Weirdness. I decided I was going to do a 'prove-it' test for AA so that they could book an engineer asap
tomorrow (Tuesday) Such a test would prove that the problem is in OR's network not chez moi.
My wife is going to be out all day on Tuesday, so won't be able to do any modem-swapping 'prove-it' tests. So tonight we swapped modems 2 and 4's connections to the network. Line 4 was in fact down at the time of the test.
We swapped the modems 2 and 4's connections over and this is the report which we got from the Firebrick :
Link status report from Firebrick after 2-4 swapped
===================================================
Gave error messages: modem 2 is connected to line 4; modem 4 is connected to line 2
then:
*** One of the links is DOWN ***
--
Interface (of FB) PPP MTU Uptime Status Phone line circuit ID‡
‒‒‒‒‒‒‒‒‒‒‒‒————— ‒‒‒ ‒‒‒ ‒‒‒‒‒— ‒‒‒‒—— —‒‒‒‒‒‒‒‒‒‒‒‒‒‒‒‒‒‒——
IF-MDM1 PPP1 1500 16d 16:32:01 Active BBEU20700042 - line 1
IF-MDM2 PPP2 1500 - Link down (Trying)
IF-MDM3 PPP3 1500 02:55:11 Active BBEU20700055 - line 3
IF-MDM4 PPP4 1500 00:03:00 Active BBEU29735631 - line 2
(‡ PPPoE connect message)
--
Firebrick named “tg”, model FB2900, queried at address 81.187.---.---
===================================================
Actually this came from a program that I wrote which processes the output of the Firebrick's PPPoE status function so that it is decorated more elaborately and has various possible error messages.
As expected, the bad line 4, which had been down before, moves to i/f 2 as seen by the Firebrick. So this proves two things: that line 4 is (still) bad (when it is connected to modem 2), and proves that modem 4 is good as it works when it has to drive line 2. So all exactly as expected.
So having achieved what we wanted to achieve, we swapped things back. Then things got weird. Firstly line 4 came
back to life, and not only that but at full speed. Inspecting the SNR-vs-tones graph showed that the bad hollow curve shape of the graph as seen in every fault situation was now absent. So somehow, I fixed the problem, just like the OR engineers who have come out have done even though we don't know how.
So at the moment we possibly have no problem, because I have fixed it. There definitely was one today, from the evidence of the above report and from the evidence of the BT copper line test error report.
I am going to update AA as to what’s going on and ask them to do their own additional line tests to see if the line fault is still out there. I can’t see it on a recent second copper line test, so it’s one of those nightmare intermittent things.