internet keeps dropping after 10pm

[#1]

hi and thanx

does this mean its a line fault or a "dont know what it is fault "

im getting more confused by the minute

ta

this pic dips

[#2]

I gather the symptoms, all the disconnections, are now the same as they were before the most recent Openreach visit? If so, there shouldn't be a problem sending Openreach out again.

A few dips in a graph of the SNRM can't really tell us what's causing them. Personally, I prefer to have one graph covering the whole day.

[#3]

@glesgajohn,

The problem at the moment is that we do not know for sure, which is why I asked @Chris for a review ... do wish that @Pettitto still covered this space!!

I can imagine that you are finding all of this a bit confusing - we are in the twilight zone of internet stability issues.  Those sizeable spikes are -3dB - the reason that the line is now appearing stable is that the TARGET SNRM looks as though it has been raised from the normal 6dB to 9 or 12dB.  Consequently the noise spikes do not cause the line to drop any more ... but that is all at the expense of sync speed.  I hope that one of the PN gurus picks this up and can offer a way forward, other than you wandering around with an AM radio.

In the meantime, watch the graphs, monitor the spikes, see if there are periods when there are none, see if there is an approximately regular interval.

@ejs is quite correct when he says that a bad line is more susceptible to problems with noise, but the noise spikes seen in that last (beautiful) graph are far from helpful in terms of addressing the drops*.  I'd welcome a more detailed assessment from @ejs though I'd suggest that if you can locate the REIN source and have it eliminated, that course of action will be more beneficial than making the lines less susceptible to the still present noise.

*That last plot shows repeated clean SNRM spikes of a similar shape and similar duration and to me clearly points to a switch source.  It could be the central heating, a fridge / freezer, laptop fan....

[#4]

---

@ejs wrote:

A few dips in a graph of the SNRM can't really tell us what's causing them. Personally, I prefer to have one graph covering the whole day.

@ejs - as ever different people have different experiences and consequently have different preferences.  Personally a 2 hour graph gives granularity to the events - at the sampling rate chosen, it can even show the shape of the RFI and deliver a precise time.  I do not know if you recall my painful REIN issue from about 5 years ago, when with the detailed plotting, I finally managed to correlate the sizeable spikes with the West Coast Main Line train timetable.  The shape of the spikes even indicated the direction of travel of the trains.

---

[#5]

@Townman - I think you are missing my point about noise susceptibility. The point is, pretty much everything emits noise. To a normal line, the vast majority of the noise will have no impact, although you might be able to spot things on a graph if you look hard enough. To a faulty line, the noise from all sorts of non-faulty perfectly normal apparatus can have a severe impact. Fix the line, and the same noise levels then have pretty much no effect.

[#6]

for the record , I stay in a culldesac no trains buses or traffick, the electrics plugged in when router drops and is stable are tv/sat.box/ and laptop, i have tried unplugging them one by one when thr router drops, there is no difference with them plugged in or not plugged in, posted a graph unstable (i think )

Its 7pm i will run configstats until 9pm and post the graphs if thats ok 

thanx

ps

would i get this from the guys router upstairs from me ,? he is on plusnet to

nothing has changed in my house since it was running stable, all is the same as it was

[#7]

Since we know your upstream speed is a problem, I think we're going to need the graphs for the upstream SNRM and speed as well.

[#8]

OK a few more bits of explanation.

The spikes are intermittent - you will need to unplug things and WAIT for a while to see if there are no more spikes or the graph shifts up.

Something happening "upstairs" could be the cause - do I conclude that you are in an apartment block of some kind?  If yes that's grim news as the problem could be close by well out of your reach.

Though nothing has "changed" in your home, current switched power supplies DO DETERIORATE OVER TIME so something which has always been there can start to be a problem.

[#9]

im posting this graph because it has started to drop again (a few times ) when it dropped the internet light on the router went to red then it booted back up then went to nred then it worked 

@Townman

Something happening "upstairs" could be the cause - do I conclude that you are in an apartment block = yes=of some kind?  If yes that's grim news as the problem could be close by well out of your reach.

Though nothing has "changed" in your home, current switched power supplies DO DETERIORATE OVER TIME so something which has always been there can start to be a problem.= i am using a different power pack on router, I have switched of sat box @ wall and tv just using laptop @ mo no power supply (just battery)

[#10]

can you cap the graph to say 15dB as the spike when it resyncs makes it useless for diagnosis

[#11]

and this one , sorry

[#12]

is that the Rx synch speed , change from 12000 to 15000 , is that what u mean ?

[#13]

nope it is the Tx noise margin

[#14]

im using routerstats 6.9 , where exactly do i set the Tx noise margin, sorry 

[#15]

http://www.vwlowen.co.uk/internet/routerstatshelp/routerstatshelp.htm shows it

I assume that you have set it up to monitor Rx Tx Noise Margin but you have it set to Adjust Axis Automatically

Edit - error corrected