Disconnections and slow speed

[#1]

Hmm, yes Jim. I think I recall seeing something similar a considerable while ago, but atm can't remember where or what it was.
Barry, yes that would be very helpful. Leave the setup as it is for now, so we can see what goes on from late this afternoon right through to tomorrow morning. Shame you haven't got the graphs from yesterday evening's later experimenting with the power supply swap, but no worry - and don't bother doing it at the moment, would prefer to see some stability. One question though about the graph on reply #98. Roughly how long was it on battery and what times was that?
Another couple of tips for you -
if you maximise the Routerstats window (you can still minimise or restore it from the taskbar/tray depending on your other settings) but keeping it as a maximised window even when it's in the tray, means the captured files will also be maximised hence showing a bit more detail which may (or may not) be handy.
The other, on the Routerstats Configuration button >Graphs >Extras tab - put a tick in the top Averages box "Plot average Rx noise margin value of the previous ...."  100 samples. That could also be useful. There is no point at this time (or probably likely in the near future) in plotting the others or averages.

[#2]

Hello Anotherone,
Re. post 98: Where the graph shoots up, that is where I unplugged the old adapter. I left if for just a couple of minutes and then plugged in the new adapter.
I've set the 'average' setting now as you said and maximised the window but it seems to just scale the window rather than show more info. Not sure if that's what you wanted.
One other thing I've changed is the router to PC cable. I noticed that by moving the cable at the back of the router, the connection light would go off. It doesn't do it with the replacement cable from my main PC so I've left the new cable in. Will have to get another cable, though I suspect there are cheap rubbish ones that should be avoided. Any tips on a reliable source would be helpful. Don't think it's affected the results though. Still get the sine-wave effect but over a longer period than before.
Routerstats has been running since about 3 o' clock this afternoon. I plan to leave it running for 24 hours or more, hopefully that will show you what's going on. Thanks for your help,
Barry.

[#3]

Hi Barry, because the window is bigger, but your SNRM scale is the same, the SNRM variations will show more clearly which can sometimes be useful. For example if you look carefully at the 2nd graph on reply #103, you'll see a step down in the SNRM at ~0744. This is probably due to something being switched on, maybe a neighbour/someone on the same DP/cable back to the Cab switching on a modem/router resulting in some increased crosstalk which is "noise" as far as your modem/router is concerned. As I said, sometimes the larger view is useful, other times not.
What sort of monitor do you use? Is it a CRT type or LCD/LED?

[#4]

Hello Anotherone,
Okay thanks. As for the monitor, all the testing is done on an old laptop so it will be LCD and my main computer has a separate LCD monitor as well, but that's out of the circuit for now. No CRT displays.
Barry.

[#5]

Thanks for that Barry. I should have read all the detail I mentioned in reply #96 that I hadn't read! then I wouldn't have asked the last question.
As for switching the laptop screen off when you aren't going to be using it for lengthy periods, eg. overnight - I don't see a problem with that, just make a note of the time you switch it off and on and post that info with the graphs if there are step changes in the SNRM at those times.

[#6]

Hello again. Just done a quick recap of the thread and a few things I'd thought I'd ask -
Exactly which models of Netgears have you got, which one was in use when you first had the problems, which is in use now and is it the same one since you moved to the test socket?
Thinking about when your problem is hopefully resolved and you move back to the normal location, have you disconnected the bell wire (terminal 3) at ALL the Extension sockets as well as the NTE5.
One other question about filters. Do all your Microfilters look like this? In other words they don't look a bit like solid block socket doublers.

[#7]

Noise margin stats are given as an integer from Texas Instruments AR7 chipset based models by default, so they are almost certainly not from one of those models.

[#8]

Hello Anotherone,
My main router is a Netgear DG834 V4. The 'emergency' backup router is a Netgear DG834 V2. I only used the later very briefly (a few minutes) so I could hopefully eliminate the router as the problem and posted the results a long time ago, far-far away somewhere in this thread.
Funny you should mention microfilters. I've just been reading up on them and see that they are not all equal. The reason I was doing this is because something happened to the router stats after picking up and replacing the handset on the phone this morning. I was looking at the stats that have been gathering over night and the trend was that the noise dB was getting lower and lower towards the morning. I decided to pick up the phone just to listen to see if the line was particularly noisy as I had experienced a noisy line in the past as mentioned somewhere near the beginning of this thread, but this had gone since rewiring and removing an old phone from the system.
Anyway, the line was quiet and I replaced the handset. The stats however jumped up suddenly at this point to a higher dB value and remained there. A little later I received a real phone call and noticed the stats dropped down again while the call took place then rose again.
I've attached all the stats so far since last evening. These are the bigger full-screen captures that you asked me to provide. You can see on the last one where the phone experiment and call took place, don't think I need to point it out.
I have a question about microfilters but will post later.
Barry.

[#9]

Regarding the filters: They are like the type in the photo with the short flying lead.
I only disconnected the bell wire at the master socket.
I'm willing to buy a couple of decent filters just to eliminate them from the line problem. There's one called XF-1e which is claimed to be one of the best. It's expensive, it uses transistors and claims this helps separate the frequencies between the phone and ADSL. Is it all hype? I know at least one of mine is a cheap ebay filter.
Barry.

[#10]

The XF-1e is a quality filter BUT under some circumstances I've seen some users report that they've not helped or even made things a bit worse as the active components can be upset by strong RFI etc. My own thoughts are I'd tend to stick to quality BT MF50 rat's tail type filters from a reputable supplier. Only if things continue to be very troublesome and all other solutions fail might I then try an XF-1e.
As for the cheap ebay filter, well it depends what it was and how cheap etc. If you had stable conditions then you could at some future date do a comparison to see how it performs. Poor filters will in general result in slower speeds and more glitches with phone use.
Leaving the superimposed slight sinewave variation aside, the daytime to dusk to dawn to daytime graphs you've posted are pretty much what I would have expected to see. However the one showing the gradual drop in SNRM from about 0815 until when presumably you picked the phone up at ~0937 is certainly highlighting an issue. My suspicion is that it's a dirty connection somewhere and it could be anywhere from the filter onwards upstream on the line. When you pick the phone up, the current flowing tends to have a slight cleaning effect on a dirty connection. Incoming ringing current will have the biggest effect, sometimes cleaning a joint sufficiently for issues to go away for an indeterminate while, and other times may cause a loss of sync. Sometimes audible noise may be heard on the line and other times not. This type of issue can be a pain to deal with when intermittent.
If you don't notice any jumps/obvious changes in SNRM if you tap the filter or gently wiggle the connections and plugs, then it tends to suggest the problem is more likely to be external.
With the incoming call that you had can you just confirm the time and approx. duration.
With regard the signwave variation, have you got the follow-on graph ~1313 I'd guess. I'd like to see if there was any changes after the call.

[#11]

Thanks for the advice and suggestions. When I did the rebooting of the router the other morning as discussed earlier, I sprayed a bit of contact cleaner (de-oxit)) into the master socket and plugged the plug in and out a few times. I think I would have done the filter as well though I can't actually be sure that I did. I certainly didn't do the sockets on the back of the router or computer, so maybe I''ll try that next time I reboot.
The phone call was where that dip is about 10:38AM. EDIT: Less than 30 seconds.
I've attached the most recent stats and won't say anything about them other than that I didn't touch anything, receive or make any phone calls or even use the computer when the "thing" happened. It seems like one step forward, 2 steps back.
Barry.

[#12]

Dealing with the phone call one quickly. I had guessed it was a very short call from the graph but did want to check.
As for the next event OMG. something crazy has certainly happened.
Now it's entirely possible that, whatever the cause, there was either a large burst of noise for some reason (any lightning at all?) or maybe an engineer working in the Cab or exchange etc, disturbed you connection which resynced whilst the disturbance was still present.
Hence the SNRM of around 17.2dB from around 1330 on (to 1511 on that graph). Temporarily I suggest you go into Routerstats Configuration and on the >Graphs >Y limits tab, in the Rx Noise box, change the Maximum value to 20.
I'll post more in a moment.

[#13]

Can you confirm what your Upstream speed was before and since this event - that's Upstream (Tx) not the 4698 downstream.

[#14]

Thanks Anotherone,
No lightning, didn't see any or hear any thunder anyway. I've upped the scale to 20 now.  Here are the most recent tx stats:
Barry.

[#15]

Now in Routerstats graphs themselves (not the Configuration) you are obviously plotting Rx Noise (tick in the box bottom LH corner) and obviously Capturing, you've got the Green Camera icon above the tick.
Now if you change tab to the Tx noise, have you got it plotting (tick in the box) if not, then tick it. If you are already plotting and not currenrtly capturing, does the graph look nice and flat or does it have a lot of noise? If it has a lot of noise and you aren't capturing, then right click in the graph and select Scheduled Capture and then also select Capture Now - only if it's not dead flat. If it's not flat post the one that covers the "event" if you were capturing, otherwise wait until there's a fair bit of info on the graph, do a Capture now and then post it.
More to follow, and I'll just look at your last post.