I think we've got a fault!

[#1]

Chris,
From experience, a line test (or any other investigation)  at the "wrong time" might not point to any problem.  The report here is that the SNRM step change happens circa 18:00 - BTOR don't do anything after 18:00.  Looking before then is not going to find the source of this apparent REIN.
This is a pronounced injection of noise, apparently at near enough the same time each day.  It would be useful to look in the area to determine where the D-Side runs to the green cabinet.  We've seen reports on here of take-a-way extractor fans causing local interference.  I have personally seen big SNRM shifts from CCTV systems as dusk falls, but not step transients like this.
There have been reports on here of neighbour's plasma TV systems causing issues.  You might want to try wandering around the neighbourhood with your radio and see if it detects anything.
Good luck with your hunting!

[#2]

That's good advice, townman (I was going to say that, but I forgot - honest!).
I've been recording the stats via RouterStats for a full day now and I observed the opposite effect (i.e. the SNRM going up in the space of 30 seconds or so by 2 or more dB). I have attached a photo of this.
Please take notice of the marked increase in SNRM by about 2dB.

[#3]

I suspect that this interference is causing the SNRM to drop more than it should (disregarding the sudden drop). The SNRM always used to drop by maybe 1-3dB. Now it drops by 5 sometimes. In addition, the errors are much higher than they used to be.
I really hope this gets sorted soon.

[#4]

Happened again tonight, at almost exactly the same time as the previous nights.

[#5]

I should mention that there has been no response to the fault ticket and it's now been over 24 hours.

[#6]

You might need to do a few BT speedtests when the problem is present and reducing the speed significantly, to provide "evidence" of the problem that will be visible on the backend system. Make sure you click for further diagnostics and enter your telephone number after the initial test finishes.

[#7]

That's good advice, but right now we won't see any speed difference as we've been told to leave the connection at it's slower sync speed. Then, we can still see the effects of the interference but the connection remains stable too.

[#8]

Still no update on the fault. What's going on, Plusnet?!?!!?

[#9]

Sorry for the wait, we'll get this picked up today.

[#10]

I seem to have closed the closed the ticket by mistake. Could it please be reopened?

[#11]

We can't reopen it, could you open a new one and I'll get it picked up?

[#12]

Done.

[#13]

Thanks, someone's getting onto that.

[#14]

Got an update. You say the speed is within normal limits. That's irrelevant. There is clear interference which needs to be sorted!
However, you do seem to have found a fault:
Results:
Problem Explanation: [CU21] The Copper Line test return OR error code 2013.
Resolution/Recommendation: Pass to BTW to investigate the error code.
BRAS: 6376
Profile Rate: 7430 832
Test File Size: 8774168 395950
File Transfer Average Throughput/Session: 4490 610

[#15]

There's still nothing happening with this. We need to get an engineer out to fix this. It's obvious there's a fault - YOU'VE FOUND IT!
I will paste in exactly what I just added to the ticket. Can somebody from Plusnet please make sure this gets fixed!
The SNRM drops by up to 3dB in the space of a minute or so between 5-6PM every night. This is NOT due to it getting dark, this is due to interference, which needs to be sorted (appears to be REIN) - that is why I have opened this fault.
When this happens, the speed becomes so slow that webpages cease to load. Since you restarted the connection this afternoon, it connected at a good speed of 8Mb, but the SNRM dropped at approx 6 and the errors created were ridiculous. CRCs were in the thousands, increasing by hundreds every minute. We had to restart the router to get the connection working again.
The DROPS are NOT the FAULT. That is ME, fixing the speed problem. YOU (Plusnet) need to FIX THE INTERFERENCE by getting an engineer out, etc.
Please look at the images attached above to see the problem.