care what time the GPS thinks it is, just where the cell phone is.(*) But that time is important for a lot of research systems.

Including IceCube. I heard a report on this last week, at which it was decided to switch from the old (the company doesn't exist anymore) receivers to the new pair of receivers. The new ones gave slightly better time resolution too, which is a plus.

The timestamp matters a great deal to IceCube, since we need to know the timing of signal pulses to nanoseconds--it matters. The GPS feeds the clock, which fans out to our detectors. To reconstruct a track using the signals from the phototubes we need to know their relative timing accurately. Most Multi-Messenger Astronomy doesn't require nanosecond timing, luckily. (In fact, different particle signals could be delayed by days wrt each other. Protons won't arrive at quite the same time as photons, and different production processes may happen at slightly different times.)

---

(*) The old Honda's instrument cluster is completely out of stock, and junk yards don't have them anymore either. I could either have yanked it and tried to find and fix the intermittently broken wire, or used a cheap GPS speedometer. I went with cheap. I don't think it's going to go nuts on Saturday.