Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[drmrbrewer]

I'll investigate the wiring in more detail.  In the meantime, routerstats lite looks useful, to see how things vary over time.  But I can't get the thing to work.  The System Log panel shows a series of "Undefined situation" lines.  The plots just run along zero.

Mike

[waltergmw]

Hi Mike and Good Morning,
That might be because you haven't set in the correct modem and password log-on details.
If you are still stuck, ask again.

House wiring is a very common fault with poor broadband performance.
You'll need to ensure your side is squeaky clean before poking the lumbering elephant !

Kind regards,
Walter

[drmrbrewer]

Hi Walter

I'm pretty sure that I had everything set up OK.  Modem/router is DG834Gv5 (which should be a benefit, seeing as "RouterStats is primarily designed to work with the Netgear DG834 series"), I double checked the login details, and the URL was fine -- it gets me to the router login page.  If it were something simple like bad login details, I'd expect a more informative error message than "undefined situation".  

I was wondering whether I perhaps have some firewall or security setting which is preventing the program from gaining access to the router; I'll check it out later but any tips would be appreciated if you think that might be the issue ;-)

I still keep going back to this thought: the suggestion is that I should double check my wiring / extensions / filters / etc, but isn't my modem/router sitting at the end of all of that -- at the end of the nest as it were.  If the wiring setup is bad, shouldn't my router stats be telling me that?  All indications are that my router is seeing a clean line.  So, surely that means that my wiring / filters / etc are all OK, in combination?

I did wonder whether perhaps it changes drastically over time -- sometimes it's OK but other times not.  Which is why it would be nice to get routerstats up and running.

Mike

[waltergmw]

Hi Mike,

A router can synchronise at a reasonable speed but will be affected by noise bursts or spikes.
Making sure you are as close to the house entry point is a good idea, and checking with the house wiring disconnected will verify that you configuration is as good as possible.
Once that has been absolutely proven you are then in a much stronger position to ask your ISP to investigate REIN (i.e. noise being picked up on BT's wires).
If you use this term it should get an engineer to look at all forms of other line difficulties as well, including cross-talk.

Kind regards,
Walter

[HPsauce]

my modem/router sitting at the end of all of that -- at the end of the nest as it were.  

That will give every excuse for BT to blame your wiring.

You MUST have the router in the BT Test Socket when running diagnostics, so that your wiring is disconnected and eliminated

[drmrbrewer]

OK.  I got routerstats working.  Ironically, it was easier to get the full version working than the "plug and play" lite version.  And the traces aren't very interesting so far.  Completely flat-lining.  No variation.  Not one pixel.

And here is a little more background on the line to our house.  I'll try to say more than less, in case something apparently trivial is actually significant.

Along our road near our house there are a couple of telephone poles, with a star configuration of wires feeding out to the various houses near each pole.  This is, I guess, pretty normal. 

My house and that of our neighbour (the one I mentioned as having a better broadband speed than me in the original post!) are served from the same pole.  Their house is admittedly closer to the pole than ours, but I doubt that would make a huge difference: we're talking at most 50 yards difference here. 

Our line from the pole is supported on the corner of their house before arching across to ours.  From our eaves, a cable runs down our wall to a rather ancient looking weather-protected junction box, entering from below. 

Another cable runs down away from the box junction box from its below, and straight down into the ground.  No evidence of any other cables emerging from the earth nearby.

Opening up the junction box is easy: there is a large thumb screw on the front; no screwdriver required.

Inside this box are eight junction contacts, but in fact no wires are connected to these.  Instead, the two cables entering the box are merely connected one to the other, with the connections using the weather protection of the junction box.

The connections (or not) of the wires are as follows:

Upper Orange to Lower Green
Upper White to Lower Blue
Upper Green and Blue not connected to anything
Lower Orange and White not connected to anything


I am not sure exactly how the wire from outside reaches the phone point to which my router is connected (it's not on an external wall).  Our digital phone is connected to a different phone point in the next room, and again it's not apparent how the line gets there -- again it is not on an external wall.  The cable outside disappears into the ground, and magically pops up somewhere in the house.

I've tried turning off and disconnecting the phone completely.  This didn't have any effect on the routerstats trace.  Not sure that it would.

Nothing else in our house is connected to the phone line.  There are a couple more phone points around the house, not all of which work.  Again, it's not apparent how the wiring gets there.  This is a 1960s house extended several times over the years.

So, that's it for now.  The only thing which strikes me as a bit odd are the connections described above: is it really normal to connect orange to green and white to blue??  The phones all work OK, so it can't be too drastically wrong, but it still seems odd.

Thanks,

Mike

[waltergmw]

Hi Mike,

So you have established that you only have one of the two pairs in the drop cable from the pole connected.
As you suggest line distance is probably irrelevant.
Just to be sure, can you see if there are any other cables leaving your dropwire on the adjacent house?
To progress a little further I suggest you look inside all your sockets that have a dial tone and try to find out how they are connected, checking to see which wire colours are used in each socket.
You may also have one socket with a big yellow capacitor. This is the master socket and would be the candidate for a NTE 5 socket if BT could be persuaded to call.
It would be worth removing all internal wires connected except for those in pins 2 and 5. If you have very old sockets take some pictures and post them here.

In an ideal world it would be good to drill a hole through your wall near the external box and place a new NTE5 just inside, but this should be done by BT Openreach.

Keep an eye on Routerstats as well, reducing the sample time to the minimum safe value, say 4 or 5 seconds.

Kind regards,
Walter

[sevenlayermuddle]

There are a couple more phone points around the house, not all of which work. 

That rings alarm bells for me.

I'm just emphasising one of the points Walter made, which is to check the wiring in each of your sockets.  If it all possible, I'd suggest getting to the bottom of these non-working sockets.

The problem is, if the reason for some non-working sockets is a broken wire somewhere, then it is quite likely that only a single wire is broken, in which case the surviving wire (of the same pair) becomes completely unbalanced, and acts as an antenna, picking up noise and feeding it into the router.

The most likely place for broken wires is where they connect to the sockets, so a careful visual inspection of each socket may yield results.  But it could also be a damaged cable, e.g. when DIY extensions became popular in the 80s people did things like running wires under carpets where they get repeatedly trodden on and the conductors eventually fracture.

[waltergmw]

@ 7LM

Again, it's not apparent how the wiring gets there.  This is a 1960s house extended several times over the years.

Yet another reason to be suspicious of internal wiring.

Kind regards,
Walter

[drmrbrewer]

Thanks, guys.  I'll do some poking around internally.  One particular oddity ("feature") we've noticed since being in the house is this.  There is one particular doorway (between the room with the router and the room where the phone is plugged in) which, when you tread on a particular spot on the carpet, a strange "plunk" noise is heard.  We haven't yet got around to investigating, but have always thought it might be phone related.  I'll dig up the carpet later to see what it is.  What on earth would cause the "plunk" noise I really don't know.  We can locate the spot on the carpet which causes this to happen, but we can't quite locate where the sound is coming from.  But if a telephone cable has been under a carpet in this area for many years, it's quite likely that by now it has worn through or been severed.

Will keep you posted.

Mike

P.S. we have plans to completely rip the house apart, alter and extend it -- but this time not in an adhoc fashion as in the past, but to pull all the random bits together.  Almost like starting again -- which appears to be the best way to go about this problem really.  Sever the phone connection at its point of entry and completely re-wire from there, either leaving the legacy wiring buried away, or strip it out completely.

[waltergmw]

Mike,

Yes an excellent idea to start again with such old wiring.
It's also much better to physically remove redundant wiring as it will all look very similar after a couple of years.

Floor noises are probably coming from floor boards.
Although the very unadventurouse hide wires under the carpet, most would probably try to put wires in the under-floor void.

Kind regards,
Walter

[drmrbrewer]

That "plunk" noise isn't a mechanical sound as far as we can tell, like creaking floorboards; it's electrical.  As if there is some malfunctioning and very puny speaker in there somewhere, and being activated (though not consistently) by applying pressure to some wires under the carpet; like some very very poor guitar simulator in a cheap kids electronic toy.  Yes it's very odd, but we have heard that a previous owner of our house was very into his DIY electrical projects.  Who knows what he installed here.  Having said all this, it does rather explain why our neighbour might be getting a better broadband speed than us -- assuming these redundant / peculiar circuits can actually have a dramatic effect on broadband performance.

Incidentally, what is an acceptable noise level in the router stats?  Is mine on the high side (see stats posted above)?

Mike
 

[sevenlayermuddle]

Incidentally, what is an acceptable noise level in the router stats?  Is mine on the high side (see stats posted above)?

There's no easy answer to that.  What matters is the ratio by which the signal exceeds the noise, and the margin by which the required Signal to Noise ratio is exceeded, the 'SNRM'.   A high SNRM is basically 'good' as it suggests the noise is low compared to the signal.    The stats you posted near the start of the thread actually show quite a high SNRM at the moment you captured them, but that's not to say it was so high at other times. 

I see there was some debate earlier in the thread as to whether you had been put on a fixed rate.  Was that ever resolved?

[drmrbrewer]

Thanks.  Given that my SNR is good, I'm still perplexed as to why my speed is so poor, however quirky my wiring.  I will however continue to investigate to see whether there is anything I can do to improve the situation.  And yes I'm pretty sure that I've been stuck on a fixed rate, apparently for my own good!

Mike

[waltergmw]

Hi Mike,

If you can do another BT Speedtest and post the full results that should help determine whether you're on a fixed speed or not.

You latest modem stats would also be useful.

Kind regards,
Walter