Can CRT have a look at my Profile please.

[#1]

Having recently seen the post regarding use of the BT Home Hub 2 on Plusnet I dug out my old one as we have a Canon printer which, when we were on BT, ran very happily over its WiFi link. When we switched to Plusnet the TG582n just refused to see it and I ended up having to relocate it downstairs on top of a filing cabinet using a 5 metre ethernet connection. I switched the two routers over using the disconnect-reconnect procedure and, yes, it came up first time and is working well. When I last re-negotiated my stay with Plusnet, in October, I had got a new contract notification saying the predicted line speed was 2.5 Mb. I had, in passing, thought that a bit slow since, when we switched from BT to Plusnet, both sides were showing the line speed as being 3 Mb. We are light users, rarely going over 6 Gb a month, so don't normally notice downloads being slow. Anyway I logged into the HH2 to see what the line looked like and was somewhat surprised to see that the SNR was 15 dB down and 17 dB up. A speed test showed that typical download speed was around 2.75 Mb. So I started reading up on posts in the Forum and decided to give the line some time to see if it would adjust itself while also accumulating data.
The BT Wholesale Line Test gave a best effort of 3.11 Mbps Down and 0.37 Mbps Up from three runs (not intentional) with pings of 44.88 and 43.00 on the first two runs (which had 2.79 Mbps DOWN) and the last run of 3.11 Down had a ping of 66.38.
The IP Profile was given as 2.85 Mbps and the Plusnet Line speed shows as 2.8 Mb.
The router stats show the modulation as being 992.3 which I believe means the line is DSLmax on 20CN. The Kitz calculator suggests that  the line could go up to a 3500 kbps Profile and the error rates don't look too horrendous. Could someone have a look, please, and see what could be done. Stats attached below.
NB. edited to update Plusnet Line Speed which has gone up to match the BT one.

[#2]

Hi Mike,
A warm welcome to the forums and thank you for providing a comprehensive set of data as an introduction to your problem.
First may I correct one of your assertions - G.992.3 is ADSL2 which means that you are on a 21CN service which (at your line length) is only marginally beneficial and ADSL2+ is certainly not going to deliver benefits.  It does mean though that you should get a bit more synch than you were anticipating.
When discussing expected speeds, it is conventional (and less confusing) to talk about synch speed, as until that is sorted, the rest simply does not matter.  The profile will follow the synch speed (88.2% thereof) and the actual data speed test will get out of the profile figure the best that the end to end systems can deliver given any constraints from exchange capacity or line conditions.
For the current synch speed, the profiles are correct.
So most important question first - why is the SNRM so high?
Is there noise on the phone line?  Dial 17070 option 2 and check it regularly.  If there is noise report a noisy line fault to your phone provider, but say nothing about broadband issues, unless it is PlusNet.
Is the router plugged into the master socket?  If not please try it there and recheck the stats.  If it is in the master socket, what kind of socket is it?  Does it have a removal lower face plate?  If yes, please remove it and try the router straight into the exposed test socket.  Again, please obtain and report the stats.
It is interesting that the error counts which are not many are only on the DS side - someone else might have a view if that is relevant.
If there is no noise I'd suspect that during the swapping over of the routers the DLM got rather aggressive over repeated line drops and all that is required is a SNRM reset to get the operational value down to 6dB.
Having the US uncapped might slightly improve your usable data speed slightly.
HTH,
Kevin

[#3]

Hi there.
I can confirm your line currently has a target of 15db for your downstream speeds which will be restricting your speed.
I'm more then happy to reset that for you but I would also recommend going through all the steps that Townman's advised you on as the target will have increased for a reason.
If you have any further questions feel free to let us know.

[#4]

Quote from: Townman

Hi Mike, A warm welcome to the forums and thank you for providing a comprehensive set of data as an introduction to your problem. First may I correct one of your assertions - G.992.3 is ADSL2 which means that you are on a 21CN service which (at your line length) is only marginally beneficial and ADSL2+ is certainly not going to deliver benefits.  It does mean though that you should get a bit more synch than you were anticipating. For the current synch speed, the profiles are correct. So most important question first - why is the SNRM so high? Is there noise on the phone line?  Is the router plugged into the master socket?  It is interesting that the error counts which are not many are only on the DS side - someone else might have a view if that is relevant. HTH, Kevin

Hi Kevin, Thanks for your post. I cannot now recall what prompted me to assume that the line was not ADSL2 and have to put it down to having originally done all the searching to assemble the data the night before only to find when I submitted the post around three in the morning that I was no longer logged in and the post was gone ! Think it was that I couldn't reconcile the sync speed with Plusnet's table of sync speeds for ADSL2. The time of day may well have been a significant factor.
I can hear no audible noise when doing the silence test and have tried both cordless and corded handsets. It is a subjective test however and I did use my best ear :0).
Yes the router is plugged into the master socket and I have on previous occasions run direct into the test socket with negligible difference in results. There still remains the filter which is plugged into the master socket of course, but I am reluctant to do any disconnect at present to give the exchange hardware reason to lower the line speed yet more.
I noticed that the errors were decidedly one-sided and further reading after your post indicate that a typical scenario for upping the SNMR and lowering the line speed could be REIN. I am making an assumption here, which could well be wrong, that errors down indicate faults found when decoding in the router and could be due to interference local to the router. The problem with that is why wouldn't the interference appear in the outgoing packets when received at the exchange? The problem is that, of necessity, the 'phone base station and router have to live in an 8 ft x 7 ft room that has both desktop PC and monitor, printer, film scanner and various other electrical bits of kit as well as the master socket and only two single sockets to supply them all. Despite having switched to the HH2 to allow me to shift the printer back out I may have to revert to the TG582n to enable me to run the routerstats software to see if anything is REINing on my parade. For the present, I have tried to reduce the equipment plugged into the socket nearest to the master socket down to the router and 'phone base station. The PC and monitor are fed from a strip plugged into the other socket at the other end of the room - bearing in mind that 'other end' means only about six feet apart and the monitor is two feet away from the router.
Mike

[#5]

Quote from: Matthew

Hi there. I can confirm your line currently has a target of 15db for your downstream speeds which will be restricting your speed. I'm more then happy to reset that for you but I would also recommend going through all the steps that Townman's advised you on as the target will have increased for a reason.

Hi Matthew,
Yes, I would appreciate a reset thank you. See my reply to Kevin above.
Still learning how this editor works. Is there a guide for it at all?
Mike

[#6]

Hi Mike.
That's been done for you now.
You'll lose connection briefly but when it comes back up the speeds should increase
It'll be interesting to see how the line performs after this. Fingers crossed what you've done so far has eliminated any local interference.

[#7]

Quote from: Matthew

Hi Mike. That's been done for you now.

Thanks for that Matthew.
Mike

[#8]

Hi Mike,
Well you have clearly done your research so consequently we are off to a flying start.  REIN, RouterStats well you have got to the next level of investigation without me needing to prompt you!
Which version of RouterStats are you using - lite or full?  There is a good deal of experience around here of running RS and good knowledge of useful tricks.
Ideally this needs to be run 24x7, using a 10 second interval sampling with 720 points per graph.  This gives good resolution and two hours per graph.  Have you set the graph options to save "Scheduled Capture"?  This will save a history of graphs which can be useful for looking back and for sharing on here.
To attach files, open the "Additional Options and Attachments..." link immediately below the edit window.
If you have questions, just ask and someone around here will have an answer.
Help with using the forums - this is based on SMF - This "regular user" guide from their support site might be useful to you - http://wiki.simplemachines.org/smf/Category:As_a_regular_user
Kevin

[#9]

Hi Kevin,
I am still running on the HH2 at present. Routerstats can't work with it as it doesn't use a pop-up for the sign in; that's why I was contemplating reverting back to the TG582n. I haven't switched router yet.
It is early days. I still have the one-sided errors to sort out, but . . .  YES, one V.Happy bunny so far with the result of the reset. I will have to see if the error rate remains moderate or whether I will need to do more work on them.
I appreciate your input Kevin, it got me to focus on what was important in the data. Will see how things go for the next day or so. Thanks for the link re: the editor.
Mike

[#10]

That looks alot better then it was before.
If anything changes let us know and we'll investigate further.

[#11]

Matthew,
Does it strike you as odd that the BT Profile shown on the speed test is less than the transfer speed attained in the test?  Is there a residual issue lurking in the wood work here?
That IP profile (2.85Mbps) is the speed this line keeps "falling back" to.
Also do you have a view on the merits of uncapping the US on this line please?
Kevin

[#12]

@Matthew
Looks like there could be a profile mismatch in BT's systems possibly somewhere there - download far faster than IP Profile
@YetAnotherMike
Advice - when you go to do anything that involves loss of sync, changing modem/routers, filters etc - using the following method -
If you have a modem/router that doesn't have an interface for you to drop the PPP session, then pull the power plug and wait a minute or two before unplugging from the line - otherwise -
Log in to the Modem/Router, in the Internet box, click Disconnect to drop the PPP Internet session (this is not the sync), wait about a minute and then power down the Modem/Router. After about another minute you can unplug it from the line.
Do whatever is needed, but stay offline for at least 10 minutes. When you re-connect and power up, Login to the Modem/Router and click Connect to establish a new PPP session.
Don't do this more than 5 times in an hour, then leave it for the rest of the day. This method is to help stop the Exchange DLM from thinking your sync loss was a dropped connection.
One further point, especially applicable to very long lines, you will get very small differences in sync speed every time you sync, so just because you get a marginally better sync speed in one situation, it doesn't necessarily mean that situation is better than another situation.

[#13]

@Townman and Anotherone
The profile on the BT Speedtester hasn't increased yet. It can sometimes take up to 4 hours for the profile on there to update to match the sync rate but the reset has been successful as the OP's seeing higher throughput.
I've also placed the order to uncap the upload as well which will complete tomorrow.

[#14]

Quote from: Anotherone

@YetAnotherMike Advice - when you go to do anything that involves loss of sync, changing modem/routers, filters etc - using the following method - If you have a modem/router that doesn't have an interface for you to drop the PPP session, then pull the power plug and wait a minute or two before unplugging from the line - otherwise - Log in to the Modem/Router, in the Internet box, click Disconnect to drop the PPP Internet session (this is not the sync), wait about a minute and then power down the Modem/Router. After about another minute you can unplug it from the line. Do whatever is needed, but stay offline for at least 10 minutes. When you re-connect and power up, Login to the Modem/Router and click Connect to establish a new PPP session. Don't do this more than 5 times in an hour, then leave it for the rest of the day. This method is to help stop the Exchange DLM from thinking your sync loss was a dropped connection.

@Anotherone
I had already spotted your advice in other threads and have been following a hybrid version with the Home Hub which has a web interface to allow one to disconnect, but no power switch which forces one to 'pull the plug' to power down. That advice certainly helped prevent the DLM driving down the sync speed and further elevating the SNRM. My line comes in underground and surfaces near the front door to the ubiquitous grey box about 18"/50cm up the wall. The master socket is mounted on the inner side across the cavity. Am I right in thinking that, short of there being a problem with underground ducting, if REIN is the source of the packet errors then it is likely to be something inside the house?
@Matthew

Quote from: Matthew

I've also placed the order to uncap the upload as well which will complete tomorrow.

Thanks for that Matthew
Mike

[#15]

ADSL uses the higher frequencies for the Downstream, whereas the Upstream uses the lower frequencies. The Downstream is always more susceptible to any kind of RFI, the most common of which is AM/MW propagation effects between dusk and dawn.
I wouldn't assume that just because your line comes in underground that any REIN or other forms of background interference etc have to be inside the house.
It only needs other lines in the same cable or on the same DP (or yours) to be suffering problems, for cross-talk to have a noticeable effect. Some errors are inevitable. Based on the stats posted above, for that short uptime, the errors aren't horrendous, but it would be nice if they were lower. Obviously your own equipment may generate RFI that causes a problem.
The best way to check out your own equipment is to get RouterStats running and monitor the SNRM, you can also look at the stats as well, to look at the error count.
Then go round and switch off (not just standby) all electronic and electrical equipment that could generate interference - mobile & cordless phones, anything with a modern switch mode power supply is a potential candidate, TV sets, set-top boxes, sky boxes, all energy saving lighting and fluorescent fittings, also discharge lighting (sodium, mercury vapour, metal halide etc - a street light nearby could also cause a problem if faulty), anything with a motor - washing machines, dishwashers, vacuum cleaners, - central heating pumps, fridges, freezers etc. Whilst the last 3 when running normally should not generate interference whilst running, switching on/off if thermostats are arcing can cause problems. Switch on (or sometimes off) spikes from the other devices can also be an issue, more so motorised devices. The list isn't exhaustive, but gives you a good idea.
Likewise stuff that may not be on, switch it on. The other thing to switch off is your monitor. Note the SNRM and error counts and the time, and then switch it off for say 10 minutes. When you switch back on, see if there was any step changes in the SNRM on RouterStats and any effect on error counts.
If you notice any significant shifts in SNRM or increases in CRC/ES errors counts, the track down the individual bit of equipment responsible. HTH.