Wiring + Stats question

I've seen a few posts recently mentioning star wiring and getting this changed. We're in a new build that I believe has star wiring to the extensions from the master socket.

I've got a SSFP Master Socket (like this), so does it matter about how the extensions are wired? Can the extension wiring still affect the broadband?

Currently sync at 1.625 Mbps, but did get 1.875Mbps for a while. Is 1.625 Mbps and the large numebr of errors acceptable given my lines stats, or should I be getting better?

Line stats: http://imgur.com/dv2kC0y.jpg

Wholesale checker results: http://i.imgur.com/c6vLqyx.jpg

Thanks,

Matt

Easiest way to see if the wiring is having an effect is to compare stats: first set when the extensions are connected and the second with bottom half of the face plate removed - that should disconnect all extensions. If an extension still work then you have a problem.

As for your stats ... the 59.9 attenuation suggests a long line and speeds are roughly right. The 3.4 dB SNR - that is low and could be the cause of errors. Forcing it to 3dB from 6dB may give an increase in apparent speed however a lot more errors will creep in and the net result can be that your overall speed is actually lower.

Assuming that is the only Master Socket, not one someone has added since and there is another somewhere, then gently remove the bottom half where the screws are. If there are wires from the back of the removed faceplate, and all extensions stop working when you plug a phone into them, then you don't have (what we call) star wiring.

Say there are three extensions. If each of the used terminals has three wires in, in a way that is star wiring, but it is star wiring from a junction box somewhere before the master that is problematical. From the faceplate terminals it is fine.

It would be an idea to upload a photo to somewhere of the back of that faceplate so we can check the wire colours and the terminals used. In particular, T2 and T5 are vital, T3 is redundant and can cause loss of speed, and the wire colours on T2 and T5 matter.

Any gain we can achieve on a long line like that is welcome .

A bit on the slow side for that attenuation. As mentioned by others, remove the faceplate on the master to isolate the extensions (check that there's no dial tone on them). Temporarily plug the modem/router into the test socket and then check the stats. If the downstream sync speed is substantially better, then there are some things that can be done to improve the performance. Simply disconnecting the ring line to all the extensions (at the master socket) can have a dramatic effect and costs nothing, but there are more sophisticate approaches.

Thanks to everyone for taking the time to reply, any improvement to the line would be fantastic!


We did have an issue when the line was first installed that the master socket was installed (by a contractor working for Openreach) to an extension. I raised a fault as we had constant disconnects, and huge variation in sync speeds. An Openreach engineer came out, spotted the problem and moved the master socket to the correct location.

Since then I've removed the ring wire (using the info on your site, RobertoS). After the fault repair and removing the ring wire the sync speed was mostly 1.8 (sometimes synced lower), but in the last few months it seems to have settled at the 1.625 figure.

I'll post stats from the test socket and pics of the wiring and hopefully something will stand out.

Edited by deleted (Wed 02-Sep-15 16:42:14)

In reply to a post by TheEulerID:

Temporarily plug the modem/router into the test socket and then check the stats. I

Test socket is not needed. On te version the OP has there is an RJ11 socket on the upper half - just remove the lower half and continue to use the RJ11 which alleviates the need for an adapter too.

Though if stats are still poor going to the next step of removing interstitial plate and testing with a microfilter into the test socket is worth while, the filters in the faceplate or one of its connectors may be bad.

In theory if they have the interstitial plate then the effects of the extensions should have been mitigated by the faceplate (interstitial) filters

I removed the front of the Master Socket last night, I think I've got a different Master Socket to the one MHC is thinking of, as I couldn't leave the RJ11 connected to the front and remove the bottom as the whole front comes off: Pic 1, Pic 2.

-Stats in test socket with filter: http://i.imgur.com/bAPlnFt.png
-Stats when back in usual socket: http://i.imgur.com/BQez6bd.png

Oddly enough I actually got slightly better speed (1.688 Mbps vs 1.656 Mpbs) when not in the test socket.


I also took pics of the wiring: http://imgur.com/a/4duRG#0

Just in case the pictures don't show it too well:
-In the bottom terminal: 2 wires (both blue with white stripes)
-In the top terminal: 2 wires (both white with blue bits)
-Not connected to anything: 2x orange/white and 2x green/white


Does this leave me with any options to improve the connection?

Thanks

Looks like your wiring isn't causing any issues with your BB sync then. The wiring on the socket looks fine and the stats are near enough the same as to make little difference.

There's no significant different in those stats. The slightly higher connection speed had a lower noise margin.

Did you check that you'd lost dial tone on the extensions (so the extensions were definitely disconnected). If so, then it's difficult to see where any improvement might come from.

I assume your locations hasn't been enabled for FTTC.

No it is a normal master socket, just the interstitial plate (the bit where the RJ11 lead plugs in) came area first in your case.

http://www.coolwebhome.co.uk/faceplate/ shows the way the small front half should come off.

But as others said it was enough to do what you did anyway.

No dial tone from any extension, so pretty sure there's nothing coming off the incoming pair before the test socket.

I managed to find a pic of my stats from back when I was getting 1.875Mps: http://i.imgur.com/hMisIJy.jpg

Guessing there's nothing more I can do then?

Sadly we've been left out of both BDUK and commercial FTTC rollout, despite our best efforts (www.cheltenham151.com)

Ah, thanks for that. Mine felt pretty stuck together and didn't fancy pulling too hard on it!

If you have one of those types of modem/routers where you can hack the target SNR margin you might squeeze another 0.5 mbps (at the possible expense of stability), but I rather think you are in the realms of requiring an uplift to the local BB infrastructure.,

In reply to a post by TheEulerID:

The slightly higher connection speed had a lower noise margin.

This may be a silly question, but the original stats had a lower noise margin but lower speed. Does lower noise margin not always mean higher speed?

Original Noise Margin.......: 3.4 - 1.625 Mbps
Test Socket Noise Margin..: 6.3 - 1.656 Mbps
Back in normal socket......: 5.4 - 1.688 Mbps
Stats from a while ago.....: 4.2 - 1.875 Mbps

Not always, the figures are very much a capture from a moment in time, so if there was a burst of noise when syncing you can have a low noise margin and lower speeds, when the noise goes away this may be reflected by you having a higher SNR figure.

On long lines like yours the signal is so weak that it may not take much variation in noise to cause a variation in connection speed. Even the choice of ADSL modem and firmware its running will have an effect.

Until recently I had an ADSL connection on a very long line.. its attenuation on downstream was in about 62db and that could get 3-400kbps so I would certainly say something is up outside the property.

I may have said it before.

Lower margin may mean higher headline or sync speed however it may also mean significantly more errors and thus a lower actual speed.

further my previous post regarding the lack of speed on your line, I have dug up some old diagnostics data from my old ADSL1 connection, it follows!

Mode: ADSL_G.dmt
Traffic Type: ATM
Status: Up
Link Power State: L0

Downstream Upstream
Line Coding(Trellis): On On
SNR Margin (0.1 dB): 38 60
Attenuation (0.1 dB): 595 315
Output Power (0.1 dBm): 181 124
Attainable Rate (Kbps): 4,848 996

Path 0 Path 1
Downstream Upstream Downstream Upstream
Rate (Kbps): 4,320 832 0 0

K (number of bytes in DMT frame): 136 27 0 0
R (number of check bytes in RS code word): 4 8 0 0
S (RS code word size in DMT frame): 1.00 4.00 0.0 0.0
D (interleaver depth): 16 4 0 0
Delay (msec): 4.00 4.00 0.0 0.0
INP (DMT symbol): 0.23 0.11 0.0 0.0

Super Frames: 44,899 44,899 0 0
Super Frame Errors: 42 1 0 0
RS Words: 3,053,132 763,045 0 0
RS Correctable Errors: 20,125 1 0 0
RS Uncorrectable Errors: 276 0 0 0

HEC Errors: 92 0 0 0
OCD Errors: 0 0 0 0
LCD Errors: 0 0 0 0
Total Cells: 7,776,749 0 0 0
Data Cells: 283,832 0 0 0
Bit Errors: 0 0 0 0

Total ES: 20 0
Total SES: 0 0
Total UAS: 0 0

Thanks for all the replies.

The connection has now, annoyingly, dropped to 1.5Mbps, not sure why as nothings changed.

Are there any recommended routers / modems that work better on longer lines / allow the hacking of the SNR Margin as mentioned?

There's some mention of routers here. Some allow you to change the target SNR margin "legitimately", in the sense the official software allows you to change it. In the case of the Netgears there was some hacked firmware.

http://www.increasebroadbandspeed.co.uk/SNR-tweak

Note that any hacking of the target SNR margin is not legitimate in the sense that the modem has to tell little lies to the DSLAM during negotiation. The DSLAM will specify the target SNR margin whilst you modem will secretly set it to something else.

Note that all this only works on the downstream side of course.

With a hacked NetGear I used to be able to reliably run at 3dB SNR margin, but my line was running at about 5mbps and was stable.

Another word of warning is that in setting the target SNR lower it may be prones to more dropouts and line errors which, whilst corrected, can still cause you performance problems.

In all, it's potentially expensive. However, perhaps you might have a look on eBay. I suspect with the advent of VDSL2 there are a lot of second hand models around.

They may be old however the 2wire 2700-HGV has an excellent modem on long lines.

With a long and potentially noisy line, tweaking te margin down to 3dB may give you a higher "apparent speed" however you can end up with increased errors and overall no improvement.

Invest £5 to £10 in a 2700 (not a 2701) and try it out. If the line is stable, the DSLAM and modem will negotiate a 3dB margin as a matter of course. When I was on ADSL, my line went to 3dB and stayed there for years and even as the SNR changes during the day the modem would hold up even when the SNR dropped to zero.

I fitted a 2700 to a line running at around 800kbps, it is now giving around 1.4 to 1.5 Mbps.

I still think you have got something else causing issue on the line here... as above I have got more speed from a longer and noisy line so yours should get more as it appears shorter!

Thanks for the tip, just ordered one off eBay

I'll see if I can get hold of a corded phone and try a quiet line test. There is some noise when using the wireless handset, but I put that down to it being wireless.

If there is some noise on the line, can i raise this as a fault (without risk of being charged for the call out)?

Should it be completely quiet, or is there an acceptable level of noise?

Unplug all ADSL hardware including microfilters and if still a noisy voice line then report it to the people you pay telephone line rental, don't mention broadband issues, just deal with it as a noisy phone line.

When engineer arrives unplug the ADSL hardware again so they can see its nothing to do with the broadband

Do not know who your ISP is... if it is not BT you may need to unlock teh 2700. Details are on this site: http://bt2700hgv.tripod.com/ir1002700HGV.htm

TO use on any ISP use the page=hurl 13 unlock method....it has never failed for me

http://bt2700hgv.tripod.com/063.htm

Scroll down to "Simple instructions for BT 270x models to be used on OpenReach 20CN/21CN broadband connections."

Ian

Thanks Ian / MHC, I'm with BT so should hopefully be fine for now, but good to know for when I change ISP.

Finally got time to upload the stats from the 2700HGV, sorry for the delay. Unfortunately it doesn't seem to have helped much, but it does seem to provide a lot more information than the homehub.

http://imgur.com/a/K2Ao4/layout/horizontal

Does any of the extra info help identify what may be causing my sync rate to be lower than it should be?

Thanks,

If you are on a new build maybe time to meet the neighbours if close by/ on same line and
ask what speeds they get.