SNR/Connection speed wobble.

[#1]

Not a massive problem - yet - but after 100+ days at an acceptable level, I've had another DLM intervention which has upped the SNR with the resulting connection speed drop - why?

Can some kind CS person have a look?ResonablePlaying silly so-and-so's

[#2]

Hi @jab1,

I have tested the line and I could not find a fault. I've gone ahead and reset your SNR margins but It could be an intermittent noise issue that needs raising. Have you tried the usual checks from your side?

[#3]

Thanks @Beyhive Didn't think you'd find a fault - they didn't last time, either.

Thanks for the reset, I'm now at 6dB d/s SNR but 12.26Mb/s speed - which I won't, sadly maintain, it's too high for my length of line.

In answer to your question - yes, I always try to solve a problem myself.

TBH - if it is a noise problem, it is not being reflected in my monitoring.

I'l live with it for a while, but just wish the problem would stop popping up.

[#4]

Hi @jab1 

It's strange that its not cropping up on testing.  I'm not all together sure what else we can do at this point. 

I'l live with it for a while, but just wish the problem would stop popping up.

Can you shout us as soon as something happens again so we can try and find an issue asap after it appears at your side? 

Thanks, 

MoR

[#5]

@MasterOfReality If I'm watching the lappy at the time, of course I will, but I'm not always in front of it.

[#6]

It didn't look like anything to do with the DLM to me. I think there must be 2 disconnections on the graph shown, at the times both the tx and rx SNRM were plotted at zero, but these do not correspond to the times of any large changes in SNRM. Unless the DLM change is on a subsequent graph not shown.

[#7]

@ejs Graph #1 was just to illustrate the 'steady' state before the massive change at 1438 on graph #2. The previous drops did not indicate any SNR change, but that one certainly did - unless I'm missing something major?

The other 'drops' were not disconnections, merely an anomaly of RSL which I have come to live with - it occasionally loses connection, sometimes I catch it, do a quick pause and restart of the graph and don't think any more of it, sometimes it happens and it can be a while before I spot it and carry out the same procedure, the graph then recovering to the steady state.

[#8]

If the DLM changes the target SNRM, the connection will re-sync at the new target SNRM. Usually.

OK it's a bit odd that the SNRM drops down to about 2.5 shortly before 14:00, then a few minutes later there's a re-sync and the SNRM is about 3dB, then there's another re-sync and again the downstream SNRM is about 3dB afterwards. Then a few minutes later the downstream SNRM jumps up to 11 and then continues rising gradually for a few minutes until it's over 14. It might have been useful to know the connection speed while the SNRM was around 3dB, but I'm guessing the speed was lower, the target SNRM wasn't 3, but the characteristics of the noise were such that it wasn't detected very well while the modem was establishing the connection. It also doesn't seem likely that the target SNRM could have been 3dB and then have the SNRM increase to over 14dB without a re-sync.

You could have tried rebooting your modem or commanding it to re-sync to restore your usual speed.

[#9]

OK, @ejs Corresponding speed graphs, just for you:

#1#2

[#10]

A question which I need a fairly simple answer to:

As nothing in my home environment has changed (to my knowledge, anyway) why did DLM suddenly decide to push my SNR up, and obviously reduce my connection speed around 1400 on the 25th, then increase  around 1100 on the 27th, only to interrupt again about 5 hours later?

I have tried to force a better ratio/speed this afternoon, but this only lasted around 3 hours, and I don't really want to spend hours attempting to resolve the problem - I currently don't have the patience or inclination.

As stated previously, this strange behaviour was after 100+ days at a reasonable 9992 Kbps connection speed, admittedly with a 12dB margin, but I was perfectly happy with that.

As an aside, why do I reconnect at something around 12250, when I have been told, and quite understand, that the most I can reasonably expect is around the 9900 mark?

The screenshot below illustrates the 'modifications' better than I can.

[#11]

There could be various reasons, but our testing isn't showing a definitive cause.

Looks like the SNR target is still at 6dB but I can definitely see the margin at 15dB. I'd be tempted to arrange another engineer to investigate this further as the SNR shouldn't be jumping around the way it is.

[#12]

@Gandalf Thanks for your rapid reply - the *bleep* thing has just interfered again - now at 6.1dB and 12323 Kbps!!!

[#13]

Yeah that may have been me testing the line. Sorry

[#14]

No problem - but that raises another query for me - why?

[#15]

When we test a line more often than not it'll drop an ADSL connection, for fibre it generally doesn't.

As the connection had dropped your router has basically resynced