This one is a bit of a shorter update, because there’s not a lot new to report. Outages remain frequent, and longer, like they did in March. Starlink announced that they would be rolling out an update that would allow the dish to change which satellite it’s connected to when it sees an obstruction. If that has made it to my dish, I haven’t noticed its effects.
And yet, even with my obstruction problems, Starlink is still nicer for most of my internet usage than my fixed wireless connection. I reconfirmed that in the middle of the month, when I felt like my Starlink connection was especially unstable. After two days of much slower internet (the large grey bands in the timeseries plot below), I returned to Starlink and have waited out the downtimes since. (Except for video calling, when the slow-but-connected fixed wireless still wins.)
I’m officially looking forward to writing my May update. Parts for a tower to raise Dishy out of tree obstruction territory should arrive on Monday. Fingers crossed that about a week from now, the shape of my outage graphs will change dramatically.
This month’s graphs are below. They cover noon on March 31 through just before midnight on April 30. A refresher on what they mean:
- Time series plot:
- Each square represents one second. There are 1200 seconds (20 minutes) in each line, so a day is 72 lines tall.
- White: connected. Red: Obstruction. Blue: Beta downtime (recently renamed “other” in the mobile app). Grey: no data (the dish rebooted, or my laptop was connected to the other network). Yellow: a special case of White, where the connection lasted for at least 30 minutes, without an outage lasting longer than 2 seconds.
- The height of each bar indicates the number of times an outage (or connection) was observed lasting the length of time indicated on the horizontal axis.
- Colors are the same as the time series plot, though now all connected durations are yellow (no more white), and all durations are plotted (not just greater than 30 minutes). Grey (no data) is also left out of this chart, mostly because the counts are so small compared to other fields, that they wouldn’t be visible.