What a difference a sync makes

[Rik]

Prior to my 300 hour re-sync, I'd been seeing more noise on my 2700 than I was used to, but had put it down to someone using an electrically noisy air-con. However, since the re-sync, the graph has gone back to it's normal flatline state. I'm guessing that, previously, I had picked up a noisy frequency. Just goes to show how temperamental ADSL is though...

[attachment deleted by admin]

[Sebby]

It does. Isn't that the point of bitswapping, though?

[Rik]

It is, though it looks like it wasn't working that well. It's a pity I hadn't capture the distribution graph.

[Sebby]

I know different routers handle it differently. I thought the 2700 was one of the better ones. Still, there's little logic to ADSL anyway.

[Rik]

There's lots of it, all fuzzy.

[Lance]

If only my noise margin line was flat like that! I've got a target noise margin of 6db, but you would never guess that from looking at the attached graphs.

[attachment deleted by admin]

[Rik]

Are you sure you've not got a Netgear connected there, Lance? 

[Lance]

Yep! I just have a problem on my line which BT tests aren't picking up. Here is what happens when it is actually synced at the proper speed (about 3mb higher than in the above graphs). The second is where it resynced with the noise on the line.



[attachment deleted by admin]

[Rik]

That's not a problem, Lance, that's a big problem!  You'll just have to move out from that power station. A 13db swing on a 2700 is crazy...

[Lance]

The problem is that it is intermittent and I know that if and when the BT engineer comes out that there would be no fault - £180 thanks very much!

I might get on to BT voice and mention intermittent crosstalk (not that we have any) and try to get a new pair...

[Rik]

Worth a try, from what I've heard. Don't you just hate intermittent faults.

[Lance]

I've got a voice engineer coming Thursday morning. Hopefully if I get a new pair, they will be better than the current one. 

[Rik]

Good luck. 

[Sebby]

Good luck, I'll be interested to see if this helps.

[MoHux]

.............. I have never used a laptop, or wifi.  So please forgive my naivety, but ............ if Lances results are as he showed when connected through his 2700, and assuming he is using his laptop, what sort of result would he get if he connected through a hot-spot??

Would that show better where in the network a fault might lie? 

Does that make sense, or shall I keep 

Mo

[Rik]

The problem would be, Mo, that it would show a completely different network, so it wouldn't serve to identify anything on his IDNet connection - if I've read you right.

[Lance]

I think you have, Rik.

To expand - it is irrelevant whether I am using my desktop or laptop, as the results are coming from my router (obviously connected to my line). If I was to connect to a hotspot, it wouldn't be my router (and in turn line) that I would be connected to.

[MoHux]

Thanks both ............. I get these 'bright' ideas sometimes.

I had been thinking, we always assume (probably rightly), that the fault lies before the signal gets to the exchange, and  I was wondering if there was any way to pinpoint just where a fault may lie, using the laptop and WiFi. 

I'll go back to sleep now .......... 

Mo

[Rik]

If there was an IDNet hotspot, it would be a great diagnostic, Mo. Sadly, they haven't got around to that yet.

[Sebby]

I had been thinking, we always assume (probably rightly), that the fault lies before the signal gets to the exchange, and  I was wondering if there was any way to pinpoint just where a fault may lie, using the laptop and WiFi. 

I know what you're saying, Mo, and it would be one way to attempt to find where the problem lies. Currently, assuming there is a NTE5 master socket, we are able to (with a fair degree of certainty) see if the fault lies in the customer's property or outside, but not any "further" than that.

[Rik]

I demand access to a test socket at the exchange!!

[Sebby]

That would be nice.

[Rik]

I'm just pondering how to drag the reel of cable back here without causing traffic chaos.   

[Lance]

50ft high poles, obviously!

[Rik]

But Sue would struggle to drag the cable up those.

[Lance]

You could get sue to attach the cable to the top and then erect the pole after! 

[Rik]

Sort of 'ship in a bottle' style?

[MoHux]

Ferret and a long bit of string, down the nearest drain??  

Mo

[Rik]

Nice thinking, Mo. How are you at training ferrets?

[MoHux]

NP, they are doing what comes naturally ............. the trick is getting it to come up the right hole!! 

Mo

[Rik]

That was the bit I was thinking about, Mo, that and following the shortest possible path.

[Lance]

Here is my noise margin graph this morning...

Shows how my fault is intermittent. It's also synced about 2-300Kpbs higher than usual!


[attachment deleted by admin]

[Rik]

Looks like mine usually does, Lance, and clearly demonstrates a fault (or intermittent noise from an external source if you're very unlucky).

[Lance]

Right ho - the engineer is off to change the pairs between the exchange and the green box (the E side). He's hopeful that will sort my problem out!

I just hope I get a better wire!

[Sebby]

Fingers crossed!

[Lance]

One thing the engineer said is that apparently there are quite a few problems with the E sides from my exchange. Don't sound too good does it!

[Rik]

No, it doesn't. I suspect that's an increasing issue with BT, it seems clear to me that there has been a lack of investment in the local loop for some time.

[Lance]

I'm now on the new E side. So far everything is nice and stable, and I'm synced at 5632. It's never normally as stable (the noise margin that is) at this sync. So far so good!

[Rik]

That is good news, Lance.  Of course, this means that your line is even better than mine now. 

Any change in attenuation?

[Lance]

Only a increase of 0.3 in attenuation, so no big change. Previously, my line wasn't too happy at anything above 5200, and only truely stable around 5120. It'll be interesting to see if this lasts and what it settles down to overnight.

[Rik]

Keep us posted. How's routerstats looking?

[Sebby]

Sounds hopeful, Lance. Let us know if it continues.

[Lance]

Keep us posted. How's routerstats looking?

Very flat! It's not budged off 6db ever since it came back after the engineer had finished.

Sounds hopeful, Lance. Let us know if it continues.

Don't worry, I will. I'm expected a couple of resyncs tonight, but hopefully it'll then find its stable level and I won't see any noise bursts.

[Rik]

It's sounding good so far though, so you may not get any re-syncs...

[Lance]

I'm expecting one or two as darkness comes and noise increases. You never know though, I may be lucky!

[Sebby]

Darkness should only cause it to fall by 1 or maybe 2dB, so it may well hang on now.

[Baz]

all this talk of re-syncing and stuff, do you all turn off your routers at night. is it best to or just let them run. is it worth while shutting it down every so often and do you benefit by doing that?

[Sebby]

Personally, I prefer to leave it running 24/7. Ultimately, I think that's what a router's for. Plus, even though I know it's not the case, I feel my line might be perceived as unstable if I run into problems and they see a sync every day.

[Rik]

I'm with you, Sebby. I only shut my router down when there's thunder around or when I'm away. (And every 300 hours, of course. )

[Baz]

300 hours because...........

[Lance]

Rik is running the SBC firmware on the 2wire 2700HGV, which for some reason initiates a reboot or resync every 300 hours

I run my router 24/7.

Over the last half hour or so, my noise margin has been going between 5 and 6, and now seems settled on 5.

[Lance]

Update on my line...

Noise margin dropped to 4db at 8:35pm and to 3db at 9:13pm. It dropped to 2db at 9:49pm and that is where it is quite happily sitting at the moment. On my line, at the current sync, that is amazing!

[Sebby]

It looks like it's all sorted then, Lance.

If you want, you could force a re-sync at night so that your SNRM is 6dB at worse, but if there are no problems, you could equally leave it alone.

With regards to the 12 day/300 hour thing, it's something that Rik and I (and undoubtedly others) experience with the single SSID 2700HG running SBC firmware. There doesn't seem to be any real reason for it. I've now got myself a dual SSID with BT firmware, and it doesn't reboot.

[Rik]

The 2700 with SBC firmware re-boots itself every 300 hours (12.5 days), Baz, which means it often picks up poor speeds on alternative re-boots due to time of day.

[Lance]

Further update, my line kept sync all night, and is now happily back to the usual 6db. Fantastic!

Shame it's not quite high enough for a 5mb profile!

edit: I've had 3,570,373 FECs and 2,245 CRCs since 15:10 yesterday.

[Rik]

Now, now, don't get greedy.

[Sebby]

You can usually try for the next profile, but stability is more important!

[Lance]

Indeed. I don't think I could sync any higher and be stable. I'll just make do 

[Lance]

Again, last night my line kept sync all night at the high sync. The noise margin dropped as low as 1db at around 10:30pm for about 10 mins but then went back up to 2db.

It does seem that getting my exchange side cable swapped has done the world of good!

[madasahatter]

The 2700 with SBC firmware re-boots itself every 300 hours (12.5 days), Baz, which means it often picks up poor speeds on alternative re-boots due to time of day.

Does anyone know exactly why it does that, or is it just one of lifes great modern mysteries? 

[Rik]

It does seem that getting my exchange side cable swapped has done the world of good!

 

[Rik]

Does anyone know exactly why it does that, or is it just one of lifes great modern mysteries? 

We don't know why, Mad, it just seems to be something in the code. Possibly deliberate, probably accidental. It does trigger a full re-boot, though, not just a re-sync, so it's possible it was done for a reason.

[Lance]

Damn. At 10:22 this morning, my line dropped sync to 4832, but it seems I now have a 9db noise margin target 

[Rik]

BT working anywhere nearby, Lance?

[Lance]

It's possible, but I'm not going out today so won't be able to see. Too much revision to do!

[Rik]

Send Nicky on an area search.

[Sebby]

It does seem that getting my exchange side cable swapped has done the world of good!

Brilliant.

[Sebby]

Damn. At 10:22 this morning, my line dropped sync to 4832, but it seems I now have a 9db noise margin target 

I don't think that's necessarily a bad thing. Once you're used to a slightly lower sync, you'll find that you worry a lot less with a target of 9dB as there's so much more margin. Rik and I both have 9dB margins and wouldn't have it any other way.

[Rik]

Yes I would. I'd like to sync at 8128 with a margin of 15db or more.

[Sebby]

So true...

[Lance]

To be honest, a sync of 4832 is more than I used to get before I had the 2wire and as it's solidly in the middle of the 4Mb profile, I'm happy

[Sebby]

I get a very similar sync to you, Lanc (usually between 4.6 and 4.8Mb). What's your attenuation, out of interest?

[Lance]

At this very moment 52.3

[Sebby]

So it sounds like you're actually doing better than me, given that my attenuation is 47.5dB.

I also had problems with a target SNRM of 6dB and it kept shooting up to 15dB, so I had it set manually at 9dB and it's solid as a rock; the SNRM moves only about 1dB at night.

[Rik]

Same here, at 6db the line was flaky. At 9db, the most variation I see normally is -2db, but generally it's only -1db. Of course, the Netgear was a different story...

[Lance]

It's just a little annoying that it was holding sync at 5632, albeit with up to the -5db drop at night. Oh well. It'll be interesting to see what happens tonight!

[Sebby]

It is frustrating at first, but you'll prefer the long-term stability.

[Rik]

At least, that's we console ourselves with.

[Lance]

It is frustrating at first, but you'll prefer the long-term stability.

Indeed. I was never expecting it to be stable at 5632 anyway!

[Sebby]

Indeed we do, Rik. It's a shame, that. One day we'll all have fibre, but I don't know when (perhaps not even in my lifetime!).  

[Lance]

The engineer said that starting in August, BT will be running more comprehensive line tests as standard, and that they should be able to pick up not only where the faults are, but what is causing them as well. It will also be standard that once they think they have fixed the fault, they will run another of these line tests to check for anything else (as opposed to the standard line test).

He also didn't think that fibre to the cabinet would be happening any time soon!

[Rik]

He also didn't think that fibre to the cabinet would be happening any time soon!

I suspect he's right.

[Sebby]

That's a shame. FTTCAB should see everyone get full sync, regardless of there still being copper/aluminium from the cabinet to homes.

[plugwash]

No, it doesn't. I suspect that's an increasing issue with BT, it seems clear to me that there has been a lack of investment in the local loop for some time.

Really the whole system could do with being redesigned and upgraded. ADSL is a dirty hack to get far far more data rate than the wiring was ever designed for. Frankly i'm impressed that it works as well as it does.

Fiber to the cabinet is a possibility and I belive is used by BT in a few places in the uk but it has problems of it's own. It will make LLU impractical and also cause issues for BT deploying new technologies on the lines. For example people on fiber to the cabinet got ADSL far later than most others. IIRC virgin media also use a fiber to the cabinet system (but they use cable TV coax rather than phone wire for the cabinet to cusomter part of the broadband connection).

Fiber to the home is very expensive to install and so we are unlikely to see it any time soon.

[Tacitus]

....they should be able to pick up not only where the faults are, but what is causing them as well. It will also be standard that once they think they have fixed the fault, they will run another of these line tests to check for anything else (as opposed to the standard line test).

Sounds good on paper but what if, as is the case where I live, the problem is basically clapped out cabling and ducts which go through a low lying area and leak like a sieve .  It's not that long ago that every time there was a heavy downpour the phone system packed up at my end of the village. 

In this case it would involve money and largish quantities of it, rather than simply fixing a joint gone bad.

Methinks that in those circumstances BT are just as likely to turn round and say you can't have a viable broadband service and hard luck.  See you in 20 years and that's probably optimistic.

[Rik]

Indeed, until there is a USO for ADSL, BT are likely to do just that. The role of Ofcom seems to have been to ensure that BT have no incentive to invest in the local loop, and a good starting point would, imo, be scrapping this worthless NGO.