Kitz Forum
Broadband Related => ADSL Issues => Topic started by: karmie on November 01, 2014, 01:51:50 AM
-
Hi guys - I am having some pretty 'annoying' problems with my broadband and would love to get some advice on what you think my best 'play' is. I was previously on Be* Annex-M and had 21mb down / 2mb up and ~12ms latency to BBC (for the whole 4 years I had them). Obviously Sky bought BE* and swapped us over, my speeds have been pretty terrible, 13mb down / 1mb up and ~45ms to BBC. I got Sky to turn down the interleave depth to the minimum and had pings of ~15ms or so for a while until the DLM kicked in and put it back to 45ms.
Here are my line stats from the Thomson TG585v7: http://pastebin.com/LPxY90JP
A dump from Routerstats summary page: http://pastebin.com/RdAQR7De
Excuse my lack of knowledge but from what I have gathered at 5pm every day my downstream FEC errors go crazy and as a result I think Sky sets my SN Margin and interleave depth ridiculously high to compensate to allow the connection to stay alive? As you can see with 1week uptime the connection is stable but the latency is just poor.. never have below 50ms within the UK which isn't so much of a problem but when you start hitting Sweden/Germany etc we are talking ~80ms solid which is a huge problem for me.
I have isolated the problem to what I BELIEVE is my internal wiring, when plugged in to the master socket the sync goes up and the line doesn't error. Obviously new filters, new RJ11 cable and tried multiple routers and at every extension point in my house we get huge erroring (and sometimes increased attenuation) - in the master socket the downstream attenuation is 15db and I didn't spot hardly any errors and I managed to sync much higher.
Ok so heres where I need the advice - if it is my internal wiring (which I believe it is) - do I need to get a phone engineer to come and rewire every socket in my home? I believe the cabling is all running through walls so are we talking huge amounts of money for this? Hypothetically I could use a long RJ11 extension from the master socket (this is at my front door) to the centre of the house but its not really a permanent solution.
I am just really confused - I was with Be* for 4 years and had a downstream attenuation of 9.5db and never had any problems in that whole period. I move to Sky and I gain 7db attenuation and apparently at the exact same time my internal wiring has all failed on me. I would gladly take 5mb down if I could have <20ms latency as before. Fibre is unfortunately not available here and my only option is to buy out my contract and move to another LLU provider but whether or not that will do any good remains to be seen.
I am so sorry for rambling - any advice is greatly appreciated. I am not worried about stability, downstream speeds (within reason) and purely looking to get the latency I used to have.
-
Rather than rewiring I suggest you first consider a filtered faceplate say £7, plus 60p for insert tool This includes a single "micro filter" for the whole system and is usually wired so that data is only available at the master socket which now has phone and modem sockets .
This will see off any interference picked up on the extensions and also "there and back" delays to data l It also means no dangly filters.
This will put the system in better shape, not sure about latency I thought this was down to the ISP's system but no doubt someone will correct me !!
Best of Luck
-
Welcome to the forum, Karnie. :)
Further to JGO's great advice, I would simply ring your ISP and explain you have slow speeds at your extension socket, but it's OK in the test port at the Master Socket.
I've literally worked on thousands of this very same scenario. The ISP should build an SFI Broadband engineering task for Openreach to visit and 'normalise' your internal wiring set-up.
Most, if not all ISP's tend to foot the bill for this kind of visit, by way of a gesture of goodwill in the hope you'll remain with them when out of contract ?.
-
Hi guys thanks for your valued input - I was put off with phoning Sky again as the agent had simply told me to phone BT and report a voice fault on the line yet when I phoned BT they did a line test and told me nothing was wrong with the line (at least the voice side of it) and told me I could have an engineer but if it worked when he got here it would cost me £150. I didn't really want to spend £150 when I knew the phone was working fine.. so I just assumed I would have to hire a telephone engineer myself and get them to rewire my entire house.
The filtered faceplate or SFI engineer sounds good - I am not averse to spending money to fix the problem but if we are talking upwards of >£200 I would want to be sure it would fix the problem.. Genuinely encourages me to move property when we are nearly in 2015 and I upload at 1mb a second and have >50ms to BBC. What is ironic is back in 2001~ on Blueyonder prior to congestion I had <5ms pings to the Blueyonder CS Servers (if anyone played Counter-Strike back in the day may remember =d) Obviously the older we get the more latency advantage we need on these young guns.. ;-)
-
One thing which is not obvious from your initial post is whether your "master socket" is an NTE5/A (first image below) or an old LJU2/1A (second image below).
If it is the latter then your ISP needs to book an Openreach engineering visit to "normalise" the wiring and to fit an NTE5/A. That service is provided FOC by Openreach, as it is in everyone's interest -- Openreach, your ISP and yourself. :)
-
Good point B'kat.
Am I correct in thinking the older version doesn't have a Test Socket inside ?
-
Its interesting because my 'master socket' at my front door has the openreach logo at the top like the NTE5 faceplate (the bit at the top anyway) and the bottom half of the panel I can remove has absolutely no input on it its just a blank white plastic panel and behind it I have the only test socket in the house.
Is this normal?
-
Karmie, The faceplate is full height with a kink so it fits over the top half-plate after the bottom half has been removed, plugging into the Test socket. This gives you a Filtered Master socket with Modem and Phone sockets on the front.
The point is that then the modem can't be plugged in elsewhere, well not without loosing most of the faceplates advantages. If you MUST have the modem elsewhere then think you need BT to extend the wiring and move the test socket - NOT a DIY job..
-
Good point B'kat.
Am I correct in thinking the older version doesn't have a Test Socket inside ?
That's right. Only the NTE5 type of master socket has a test socket inside. With the older type of socket, there was no demarcation point where BT's wiring responsibility ended - all the telephone wiring was BT's responsibility.
-
Going back to Karmie's first post, I was on the BE/O2 network until Sky took over the broadband. Over the last week my sync rate has dropped from about 5Mbs down to 4Mbs, with a noise margin of 6.9dB. I did some tests with a Sky chap last night, and have a BT engineer due to call tomorrow. I don't think the BT chap will get anywhere as I think my broadband tranmit level has been reduced at the exchange, or my line has suddenly been hit with increased cross talk. I suspect I have been fortunate that I have had the line speed I have had for years, as most my neighbours broadband lines have had slower speeds than mine.........
-
Thanks Roseway.
-
Its interesting because my 'master socket' at my front door has the openreach logo at the top like the NTE5 faceplate (the bit at the top anyway) and the bottom half of the panel I can remove has absolutely no input on it its just a blank white plastic panel and behind it I have the only test socket in the house.
Is this normal?
So it sounds like you do have an NTE5 master socket, along with other extensions in the property.
But from what you say, these extensions are not wired from the lower removable section of the NTE5. If so then this is wrong and would be worth having corrected.
-
Yup, it should all be covered under the 'normalisation' process. You have what is technically termed as a 'Bridged Tap', more commonly termed ..... 'Star Wiring'.
As I've said before, ring your ISP back and request an SFI BB Engineering visit. Tell them not to be so damned stupid if they ask you to ring BT to raise a 'Voice' fault ??
-
Its interesting because my 'master socket' at my front door has the openreach logo at the top like the NTE5 faceplate (the bit at the top anyway) and the bottom half of the panel I can remove has absolutely no input on it its just a blank white plastic panel and behind it I have the only test socket in the house.
Is this normal?
In my absence, I see that you have been given the best possible advice. :)
However in answer to your question, above, I would probably have to say "Hmm . . . uncertain." :-\ Perhaps I am "seeing" something different to what you have described but I am wondering if your demarcation point is actually an NTE5/B. If you have the means to take a photograph (with the lower front plate in place and then with it removed), that would be the simplest way for us to understand exactly what is fitted.
-
Once again I want to thank you all so much for your valued input! First thing tomorrow I will phone and request an SFI BB Engineering visit and regardless of the cost hopefully get the problem sorted.
Interesting to hear about the other Be->Sky forced migrant - is it possible the signal is weaker or something? When I got swapped over from Be* - Sky my attenuation went from 9.5db downstream to 15db in Test socket and 16.5db at my 'extension'. It coincided with about a 2 hour downtime when a BT Openreach van was at my cabinet (which is a stones throw from my house.)
EDIT:: Included photos as per Burakkucat request! One thing to note is the dial tone on all my extensions dropped and obviously the Internet when the plastic panel is removed it unplugs the thing from the test socket and it all goes down :(
Front: http://i.imgur.com/MyvIORp.jpg
Inside: http://i.imgur.com/MQVarAe.jpg
Sorry for the blurry picture, have a terrible old iphone camera :p Hopefully someone can make out what that is!
It is worth mentioning that we have 4 sockets in the home, this one is in the fuse box cupboard at the front door, one in kitchen, living room and bedroom. None of the other sockets have any test sockets inside of them.
-
Thank you for those photographs. Yes you do, indeed, have an NTE5/B. And that is the demarcation point between Openreach's wiring and your wiring.
The removable, lower, front plate and all the wiring connected to its rear via the IDCs are actually your responsibility. Unfortunately, as you do have a valid demarcation point, any work that Openreach perform will be chargeable. :(
In view of the above, it might be worthwhile waiting to see what Black Sheep has to say before you contact your ISP.
-
Yup, B*Cat has gotten to the bottom of the issue. I thought from your previous post you had star-wiring before the faceplate, but now you have posted the photo's, it shows this not to be the case. NTE5B's are quite uncommon in residential premises.
TBH, you still have star-wiring on the front-plate (ie- 2 cables attached), and the bell-wire is also connected on both these circuits. You could either chance your arm with your ISP and say you have greater speeds in the test socket than at the extension socket, and hope they pick up the engineering tab ?
Or, completely disconnect the extension wire that doesn't feed your Hub/Router on the front-plate, then pull off the wire connected to terminal 3 of the wire that does feed your router at the same front-plate, and see how much speed increase you obtain ??
Hope this kinda makes sense, got to be a quick reply as working.
-
From what I can see of the IDCs on the back of the lower front face-plate of the NTE5/B, it looks as if there are wires to all six of the IDCs, twice over.
If willing to do so, an appropriate experiment would be to remove both wires (originating from each of the two cables) from IDCs 1, 3, 4 & 6, just leaving the two at IDCs 2 & 5 connected. Then go to every extension socket and carefully remove all the wires that may be connected to IDCs 1, 3, 4 & 6, just leaving wires to IDCs 2 & 5 (as before).
(Of course, if you are located in the county of Surrey then it should be possible to arrange for our Walter's Wheelbarrow to trundle around to your address and to perform a wiring optimisation. ;) )
-
Hi guys - sorry I have been MIA from this thread but just want to update as to what went on.
Once again I am backed into a corner and the main issue is that I have a BT landline and Sky ADSL (Fibre is unfortunately unavailable in my area.) I had a local phone engineer come out to fix our internal wiring which I thought was at fault and he said there was nothing wrong with our internal wiring from his point of view and left (without charge, I might add.)
I spoke with Sky who have said: Your line is getting 13mb down and 1mb up which is more than your line should be capable of and that they are within their outlined 'speed estimation' and as such can't help and told me to contact BT. I contact BT who says because I can make and recieve landline calls that nothing is wrong with the line. Where on earth do I turn to now?
I started running ROUTERSTATS on my line and EVERY day only between the hours of 5pm - 11pm the line is dropping sync (once an hour on average), frequent CRC/HEC Downstream errors in these periods only (100 crc /minute and 20 hec/minute downstream) and the downstream margin jumps from 7db to between 9 and 13db between these periods and the sync drops from 16500 to ~13000. I unfortunately have lost my routerstats for the day as I rebooted and it doesn't log automatically but I turned on logging now and will monitor. I will post routerstats from now (the bad period) and then later after 11 to show the difference.
I am considering just purchasing a second line to be installed and new broadband install - would that solve the problem? I don't see where Sky get off telling me the line isn't capable of 13mb down / 1mb up when I had 22mb down and 2.5mb up with Be* Annex M for 4.5 years with router uptimes of 6months on 3db profile and fastpath. ???
Here is my current router stats (during bad period): http://pastebin.com/GiV4Lffq
What would you guys do? Who do I turn to? :no: I just want stability on a fastpath line.
-
Hi karmie
I note you say you have a TG585v7. Can you download DSLstats (http://www.s446074245.websitehome.co.uk/downloads.html) please and set that running. You can get a bit more info from DSLstats and I prefer it to routerstats.
I don't see where Sky get off telling me the line isn't capable of 13mb down / 1mb up when I had 22mb down and 2.5mb up with Be* Annex M for 4.5 years with router uptimes of 6months on 3db profile and fastpath
From your attenuation you should be getting a lot more speed than you are doing, but I would like to see a DSLstats/Routerstats graph just to see how much your SNRm is varying.. and to see if there any pattern.
I contact BT who says because I can make and recieve landline calls that nothing is wrong with the line.
I think I'd leave that avenue for now, because its likely you will get charged. The fault appears to be with your DSL, and therefore its your ISPs responsibility.
I don't see where Sky get off telling me the line isn't capable of 13mb down / 1mb up when I had 22mb down and 2.5mb up with Be* Annex M for 4.5 years
It is possible that you have developed a line fault, but it does seem rather odd that it occured at the time of migration over to SKY.
Im not a fan of SKYs DLM, it works backwards.. in otherwords you have to prove that the line is stable at 'a' before it will move you to 'b'. Other DLM's will try you at say e and if the line is unstable move you down to 'd'.
The reason Im not a fan is I put my parents on Sky thinking it was cheap for them as they had Sky TV. They never ever got more than 2.6 Mbps on that line if they were lucky and the latency was appaulling - something like 70+ms to my server. After several months SKy did nothing to help and fobbed my dad off numerous times.. and in the end I moved them to Plusnet in March this year. The line immediately synced in the 8Mbps region with latency of 16ms. The line is/was a tad unstable and I expected the BT DLM to cut in quite soon, but in actual fact it held at 8Mbps for a few months before the BT DLM applied Interleaving. Ive not checked their stats for a while, but last time I looked about 6wks or so ago and it was syncing at around 7Mbps and with a latency of 22ms. As far as my parents are concerned it now works and they are more than happy because at least now they can stream TV and do their weekly Skype video chats to my relatives in Oz without jitter, stutter and lag that they got when on Sky. I moved 3 people away from Sky in March/April - all 3 are much happier.
My daughter will spit at the name Sky and still blames her b/f for being conned into going over to Sky by signing up with one of those guys who had a stand in the centre of the shopping mall. She'd been used to an Annex_M 24/2.6 Mbps BE* so going to Sky was a bit a of a culture shock especially after the way they dealt with a genuine line fault at the beginning of the year. :(
and my only option is to buy out my contract and move to another LLU provider but whether or not that will do any good remains to be seen.
Dont let them lead you to believe that. You moved over from BE, you are NOT subject to a new contract. This is something Ive seen them try with several BE and O2 users. Be firm about it. Unless you took out a new sky package and changed what you got from BE then there is no new contract. They tried the same with my daughter, I think I mentioned it on here and I threatened them with OFCOM when they started stalling and messing my daughter about after she requested a MAC key. It took a several week but we got there in the end without penalty. A friend of mine had similar issues with O2 -> PN. He too lost speed (and static IP) gaining only additional latency after his line go moved over to Sky. Once he firmly said he wasnt in contract they relented, hes also much happier now with PN (No hes not one of the 3 I moved).
Anyhow me moaning about Sky isnt going to get you anywhere. The first step is seeing some line monitoring to see if we can see any pattern on DSLstats.
-
Hi Kitz
Thanks very much for your detailed reply. I am aware Sky get a lot of heat from everyone and I just assumed it was due to their reach/customer base being so large and aimed at casual use that the service was always going to be poorer (ie much stricter DLM to present least possible problems and tech support load.)
I have installed DSLstats and will let it run for 24 hours from now. If I reboot the router now (after midnight) I will sync at 16,000+ but I will just leave it as it is. Sky have turned off DLM on my line at my request to allow me to run fastpath (against their advice as the line is instable between 6pm-11pm but I only had 1 disconnect this evening.)
Will post back in 24h with results to see if anyone can suggest anything and mainly who I should turn to SKY or BT. I am assuming it is Sky based on what you have said Kitz and I don't doubt your judgement. Is it possible when swapped from Sky -> Be I got put in a different 'line card' or something at the exchange that is broken? Is crosstalk possible on ADSL? Only between 6pm-11pm does the line show ANY problems - it is worth mentioning my line attenuation went from 9db downstream to 16db upon moving to Sky? Does that mean anything?
Thanks again!
-
Is it possible when swapped from Sky -> Be I got put in a different 'line card'
Yes, you will have been. Customers were moved from the BE* MSAN on to the Sky MSAN. You can tell that you are now on a Sky SVBN from this:
Vendor ID (Local/Remote): TMMB / GSPNSky use Globespan. The BE* MSAN's were Alcatel which would show up as blank or ----
Is crosstalk possible on ADSL?
Yes. I used to see it when I was on adsl2+ with Be. It is harder to spot though and the losses arent anywhere near like they are with VDSL.
I was one of the first on the new Be MSAN, first day it went live. Several years later I'd lost about 3 Mbps to crosstalk. It was a very gradual thing though as more users come onboard. However, that said, remember its the very short lines that suffer with crosstalk the most.
I feel the loss in attenuation may be more telling. Your attenuation should not have changed by that much. At 9dB, your attenuation is slightly less than mine was and although I started on the full 24/2.6, by the time crosstalk finished with me, I was down to about 21Mbps... so your Be stats seem about right and you will likely had a good dose of crosstalk already, so I would not attribute the further loss down to 13Mbps as crosstalk.
The fact that Sky doesnt do Annex_M means that your downstream speed should have been more than when on Be, not than less, so I think you have every right to be mighty fed up at only getting 13Mbps :(
Only between 6pm-11pm does the line show ANY problems
This could be REIN (http://www.kitz.co.uk/adsl/rein.htm), but I think we need to see more info from DSLstats before jumping to conclusions... and it shouldnt really affect your attenuation either. :-\
-
Ummm Ive just had a look at your first set of stats and this jumped straight out at me
Output Power (Up/Down) [dBm]: 10.5 / 1.5
Line Attenuation (Up/Down) [dB]: 8.0 / 16.5
Downstream Ouput power on adsl2+ should be circa 18-20 dBm for a line to achieve its full potential. If that figure is correct, then you are not getting full power and its no wonder your attenuation has dropped.
Are you still using the same router as when you were on Be? Do you have some old Be stats to check what output power was? I want to eliminate that its not the router misreporting.
-
Try turning the power off to your Modem/Router over night and re-connect it in the morning.
-
Hi again guys
@Kitz - I believe the Thomson TG585v7 is known to misreport output power - i tend to see the figure at 0 or 25.5db for the downstream output power.
@Tickmike - What would turning the router off overnight do if you don't mind me asking? It would be a pain to do as the internet is almost constantly in use by someone in my household but if it has a chance at helping it will most certainly be done!
Have had DSLstats running for over 24h now and the stats basically reflect what Routerstats shows - interestingly 7:20PM 3 nights in a row was when the drops began and the line becomes stable at 11pm. I am going to try to figure out how to convey all the DSLstats information in readable format.. please note the router was power cycled at like 4am at the start of the graph when the sync jumps to 16200kbps and hasn't been touched since.
Link to DSLstats album: http://imgur.com/a/MXSvr#0
Other hopefully relevant info: http://pastebin.com/prrkeeBB
It is the 3rd night in a row that the connection goes haywire at 7:20pm and I can't for the life of me figure out why - nothing is turned on at our end no phone calls made/received at that time. Can any of you clever folk make anything of this?
Thank you again so much for all the help - even if I know who I have to pay to fix it that would be a start because right now Sky doesn't want to know and neither do BT. Its this rotation =
SKY: Does your broadband work and are you getting the speeds we quote? "yes" - Sky not interested.
BT: Does your telephone line make and receive telephone calls? "yes" - BT not interested.
:'(
EDIT: On reading back my post I wanted to make sure you guys know what we have 'tried' to eliminate the problem:
1) Tried with the SKY SR102 and Thomson TG585v7 with equal results, Thomson chipset seems to hold sync better during bad period.
2) Replaced RJ11 cable and filter between router and phone socket. The only equipment on the line is a singular phone plugged into said new filter. The exact same issues occur with the telephone unplugged.
3) We are using an extension socket which was 'tested' and deemed to be perfectly fine - unfortunately with the 'master' socket at our front door (which has no regular socket, just a test behind a plastic faceplate) if we use the 'Test' socket it disconnects every other socket in the house as the faceplate that plugs into the socket is what the extensions are run from(?) This is an issue because we need a phone in the office and it would be an absolute nightmare to run 50metres of RJ11 extension up the stairs to then have ONE landline phone in the household.
-
Going haywire at 7.20 PM could be someone near by switching on say, a new plasma TV for the evening. I get something similar but it is only 0.5 dB or so.
First line of defence - do you have a filtered faceplate correctly connected ? These plug into the test socket and have both a modem- and a phone socket on the front. All the extensions are wired to the faceplate and are phone only.
With the simplification of wiring suggested by Bkat/BS this should see off the extensions acting as an aerial picking up interference and unbalance due to the bell wire making it worse. It also eliminates the need for individual microfilters.
IF you don't have one maybe Bkat can advise - would BT modernise the connection free ?
-
with the 'master' socket at our front door (which has no regular socket, just a test behind a plastic faceplate) if we use the 'Test' socket it disconnects every other socket in the house as the faceplate that plugs into the socket is what the extensions are run from(?)
A photo would help explain this, but wouldn't the new iPlate solve the problem as it can be "slipped in" without disconnecting anything and provides a nice clean ADSL socket filtered from all the extension wiring.
Here's how it's installed. Note this sequence uses the now-defunct Mk1, but the principle is the same with the Mk2.
(https://forum.kitz.co.uk/proxy.php?request=http%3A%2F%2Fwww.ispreview.co.uk%2Farticles%2Fadsltips%2Fbtiplate_install.jpg&hash=29e30f3ac68b30b28ceaffd9dddd6ba86df4ea05)
-
NO ! not the I plate, a filtered faceplate eg :=
http://www.run-it-direct.co.uk/btvdslfaceplate.html
For ADSL you don't have to worry which version, only applies to FFTC.
-
Hi guys
I included photos of the socket in an earlier post on Page 1 but it was a while ago so here they are again:
Front: http://i.imgur.com/MyvIORp.jpg
Inside: http://i.imgur.com/MQVarAe.jpg
With FTTC anticipated here within the next 3-6 months (according to the checker my cabinet is ready but not accepting orders yet) if I got this plate fitted would it cause problems with that in future?
Regarding the bell wire it is certainly something I could do as I wouldn't need a crimping tool, is this correct? I am just cutting 2 wires from one of the IDC's at the master? Would BT scold me for this? Is the face plate fitted at my extension (in my living room) or at the master at the front door? If I fit it to the master at the front door does that mean I then have to plug my router in to the DSL socket on it at my front door and run a long RJ11 cable to the center of my house (really don't want to do that.)
Sorry for all the questions as this advanced stuff is extremely new to me. Thanks again!
-
I included photos of the socket in an earlier post on Page 1
Sorry, missed those. :-[
So an iPlate looks ideal......... (JGO do explain why not)
With the router plugged in locally to that you can use wireless or LAN cable to extend elsewhere.
Or to relocate the router use the unfiltered terminators on the iPlate and run a twisted-pair cable (ordinary 4-core CW1308 phone cable is fine) to a new socket solely for the router; that could be an RJ11 outlet to prevent use by BT-plug phones.
-
I believe the Thomson TG585v7 is known to misreport output power - i tend to see the figure at 0 or 25.5db for the downstream output power.
In a way thats a shame, as it could have explained the drop in attenuation. Output power is a measurement of power transmitted from the MSAN, which affects the strength of the adsl signal and can also make the line more vulnerable to fluctuations in SNR if its too low.
Leaving the router off overnight wont make any difference to the figure, but I have known some routers mis-report, which is why I asked if you can recall if it ever seemed to record properly when on BE.
For adsl2+ it should be 18/19/20dBm, and the DSLAM can ramp the power up a bit if theres lots of bit-swap. Anything less and you are unlikely to get full bitloading, its seldom you ever see it at more than 20dBm.
If you see 0, then yes its false reporting, because the line just would not work if that figure was true :(
Have had DSLstats running for over 24h now
Thank you. At first glance of the graph it is showing symptoms of REIN.
Closer inspection of the SNRm shows some odd upstream spikes at circa 18:05, 19:20 and 20:05 which we sometimes see if there is a line fault and the telephone rings. However, I note you say that the phone wasnt in use, so taken with the downstream variance, it could possibly be SHINE (http://www.kitz.co.uk/adsl/rein.htm#SHINE).
Just for the sake of completeness, can you advise if the line lost sync at those times? The 'Connection Speed' tab may show this as its unlikely to have retrained at exactly the same sync speed. Whilst at the Connection speed tab, it may be interesting to tick the 2 check boxes to also record the max attainable values.
REIN and SHINE (http://www.kitz.co.uk/adsl/rein.htm) do sometimes go together. From the graphs its possible that your upstream is affected by SHINE when 'something' is first switched on and shows a spike, whilst the downstream is affected for the period of time the cause remains switched on.
Your CRCs during this period are going crazy. From the graphs, it looks like the figures are clipped at 30, but in reality are much more (You can change the clipping value and/or turn it off, but the graph shows the effects either way).
Your bit loading graph, shows that you possibly do have a small amount of crosstalk. If your SNR per tone follows the same pattern, then its centered around tone 225 which is the 7/8Mb speeds.. and is exactly where Id expect a short line to see the most effects, so all looks normalish as far as crosstalk is concerned.
Whether its the state of your internal wiring that is causing the connection to be more sensitive to REIN I dont know. The bell wire is the usual culprit for acting as an antennae for REIN.
There are many causes of REIN, it doesn't necessarily have to be in your own house, weve seen many cases where its neighbours equipment that has been the culprit.
if I got this plate fitted would it cause problems with that in future?
I'll leave the other guys to discuss the internal wiring & face plates as they are already on it... but personally speaking if you can get it normalised to an SSFP, then no it wont affect VDSL. In fact BToR Engineers will normally fit an SSFP Mk3 as part of a managed FTTC install. I had previously installed my own SSFP several years before getting FTTC. No BT wont scold you for removing the bell/ring wire.
Im not too sure if its a good idea to get FFTC yet though whilst the line is in its current state. (noted that its ready but not available). I had a fault develop literally the day after I ordered FTTC. (FTTC and the Sky takeover of the BE MSANs happened at the same time here). Its going to be much harder when on FTTC to prove that your line isnt performing as it was.
I dont know the exact location of your cab, but based on the similarity of your BE adsl2+ stats to my own, then Id anticipate you should easily get 80Mpbs plus on vdsl. What was a very obvious fault on adsl2+, became much harder to prove on vdsl, because you should get lots of 'spare' SNRm for such a short line. This can mask SNRm variances and errors, that wont really affect the line until VDSL crosstalk also starts to kick in. By which time it will just all be lumped together and you will be told nothing they can do for crosstalk.
-
Karmi,
1) MK1 faceplate was designed for ADSL only and the RF3 filter had inadequate bandwidth (in Hz) for full bandwidth VDSL. Depending on your available FFTC speed this may not matter - see Les-70's posts.
2) Cutting bell wires NO - pull out one of the connector wires and swear it fell out ! The bell wire is only used for pulse dialing so obsolescent.
3) AIUI you need a xDSL extension from faceplate to modem, or locate modem adjacent to faceplate and long Ethernet cable ? The second is preferable but it may be worth trying the first, cheaper and more flexible cable.
Do bear in mind that this assumes interference is your problem - although sooner or later it probably will be
HPsauce
I plate (I for Interstitial ) is sandwiched between existing faceplate and box. It has no NO "microfilter".
FFPs have a repacement faceplate with 2 sockets phone and modem and microfilter.
Both have bell wire choke and an RF3 type device.
-
I plate (I for Interstitial ) is sandwiched between existing faceplate and box. It has no NO "microfilter".
I'm afraid you are 100% wrong there, or are thinking of some other (or older) product. Please go and check and make sure you look at the Mk2 version with an RJ11 socket at the top. ;)
(and yes I do have one myself)wouldn't the new iPlate solve the problem
I've just realised the installation image I linked to used the older version without the socket. :blush: (original post now edited)
Here's a closeup of the iPlate also showing where to connect wires for an unfiltered ADSL extension.:
(https://forum.kitz.co.uk/proxy.php?request=http%3A%2F%2Fwww.run-it-direct.co.uk%2Fimages%2Fmk2vdslidcblockbranded.jpg&hash=76bbfe31087d09a79e491230c316f4b8216b8d77)
And here's the internals, showing the filters etc. :
(https://forum.kitz.co.uk/proxy.php?request=http%3A%2F%2Fwww.run-it-direct.co.uk%2Fimages%2Fmk2vdslpcb.jpg&hash=f5a3c3fc4dc054d69282fe8ef1d861c7b0284671)
-
@Kitz thanks for the very detailed analysis - I can confirm the line DOES drop and resync immediately, usually within 10-20 seconds and sometimes will drop a number of times over a matter of minutes and then remain stable for a few hours. Between midnight and noon the line has not dropped once nor have I seen many CRC errors (which at peak times can reach around 120 per 30second sample)
I am a little confused regarding iPlate or the NTE5 Faceplate? I can order one ASAP and read up on how to fit it (provided it is compatible with my master socket, I linked images in last post?)
For now I will remove the bell wire as PERHAPS it would help ease the issue if this is conducting interference around the line as it clearly seems to be an interference issue whether it is in my property or elsewhere. Am I correct in saying it is simply removing the Orange and White wire at the master? Should I do this for all 3 of my extensions too or initially just at the master and take it from there? This will be the first thing I do when I get in this evening.
Out of curiousity, if I got BT to install a new line for me with a master socket in a reasonable location (like my living room) and ordered it with BT broadband (or Plusnet) would you think that could be a way around this issue? Its absolutely bizarre how flawless my connection has been in this property, since 2007 up until 2014 the connection was rock solid on fastpath with sub 10db attenuation and always maximum syncs and throughput. I was amazed when we moved in to the property at just how good the line stats were and the moment I moved over to Sky suddenly my line is a disgrace. I just cannot see how my line managed to fail + my attenuation increased to around 16db within the same week I got forced to move over to Sky.
-
For now I will remove the bell wire as PERHAPS it would help ease the issue if this is conducting interference around the line as it clearly seems to be an interference issue whether it is in my property or elsewhere. Am I correct in saying it is simply removing the Orange and White wire at the master? Should I do this for all 3 of my extensions too or initially just at the master and take it from there?
The colours mean nothing, it's what connector they are in.
Connections 2 and 5 are used for the phone signal, all others should be unused in a standard domestic situation.
I'd remove them everywhere in case there is equipment that can feed interference back onto the main wires, but just at the master is a reasonable start.
Note that the bell wire has a role in phones ringing (hence it's name) and some (very few these days) handsets will require it in order to ring. The good news is that you can avoid using it simply by having the phone plugged into a standard plug-in ADSL filter block as they (usually) recreate it locally.
In most situations you will have that anyway so need to do nothing, but installations with a central filtered faceplate can require this "fix".
-
So on my master and extensions if I remove the bell wire (I will check out some Google info) and provided I don't touch anything at IDC's 2 & 5 it should be all functioning? I only use 1 landline phone and it is a newer electronic wireless handset type from BT bought last year so it definitely does not need the bell wire. It is a great point to mention that the filters can recreate it locally if the need was ever there.
Regarding the Faceplate - by looking at the image of my master would that be considered a standard socket to install it into or would I have to install a seperate master box next to it and go down the whole "crimping" route (not sure if that is correct terminology) - I am all for snipping a couple of wires but I would sooner pay for the more complex jobs to be done. We had a local bloke out who said our internal wiring was all fine and I am assuming if he could have done anything to help he would have done so (as it would have made him some money as he didn't even ask a callout fee..)
-
HPSauce.
Yes, what I was describing was the first ( and as far as I knew only) I plate. (IMHO it was discredited by being advertised as "Magic Accelerator".)
I think the important point is that such a device doesn't remove pickup on the extensions or stub filtering by them and so isn't suitable for karmie ?
-
Well not exactly, the iPlate Mk1 does filter the bell wire to extensions so you are still misunderstanding.
The Mk2 is a significantly different device, they really are not comparable.
The master NTE5A here has, unusually, no "user" phone socket, just a blank lower faceplate. Installing a Mk2 iPlate would supply an unfiltered RJ11 socket for broadband (could be used for phone too with appropriate adapters/filtering) and filter all the extensions.
-
I think I get what both of you are saying and are both right in certain respects. :)
The MK2 version is vastly superior to the i-plate for filtering for filtering the lower telephone jack. It should also isolate the bell wire.
This (http://www.run-it-direct.co.uk/btvdslfaceplate.html) mean putting the blank face plate back on to the front bottom. Anything that would have come out of the bottom section (if it wasnt blanked) would be low pass filtered & anything coming out of the top socket would be unfiltered for adsl purposes (as would anything via the IDC connectors), so both voice and dsl would still be carried onwards to Karmies extension... which would still need a separate filter.
I isolated the bell wire when fitting an adsl filtered faceplate (http://www.kitz.co.uk/adsl/socket.htm)
-
Just putting my two penneth' in regarding terminology.
The i-Plate was the forerunner and was indeed touted as an 'Accelerator'. The subsequent MK1-3 SSFP's are not referred to as i-Plates between engineering staff. :)
-
Indeed, the common terminology doesn't help; as I said:
The Mk2 is a significantly different device, they really are not comparable.
So, BS, would you like to suggest suitable terms for us mere mortals to use when referring to the later "SSFP"'s?
-
Well not exactly, the iPlate Mk1 does filter the bell wire to extensions so you are still misunderstanding.
NO - choking the bell wire doesn't stop common mode interference pickup on the other two or them acting as a stub filter.
The subsequent MK1-3 SSFP's are not referred to as i-Plates between engineering staff. :)
Thanks for that BS.
I take it they are by marketing people ? :)
The important issue is advising karmie. Something from 7.20 to 11 pm each evening sounds domestic - probably TV (if you are a masochist !) or is there some plausible line mechanism ?
As he is going to FFTC it needs the later SSFPs but an estimate of speed seems to be needed first to decide which device ?
---
Edited by admin fixed smiley which was bugging me
-
I can if you would like me to ............ how about something along the lines of MK1, MK2 and MK3. ;) :)
The difference between the components I couldn't comment on, but identification is very easy. The original i-Plate would connect as a faceplate, whereas the subsequent SSFP's are actually sandwiched between the back-plate and the front-plate.
The MK1 SSFP has no markings on it other than the embossed 'Openreach logo', the MK2 and MK3 have this version number printed in the top right-hand corner of the SSFP.
HTH.
-
The subsequent MK1-3 SSFP's are not referred to as i-Plates between engineering staff. :)
Thanks for that BS.
I take it they are by marketing people ? :)
I wouldn't know, JGO .......... I stick to what Openreach engineering and our Logistics team, refer to them as. ;) ;D
---
Edited by admin - fixed smiley
-
choking the bell wire doesn't stop common mode interference pickup on the other two or them acting as a stub filter.
So for now just pull the bell wires. Cheap and simple no special tools needed. Could replace the blanking plate with a filtered faceplate but I dont see the point if theres nothing coming from it.
If /when karmie gets FTTC hopefully they still will do managed install so BT could normalise the box and perhaps even do a locate for free (if its a BToR guy and you offer a cuppa and biccies)
-
May I now make a few comments, please?
- Earlier in this thread I asked if some photographs would be possible. They were and they showed that there is an NTE5/B present. An NTE5/B can be converted to an NTE5/A by swapping the lower front face-plate. The lower front face-plate of either an NTE5/A or an NTE5/B is classified as the end user's property.
- I note there is a minor disagreement over terminology. Beattie's marketing originally announced The iPlate. Please note the definite article. The iPlate, as far as I am concerned, was a good idea blown up into a marketing con. It contains just a choke in series with what we call "the bell wire". (Subsequent newer versions of the NTE5/A actually have such a choke fitted to the back of the lower front face-plate.) The iPlate and the subsequent Service Specific Face Plates (SSFPs) are, ipso facto, interstitial plates -- they are fitted between the back (highly technical word alert) gubbins of the NTE5 (which is Openreach's domain) and the lower front face-plate, either an A or a B (as in karmie's case) (which is the EU's domain).
- Earlier in this thread I commented that it was not exactly clear from one of the photographs but it seemed as if there were two cables connected via their wires to the IDCs (of the B type lower front face-plate) and it appeared (to me) that all six wires (of the three pairs in each cable) were connected to the IDCs of the lower front face-plate.
Some very sage advice has been given (b*cat performs his best Japanese-style bow) by (in no particular order) JGO, HPsauce, Kitz and Black Sheep . . . However I do not think we are in full possession of the facts regarding karmie's infrastructure.
My suggestion, please, is for karmie to take a further series of photographs (when convenient). The photographs should clearly show:
- The wires present and connected to the back of the lower front face-plate (at the NTE5/B) -- we need to see which colour wire(s) is/are connected to each IDC.
- The wires present and connected at each of the (two?) extension sockets -- we need to see which colour wire is connected to each IDC or screw terminal, as appropriate.
Once the above has been seen and duly considered, then we can suggest one (or more) possible plan(s) of action.
Finally, would it be asking to much to know karmie's location? Just by county. If it is Surrey, then Walter (I am sure) would be willing to trundle his wheelbarrow to the required address and make any require adjustments. If it is Lancashire . . . etc. If it is Suffolk a certain grumpy old black cat might be able to plod there and stick his paw in . . . ;)
-
Well not exactly, the iPlate Mk1 does filter the bell wire to extensions so you are still misunderstanding.
NO - choking the bell wire doesn't stop common mode interference pickup on the other two or them acting as a stub filter.
It's far from perfect though it does help, but then I wasn't ever suggesting that it was used in this situation. ;)
In fact I never ever thought it was useful to me as simpler alternatives existed for anyone with a modicum of familiarity with phone wiring. 8)
As Kitz said, just disconnecting the bell wire everywhere that's possible will be as good as anything for now. :graduate:
-
Hi - had a lot to catch up on...
My location is Forfar in Angus - unfortunately we have had FTTC pushed back since mid 2012 and even though we were told our cabinet was ready they haven't taken any orders and estimate it will be before the middle of next year...
burakkucat has brought up an excellent point which I completely overlooked - if I were to take photos with a proper camera (not an AWFUL Iphone one) and proper lighting I think it would be much easier to determine a solution and what wires to remove for the bell wire fix and if possible what faceplate I can install myself given my setup - I will take the photos with a proper camera and upload them later in the day. From what I have seen online and on videos the faceplate install does not seem overly tricky provided you have the correct setup initially? I was reading up about the NTE5/B faceplate and it seems super strange how we have one as it is a regular 3 story detached house which is around 100 years old. House does not have an alarm and did not when we bought it in 2007 either.
Thanks again for all the input although some of it doesn't mean much to me it puts me on the right track to Google the rest.
-
My location is Forfar in Angus
Unfortunately, that is well out of the area for what we call 'Walter's Wheelbarrow', however if you get stuck or need something clarifying just yell and we will do our best to explain or advise further.
-
Hi guys - I just took a gamble and did what was suggested on Page 2 after a little time Googling to make sure I wouldn't do any damage.
Everywhere I read it just said leave IDC 2 & 5 connected and remove everything else so starting at the Master I worked my way through each of the 3 extensions and removed the cabling from IDC #1,3,4,6 and bent them round a screwdriver to keep them out of the way (pro tip from Youtube) and reseated them and tested with a corded filtered phone each time to make sure it was working. It is worth mentioning the phone line sounds WAY clearer than before which is hopefully a good thing.
I did take some photos but I don't have the cardreader on my PC and someone is away with my laptop until tomorrow (oversight on my part) - to simply word what was there (excuse if I get the wrong terminology as you know I am a rookie!)
NTE5B / Master socket: IDC #1,2,3,4,5,6 both had 2 cables crimped in each
Extension #1 (closest to Master and FULL of cobwebs, untouched for a LONG time): IDC #1,2,3,4,5,6 both had 2 cables crimped in each
Extension #2 (second closest to Master FULL of cobwebs, untouched for a LONG time): IDC #1,2,3,4,5,6 both had 2 cables crimped in each
Extension #3 (living room, router has ALWAYS run from this extension, PRISTINE inside): IDC #1,2,3,4,5,6 had 1 singular cable crimped in each
All I did was remove cables from all except #2 , 5 but where applicable I left both of the cables in if it was previously like that. Does that make sense? So Master,Ext #1 and #2 have 2 cables going into 2,5 ports and Extension #3 has 1 as it was when I started.
I would like to mention since doing this I have router uptime of 1hr 45mins with ZERO CRC,FEC,HEC,ES synced at 16383/1031 and attenuation has dropped from 16.5db to 14.0db downstream and the 9.0db upstream is now 7.5db. My router is reporting 10.5db SNR Margin down and 8.0db up and both have been rock solid (within .5db) for nearly 2 hours (at no point since I moved to Sky has my SNRM been this stable, usually jumps between 5-13.5db) - the only thing that concerns me is why I am syncing at 16383 on 14db attenuation? My Sky profile online says I am on a 16.2mb profile - is it possible that is capped? I currently have DLM entirely disabled on my line at my request and requested to manually enforce fastpath even though I had an unstable connection; my logic being the line was fine until evening times like 5pm-11pm but was fine during the daytime/late.
OBVIOUSLY 2 hours is not a great sample size but it is the best 2 hour sample my line has had since we moved to Sky. 0 errors or drastic SNRM fluctuations in 2 hours and a 'max' sync from what I can tell - I will run DSLstats for another 24hours and see how things pan out.
2 Quick questions for you smart folk:
1) Is my SNRM bad? 10.5db down and 8.0db up - I am sure this figure used to be lower. Am I correct in saying the 3db SNR profile on Be* is totally different to SNRM and higher SNRM = better stability or more room for error and lower SNR = higher syncs at the cost of some stability?
2) This is a Sky question - are the syncs capped based on 'profiles'? What else could cause such a poor sync speed with my attenuation (14.0db down/9.0db up)? Could it have something to do with the output power being low? I know the thomson tg585v7 misreports output power but AFAIK it reports it as 0dBm as opposed to the 1.5dBm it is reporting. I could try to plug in the Sky unit and check on that as I have no DLM on my line so I can disconnect it willy nilly :-D Would you guys suggest if the line is rock solid for 48h or so that I should contact Sky regarding the 16400 sync and see if they can retrain it or whatever?
Once again THANK YOU for all the help, I am praying this is the end of the road for these issues. I have no problem with the odd drop once a day but the connection was horrific on peak times before this.
-
GOOD ! nice to hear of someone's success. Often there is a resounding silence !
I would be inclined to make a final assessment after a week as patterns of use and possible interference will change through the week. BTW did you get my PM ?
-
As to the mis-reporting of downstream power, I have a vague feeling that it can be found by telnet to the router. Think results same as normal plus some extras.
-
Hi guys - I just took a gamble and did what was suggested on Page 2 after a little time Googling to make sure I wouldn't do any damage.
Everywhere I read it just said leave IDC 2 & 5 connected and remove everything else so starting at the Master I worked my way through each of the 3 extensions and removed the cabling from IDC #1,3,4,6 and bent them round a screwdriver to keep them out of the way (pro tip from Youtube) and reseated them and tested with a corded filtered phone each time to make sure it was working. It is worth mentioning the phone line sounds WAY clearer than before which is hopefully a good thing.
I did take some photos but I don't have the cardreader on my PC and someone is away with my laptop until tomorrow (oversight on my part) - to simply word what was there (excuse if I get the wrong terminology as you know I am a rookie!)
NTE5B / Master socket: IDC #1,2,3,4,5,6 both had 2 cables crimped in each
Extension #1 (closest to Master and FULL of cobwebs, untouched for a LONG time): IDC #1,2,3,4,5,6 both had 2 cables crimped in each
Extension #2 (second closest to Master FULL of cobwebs, untouched for a LONG time): IDC #1,2,3,4,5,6 both had 2 cables crimped in each
Extension #3 (living room, router has ALWAYS run from this extension, PRISTINE inside): IDC #1,2,3,4,5,6 had 1 singular cable crimped in each
All I did was remove cables from all except #2 , 5 but where applicable I left both of the cables in if it was previously like that. Does that make sense? So Master,Ext #1 and #2 have 2 cables going into 2,5 ports and Extension #3 has 1 as it was when I started.
That is certainly an improvement. I can almost picture your wiring infrastructure but there is still something missing that makes me :hmm:
Let us first consider the extension cables connected via the IDCs on the lower front face-plate of the NTE5/B. There are two cables present with wires only connected to IDC2 & IDC5. That is good. The presence of two cables implies that there must be a minimum of (at least) two extension sockets connected to the circuit.
I now see mention of three extension sockets. That fact is consistent with the two cables connected via the lower front face-plate of the NTE5/B.
Extension socket no. 1 is described as having two cables present with wires only connected to IDC2 & IDC5. That is good. We can tentatively deduce that one cable is the link between the NTE5/B and extension socket no. 1 whilst the other cable links to another of the extension sockets.
Extension socket no. 2 is described as having two cables present with wires only connected to IDC2 & IDC5. Ah, now what does that imply? We can tentatively deduce that one cable might be the link between the NTE5/B and extension socket no. 2 or it might be the link between extension socket no. 1 and extension socket no. 2. An uncertainty. Now what about the second cable at extension socket no. 2? It could be the link between extension socket no. 2 and extension socket no. 3 or it might be a link to something else, as of yet, unknown!
Extension socket no. 3 is described as having just one cable present with wires only connected to IDC2 & IDC5. That is good. We can deduce that socket is the end of either a chain of extension sockets or it is a socket fed directly by one of the two cables present at the NTE5/B.
Thinking it through and considering all the permutations of cables present at each location, there must be (or there was) something else connect to the circuit. Perhaps there is an extension socket no. 4? Or perhaps there was an extension socket no. 4? Perhaps there is evidence of the end of a cable, somewhere, which has been severed at sometime in the past?
The only way I could obtain a clear picture of the circuit would be to disconnect one of the two cables from the lower front face-plate at the NTE5/B and then test at each extension socket to see which is energised and which is not. There are so many possible permutations as to how your local wiring is connected that I am not going to list them all.
I will, however, mention just one of them. It could be possible that one of the two cables connected to the lower front face-plate of the NTE5/B is unconnected . . . it just winds its way off elsewhere into the property to a point where it has been disconnected or severed. If that is the case, then it should be left disconnected from the lower front face-plate of the NTE5/B. If that were true, then it means that the second of the two cables is the link to extension socket no.1 (or extension socket no. 2), which then is linked to extension socket no. 2 (or extension socket no. 1), which finally is linked to extension socket no. 3. Essentially a daisy-chain of three extension sockets. (See how confusing simple speculation can become. ::) )
-
Sounding good so far. :thumbs: Hopefully things are still the same today?
1) Is my SNRM bad? 10.5db down and 8.0db up - I am sure this figure used to be lower. Am I correct in saying the 3db SNR profile on Be* is totally different to SNRM and higher SNRM = better stability or more room for error and lower SNR = higher syncs at the cost of some stabilit
Without seeing full stats its hard to say. Also a set of stats immediately after a resync is the best way to check your SNR Margin. The 10dB sounds like it may be Skys DLM keeping your SNRm high. :hmm: Be* allowed a Target SNR Margin of 3dB if your line was stable.
Whilst related SNR, SNR Margin and Target SNR Margin are three different things.
If you want to know whats going on , theres a little bit of light reading on this page, it explains the difference between the 3. Ive tried to tone it down the techy jargon to a level that is factual yet understandable to most people. Start with ~ What is SNR? (http://www.kitz.co.uk/adsl/linestats.htm#SNR).
2) This is a Sky question - are the syncs capped based on 'profiles'? What else could cause such a poor sync speed with my attenuation (14.0db down/9.0db up)? Could it have something to do with the output power being low? I know the thomson tg585v7 misreports output power but AFAIK it reports it as 0dBm as opposed to the 1.5dBm it is reporting.
Sky DLM has profiles, which basically contains 2 parameters - [Target] SNR Margin & Interleave.
When you retrain (reboot) your router, then the line will sync as fast as it can with '[target x]'dB of SNR Margin. SNR Margin will continue to fluctuate and your router will report SNRm changes as they occur.
Having a high target SNRm will restrict the speeds you are capable of acheiving.
If your power output is too low, then yes it may slightly affect the Signal Attenuation measurement. Attenuation (http://www.kitz.co.uk/adsl/linestats.htm#attenuation) is a measurement of signal strength loss.
Less power means less signal strength, which in turn affects other parameters such as SNR and bitloading. The less bits loaded then the lower the sync speed. It also means that a line could become more susceptible to noise.
I could try to plug in the Sky unit and check on that as I have no DLM on my line so I can disconnect it willy nilly
Yes it would be a good idea to check. Although Im not convinced about you not having DLM on the line. :-\
Also as suggested earlier, you could also see what measurement you get from TG585 via the telnet CLI to make sure its not just a GUI reporting error. Ive forgotten off the top of my head what the cmd is, but you shouldnt need to this as DSLstats should give you the raw data under:
Telnet Data -> Connection Stats.
-
Excellent news.
Just to echo and support what Burakkucat has said, tracing and verifying all the remaining cables (in 2 and 5) and totally disconnecting all the unknown/unneeded ones would be good.
A final point is that, ideally, the phone cables (2 and 5) should be a single twisted pair. e.g. orange/white and white/orange. A mixture of colours will be in separate pairs, not twisted together, so even more prone to interference.
-
A final point
Never say that....... :blush:
Anyone who connects all 6 wires in the cable is almost certainly not a professional and doesn't really know what they're doing, so this is probably an amateur/DIY set of extensions.
As such almost anything could be wrong. A couple more things to check for:
1. Correct extension sockets; look on the faceplates to make sure they're not master sockets reused incorrectly.
2. Star wiring. Ideally all the extensions should run in a single chain with no branches, forks or stars.
I think you were asked above to describe the wiring in a bit more detail, so if you can work out the wiring topology, starting at the master socket, we may be able to advise.
With modern cordless handsets and wireless routers it's now almost always possible to eliminate any star wiring by simplifying to a core set of sockets that are genuinely needed.
-
DSLstats of the last 24hours: http://imgur.com/a/7hwCd#0
Connection has held sync for over 24hours (which is a first!), 0 FEC/HEC/ES and 21 downstream CRC errors per 24 hours (0.89/hour.)
Regarding Burakkucats post I believe the missing link COULD be the secondary phone line we used to use which was cancelled around the same time we got forced on to Sky. We used to have a seperate business line but then moved over to the BT service that allows you multiple numbers with different ringtones - this could be the missing piece to the wiring? Hopefully the wiring adjustment will allow the line to remain in its current condition because I have absolutely no complaints since then and it doesn't need to be 100% perfect provided I could get it more stable in the evenings (touch wood on that one.)
Kitz, I am not sure about the DLM thing - the guy I spoke to said he would 'manually configure my profile' and it could only be adjusted by ringing them back up. The line was dropping 10+ times a day for 3 days and still remained on fastpath with the lowest interleave depth and before DLM would jump at any chance to interleave the line - unless they are able to decrease DLM sensitivity somewhat. As you can see from the telnet stats below the output power is reporting extremely low.. I will plug in the SR102 Sky box and see what it reports after this post. If I were to phone Sky and have them manually adjust the profile or set the target SNR Margin lower to increase the sync speed could it cause instability issues again or are my line stats reasonable since the wiring modification? I would obviously love to have a couple of extra mbits down but I wouldn't risk stabilty over it.
TELNET DATA:
Bearers generic info DS US
Payload rate [kbps]: 16383 1031
Attenuation [dB]: 14.0 7.5
Margins [dB]: 10.5 7.0
Output power [dBm]: 1.5 10.5
HP - I am assuming you mean I should keep the phone cables in 2&5 'wrapped' together and away from all the other 'dead' cables in the socket? I made a point of seperating all the non 2&5 cables so they didn't touch ANYTHING the moment they left the sheath/split in the cabling - they are nowhere near the live 2&5 socket cables (other than inside the white wire sheath or however you say) - I wrapped them up and tucked them out of the way. Could have snipped them with scissors too I guess? If the line does degrade or it is necessary I could most likely do a better job of tidying the cabling up. I would absolutley hate to pull out these additional cables in 2&5 and not have the crimping tool to fix them on hand immediately if needed though.
Thanks again for all your help guys and I now am beginning to see a light at the end of tunnel! Provided everything remains as it has been for the past 30hours I would have no complaints!
-
HP - I am assuming you mean I should keep the phone cables in 2&5 'wrapped' together and away from all the other 'dead' cables in the socket?
No, not really, I'm talking about the way the cables are arranged inside the outer sheath as they run around your house as that's where they pick up interference like a huge radio aerial. The colours used will tell anyone who knows about phone cables whether that is correct or not.
What are they?
-
DSLstats of the last 24hours: http://imgur.com/a/7hwCd#0
Connection has held sync for over 24hours (which is a first!), 0 FEC/HEC/ES and 21 downstream CRC errors per 24 hours (0.89/hour.)
Those graphs look really good. :clap2:
Regarding Burakkucats post I believe the missing link COULD be the secondary phone line we used to use which was cancelled around the same time we got forced on to Sky. We used to have a seperate business line but then moved over to the BT service that allows you multiple numbers with different ringtones - this could be the missing piece to the wiring?
No, I don't think that is relevant. :no: You have provided clear evidence of an unknown length of cabling that is still connected into the circuit. Try as I might, I cannot think of any rational scenario other than there is another extension socket somewhere of which you are unaware or there was another extension socket which has subsequently been removed, the cable to it severed off flush to a surface and then decorated over.
I would absolutley hate to pull out these additional cables in 2&5 and not have the crimping tool to fix them on hand immediately if needed though.
No, please do not attempt any other modifications. I do not want you to start trying things whilst we are essentially "blind". Sight of further photographs of the infrastructure, when convenient, may provide us with further information.
Perhaps your new found line stability may prove to be adequate -- until a VDSL2 based service becomes available to you, at some time in the future.
-
I'm a bit happier with your attenuation being at 14dB which is nearer the realms which I'd expect. Also, when you were on Be, you had Annex_M, so that would have meant less downstream bins, so it could have affected your attenuation slightly by say 1dB when going back to Annex_A
On the subject of bins and frequencies there is a slim chance that your internal wiring may have been picking up interference in one of the tones that Be didnt use. I wouldnt swear on it though because i wouldnt have expected the impact to be quite so bad as it was for just swapping between Annex A/M. However, it is about the only explanation I can come up with thus far.. but Im not happy enough about that excuse for me to say yeah that will be what it is. :hmm:
-----------
Wooooooooah.. Ive just had a lightbulb moment. :idea: A convo me and rizla had several years back about Skys network and how they used ASSIA kit. Damnit I cant find it now. Cant recall if it was in the forum or via PM or via mail :(
But.. I think somewhere in the back of my mind I recall trying to look into ASSIA and what it configures. It can do some of the usual stuff like QoS, but the flash of inspiration was recalling something about cutting power back on the dslam.
Rizla are you out there, can you remember?
BBL if I can find that convo.
-
Sorry I cant find the convo... I think it must have been personal mail via my ipad, rather than site mail.. in which case I wont have a copy. :(
I have however found this thread, which touches on the subject of ASSIA, cutting back power, and how it was restricting the very good lines to speeds below their true capacity. The bit half way down makes interesting reading.
What we do know for sure now though is that Sky most definitely does not use the same DLM as BT. It works backwards and because of this its why some lines were getting stuck and never reaching their full potential.
http://forum.kitz.co.uk/index.php?topic=10975.0
-
Sorry for answering myself again. But after having re-read that thread.. IF that bug still applies to Skys DLM, and if its still affecting the occasional short line.. then I think we may have found what your problem is.
See how snadge mentions seeing a line what should run at 22Mbps ending up at 13Mbps and quote "what about 20Mb lines that end up at 16Mb" Sound familiar?
As I mentioned several posts back, a very low power output will make the line more susceptible to interference and noise. The signal is weaker so errors that a full strength line can cope with are more likely to materialise. I suspect that you may well have had the wiring issue for a while.. but it only became apparent now because you're not running at full pelt.
I'd love to know if Rizla knows if that bug ever got fixed. He's probably the best man to answer as it was him that discovered the bug in the first place. Unfortunately he often works away/abroad so may not pick up on this for a while.. but I'll try contacting him later via other means.
-
Hi again
Line has had an uptime of over 2 days and 9 hours with a 24 hour average of: 0 FEC/HEC/ES and downstream CRC errors of 5.45 per day / upstream CRC of 3.23 per day. Seems like a drastic improvement and in its current condition I would say I am perfectly happy with how it is currently operating.
Kitz - I think your estimation is spot on with regards to the output power and Sky. I completely agree the internal wiring here was evidently always not great but it only became evident upon the swapover to Sky and lower power output as we had been running fine for nearly 5 years. I think it is absolutely shocking that they are trying to save money to that degree by causing other people misery and most importantly in some circumstances having customers PAY out their own pockets believing and being told their lines are at fault when in reality a stronger signal would allow the service to operate even under suboptimal conditions. As a test I did plug in the SKY SR102 in an attempt to extract output power data from it but was only presented with these statistics:
Broadband Link Downstream Upstream
Connection Speed 16383 kbps 1031 kbps
Line Attenuation 14.0 dB 7.0 dB
Noise Margin 10.5 dB 7.5 dB
Perhaps there is a way to extract further information from the SR102 but presently I plan on remaining with the Bebox and perhaps giving Sky a ring to ask them to manually reprofile the target SNR which may allow slightly higher sync speeds? I don't believe the small amount of erroring or fluctuations on the SNRM would present any problems if the line were to sync at a higher rate (provided it is not being restricted by the output power issue.) I would no longer consider this a pressing matter though as a small downstream/upstream increase would merely be the icing on the cake as the main goal was to have a stable fastpath line which I *believe* we now have :-)
Thank you all so much for the excellent advice presented over this thread I was completely fed up of playing ping pong with Sky and BT and was ready to just cave in and get a new line installed with a new provider. I should have known something was up when they cut the price we were paying with BE* (20 odd quid a month) down to £7.50 - with Sky you definitely get what you pay for..
-
Without reviewing all the previous posts to this thread, I suspect that your line will now behave with a target SNRM of 6 dB.
-
Ive not been around for a few days, so only just picking up on this now, but from all the symptoms it does look like this is what happened to you. I cant think of anything else that would make as much sense.
The line certainly seems in much better condition now, and asking them to re profile the target SNR seems like a good move.
Unfortunately it turned out I never got chance to contact rizla before I went AWOL, but I have now pointed out this thread to him and Im sure that if he has anything of value to add, then when he has time he will do so.
-
AFAIK the DLM "sticking" is no longer a widespread issue, in fact Sky's DLM is now pretty much top notch these days (much MUCH better than BTs). The offending module in the ASSIA suite has long since been replaced (best part of 18 months ago when they trialled g.INP) and Sky's DLM is now pretty aggressive in terms of maxing out lines.
As the OP moved from Be and is now (presumably) on BB unlimited then he will be (should be) talking to exactly the same people he was at Be - the Bulgarians run the support seperately for "Broadband Pro", which is defined as either the 80/20 VDSL2 service or the BB unlimited service for ex-Be users so I don't see where the moaning about them is coming from. Personally I found the Bulgarians to be bloody useless but the OPs support hasn't changed at all so I dunno what the issue is there.
NB ex-O2 and "normal" BB unlimited customers don't get the "premium" support and have to make do with normal Sky support.
The problem is that there is no way to manually set a profile and have it "stick" these days - DLM cannot be stopped on a single line, the module doesn't work that way so if the OP wants the line to max out (after whatever issues he's had) then he's just going to have to wait a bit. If the error counts are accurate then DLM will drop the margin to 3dB and rely on g.INP to sort out transient issues. If not then it'll probably settle at 7dB.
YMMV but I've found Sky to be better than Be in just about every way - not having to put up with Telefonica's shonky bargain basement routing was (and is) a big plus. In addition, unlike BT Retail/Plusnet Sky has decent peering and doesn't just rely on L3 to do everything ;)
-
Thanks for the input Rizla...
... and as always, you and I will continue to agree to disagree re BE support and Skys DLM ;D
-
You wouldn't disagree about the DLM if you saw how it worked on my line :D I think snadge got 3-4Mbps extra with it too. Seriously good with the g.INP module kitz, really is.
I just found that when you had a specific issue and could demonstrate it at will (like Telefonica's shonky routing) the Bulgarians would be more interested in how it looked on the forums rather than addressing the issue. No real difference from the (pre-Telefonica) days when Brett was in charge - all about image, not much substance IMHO. I wasn't best pleased to see them becoming my "support" again on Sky and told them that in no uncertain fashion :)
For me the top 5 ISPs I've had (in descending order) are :
1) UK Online (Easynet) by a country mile - one issue sorted by Easynet escalations (Dan the man);
2) AAISP - expensive as hell but no muppets/scripts anywhere;
3) Sky - on the three occasions I've had to contact tech support (DLM stuck/VDSL self-install fault/problem with Sky TV box) they've been fine. Admittedly I phone them during the day so YMMV;
4) Zen - despite being robbed (CC fraud) by one of their staff & them totally ignoring their own complaints procedure (twice) they were reliable;
5) Be - awful both times although slightly less awful the second time. I had more tickets open with Be than with all the other mentioned ISPs put together ::)