Print

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

[guest]

It would be useful to see toulouse. Your upstream attenuation is broadly similar on each line, you clearly don't have any local REIN-type stuff going on as the line which syncs faster is stable.

If the Plusnet line syncs up to a reasonable upstream rate then that will show its probably not local wiring and that line is routed differently to the Be/O2 line and/or there's some issue on D or E side cabling.

At that point its time to get someone to manually tweak the Plusnet line (alter INP/etc) so its stable.

For now we're guessing, all we can go on is the upstream rate is lower than it should be, the attenuation on d/s is significantly higher on the Plusnet line and the Plusnet line is not stable.

That could be a local wiring problem like a split pair/short in the communal distribution point or it could be the line takes a different route and/or has problems further down the line.

[guest]

It's the BT DLM system on the upstream surely? It doesn't look like a stuck profile to me and its not the old ADSLMax upstream rate of 448kbps so its likely something is generating lots of errors on upstream. toulouse also mentions the line isn't very stable to say the least

I reckon he has local wiring problems.

If you look at the quoted stats you'll see that the upstream noise margin is very high, and the upstream speed is 440 kbps in both sets of stats. That makes it more likely that the upstream speed is capped by Plusnet, and not the result of a very noisy connection (in my opinion).

I don't know about the BT DLM but the Sky DLM will "cap" upload if it sees a certain error count per hour (to be more accurate it progressively increases the INP until the error count falls below the threshold). That results in an apparently high margin which (almost) invariably leads to local wiring/REIN issues being the cause of the DLM problem. Downstream DLM on Sky is just broken

Its unlikely to be local REIN as the O2 line looks fine and if toulouse hadn't said the line was unstable I'd have just thought "one of those things".

Wiring still looks favourite to me.

[toulouse]

I have now requested, via my outstanding Support ticket, the uncapping of the Upstream rate, which I think will take a couple of days to implement

In the meantime, for those of you who haven't yet fallen asleep with boredom from this thread - the Routerstats Noise Margin trace is still running at something between 6 and 7. There was a small increase from around 6.3 to 6.9 for a while earlier on, but other than that it all looks pretty flat. I'll keep that running overnight and report back with any other changes.

[guest]

If the line is unbalanced (ie wiring problems) then I'd expect to see your Plusnet line have noticeably lower noise margins after dark than the O2 line.

Simple reason is that after dark MF radio bounces off the ionosphere and travels further which increases interference to ADSL signals - especially overhead lines.

Now if the line is nicely balanced the MF "noise" (radio) on both the A and B wires of your line should be in-phase and will be largely cancelled out at the input of your router (common-mode rejection ratio - CMRR is the techie term).

If the line isn't anywhere near balanced then the "noise" on each wire will be out of phase and will cause problems.

Anyway lets see what happens when Plusnet do what you ask.

[toulouse]

Just a  quick update for anyone who may still be interested.

I have had a message back from Plusnet, advising that an order has been placed to uncap my Upstream rate, which I'm told should kick in on Monday.

No spikes or major changes in any of the Routerstats data, but I will continue to monitor it, and report back with anything which I think is significant.

Thanks to all for your input and suggestions

TTFN



toulouse

[guest]

Will be interesting to see :

a) if your upstream increases a lot;

b) if it makes the line less stable.

The problem with DLM/profiles etc is that it makes any sort of "forum diagnosis" a real pain. I've helped people on the Sky forums who (from router line stats) seemed to have a maxed-out line. Noise margin looks right etc. In a lot of cases it turns out the Sky DLM/profiling system has slapped on ludicrous levels of INP (Impulse Noise Protection) and when that gets changed the stats look a whole lot different.

INP is only useful for UDP traffic anyway, TCP will retransmit so at the TCP/IP layer it'd affect DNS lookups, gaming and a bit of streaming for most people. DNS will retry, games are normally designed to cope with packet loss but streaming usually ends up rebuffering as the video is compressed and can't recover until a key frame is received. I'm not convinced INP is of any use in anything other than streaming, which of course is the holy grail for Sky (satellites are fragile, expensive and can't be repaired commercially). Set INP and encode video key frames/buffering right and you'd have a rock solid platform with the low overhead UDP brings.

Sorry went well off-topic there

[toulouse]

Hey rizla, no problem. Your input is very interesting - you will excuse me for saying that you certainly seem to be a REAL techie.

I guess we'll have to wait until Monday to see what my line does once the brakes have been taken off the Upstream. Do you think that I may see an increase on the Downstream as well, or shall we just wait and see ? Just out of interest, within the Plusnet Member Centre, under Connection Details - High Speed Broadband up until quite recently my profile has been showing as something like '10.1'. I did notice yesterday during my many dips in and out of that page, that the same 'profile' is now showing as '11'. I'm not quite sure what that relates to, as they seem to have changed the format of this indicator some time in the recent past. It always used to show the BT Bras profile (or something very similar), as far as I'm aware.

Anyway, thanks for your remarks, for now.

[guest]

I used to design military radar systems - ESM/ECM/missiles*. Then I did some programming, then telecomms, then network security then datacentre work, then hotels/whatever I could get, and so on. The UK is largely screwed for anyone doing electrical/electronics design activities these days other than aero engines in Derby (problems looming), some foreign-owned "defence" companies and some ESA stuff on the South Coast. Go North of Brum and there's virtually nothing now. Last design work I did was for a German supermarket chain putting a datacentre in Belarus. Not something I'd recommend. до свидания pet or what?

I dunno what I'd call myself these days, certainly not an engineer. A tinkerer maybe

*anyone in London with Rapier FSC missile batteries nearby for the Olympic nonsense - I redesigned that so that two missiles in the air at once actually works. However if the link from the command transmitter fails then things get a little "hairy". Pretty sure google will lead you to some lovely videos of the range at Benbecula with missiles "looping the loop". Rather you than me

[toulouse]

Nice one !!!

[toulouse]

Hiya guys (and gals if applicable),

It's time for me to 'fess up. This morning I plugged a corded phone into the extension socket, which I had previously been told by 2 separate BT engineers was dead. And guess what.....yep, there's a dialling tone on that socket. This kind of suggests to me at  least that that socket is something other than 'dead'.

What to do now !!!

Perhaps, I'll wait until Monday when the uncapped upstream rate kicks in, and then see what's what. Or maybe I ought to get a small screwdriver out and have a poke around - what do you all think ?


TTFN

toulouse

[HPsauce]

get a small screwdriver out and have a poke around

Definitely. ASAP.

But first dial 17070 from that socket and check what line it's on.

[toulouse]

I've had another thought (with my brain) - how about a filtered faceplate ? I did use to use one obtained from Clarity some years ago, but during the course of trying to track down various problems on this particular line it was removed and never returned. I'm right in thinking that that would 'automagically' bypass any extension wiring, aren't I ?

 

[HPsauce]

It depends where the extension is wired in. Not unknown for it to bypass the faceplate. 
17070 and screwdrivers are the best initial investigatory tools.
Fit the solution to the problem, once you've identified it. 

[toulouse]

Ok, thanks HPSauce. I'll take a little look as you suggest.

[guest]

If you don't use the extension (which you obviously don't) I'd be inclined to disconnect it at the NTE. If that improves the Plusnet line then you can have a think about whether you actually want the extension for future use.

DECT phones have decent coverage, I ripped out all out "internal" wiring years ago and just got a DECT pack with 3 handsets.

I wonder if the bell wire is connected on the "rogue" extension wiring, it'd explain a lot.....