Target SNR Margin

[#1]

If on restarting my router shows a SN Margin of 9.0dB, does that mean my target margin has been set to 9.0dB?
The margin has remained around this level for the past 4 hours at a below normal sync rate of 3104Kbps.
Edit: The router did resync from 3488 down to 3104 after about 1 hour, with the same SN Margin.

[#2]

Quote from: MrT

If on restarting my router shows a SN Margin of 9.0dB, does that mean my target margin has been set to 9.0dB?

Most likely, yes.

[#3]

Quote from: walker23

Most likely, yes.

Given I'd had a stable connection for about 7 days prior to unknown events causing a loss of sync, that seems a little severe!!

[#4]

Were there a lot of CRC errors before sync was lost? The loss of sync might have resulted from the DLM changing target noise margin to reduce the rate at which CRC errors accumulated.

[#5]

Quote from: spraxyt

Were there a lot of CRC errors before sync was lost?

I don't know the answer to that one but I usually have a low CRC and HEC total. The FEC total is sometimes high but I guess that's not a problem.
Stats after 5 hours are attached

[#6]

That's equivalent to 1100 CRC errors per day; from my experience I think the DLM gets concerned if it's much above 3000 per day.

[#7]

Quote from: spraxyt

I think the DLM gets concerned if it's much above 3000 per day.

Unknown events caused another resync at 1am this morning at a slightly lower speed of 2976Kbps with a SNM of 9dB. The stats in the 9 hours since then show only a 'few' CRCs. The 22 Error Seconds might be a cause for concern though?
I'm monitoring with RouterStats today - so far the margin is 10.5dB with a few jumps up to 11dB.

[#8]

I wonder if some of these unexplained 're-syncs' in the early hours result from BT doing work in the exchange? For example is your exchange being updated for 21CN/ADSL2+ (perhaps some months away from Ready for Service)?
With regard to the error seconds I suspect the previous value (10ish) wasn't reset when the router re-synced. I doubt the value matters at that level.

[#9]

My exchange (EMHINCK) is already 21CN enabled. Some areas already have FTTC. I don't know whether any additional rollout is under way.
After 6 hours with SNM steady at 10.5dB, I've just had another resync (at a slightly higher sync rate, as expected for the time of day). The SNM is now wobbling between 9dB and 11dB. I cannot think of any reason as to why.

[#10]

If one looks closely at that plot it looks like there was a genuine burst of noise shortly before sync was lost with Rx noise margin falling substantially before becoming zero. The latter value was probably "no sample available" while the router was re-syncing.
The noise plot is more lively after re-sync though that might simply reflect small variations about 9¾ and 10¾dB with the router reporting the nearest ½dB. The upward step at 18:22 looks like a source of electronic noise disappearing. In my experience noise margin plots tend to be more lively in the evening.
Unfortunately seeing what happens is easy, as is speculating about the source, but actually finding the source not so.

[#11]

Did you spot it went negative? I don't think there's a "no value" plot on the graph as routerstats now shows a marker where that happens.

[#12]

Over recent weeks, I've noticed quite a few 2dB steps (both up and down) in the noise margin. None of these can be linked to events such as phone calls or timeswitches in my property. I think they may well be due to external events which makes them even more difficult, if not impossible, to trace.
Such events seem to be more often than they were. I just wonder whether our miserable dry summer has a part to play. That said, my line attenuation has remained at 56dB so there doesn't seem to be any dodgy cable joints anywhere.