Stardate
20021004.1035 (Ship's log): The RR tech came yesterday. Before coming over, he'd spent some time testing things from the other end. At the same time he was pinging my modem and seeing losses, he was pinging another modem on the same circuit which was working perfectly.
As soon as he walked in and saw what I had, he said he knew what was really wrong, and it was something I would never have suspected. Indeed, I was incredulous as soon as he suggested that the modem's power brick was actually at fault.
At the suggestion of several readers, I had downloaded and started running "pingplot" a couple of hours before he showed up, so when he replaced the power brick, pingplot showed that the intermittent failures went away. Then he switched back to the old power brick (same modem, all this time) and the modem couldn't even connect. So new power brick it was.
The only thing I can conceive of is that the old power brick wasn't providing clean power.
Unfortunately, he decided he also wanted to replace the modem, just on general principles. Now performance is a totally digital step function: either it works perfectly or it's totally out for extended periods. It went out around 8:00PDT last night and stayed out for about five minutes, until I got impatient and power-cycled the modem. It went out again at 8:15 but came up on its own after a couple of minutes. It also went out a bit after 10:00, and stayed out until I power cycled the modem.
It also went out just before 5:00 AM PDT this morning and stayed out until I noticed it at 10:15, when I power cycled it. I've put in a call; it's clear that this modem isn't any good and will have to be replaced again. With any luck they'll be able to squeeze me in this afternoon. If not, it won't be until Monday at the earliest, and it's likely that there will be further cases where the link goes down until I notice it.
Please note that you cannot help me notice it with email because my email travels through the same cable modem, so I will not receive your email until after I've noticed it and fixed it.
include
+force_include -force_exclude
|