Print

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

[MartinGoose]

After a period of many months since last summer when my REIN issues were cleared up (neighbour's exercise treadmill!), problems have set in. Over most of that period I have had a stable connection (ADSL2+) of just under 12MB/s @ 3dB SNR with a downstream attenuation of just under 40dB that was stable and ran for weeks without a resync. All this as reported by my 2Wire BT2700HGV running version 5.29.107.12 firmware.

Over the last three weeks the line has become noisy and 7MB/s @ 6bB is now the norm. Monitoring with Routerstats shows sudden jumps in SNR both up and down. Overnight SNR drops far more than previously (6-7dB).

I have done all the swapping of hardware, filters etc to prove my own equipment and opened a ticket with my ISP Plusnet. This has not got beyond the 'please carry out the following tests' stage yet. There is only 1 metre of cable between my master socket and where the Openreach cable comes out of the ground. The BT2700HGV plugs directly into the filtered faceplate and then its Ethernet all the way from there.

It occurs to me that this degraded performance might nevertheless be within the acceptable range for my line (copper all the way to the exchange) albeit much less than previously. Any views on how hard to push this with my ISP?

[burakkucat]

Hello Martin. I certainly do remember you from last summer's treadmill issue.

Over the last three weeks the line has become noisy <snip>

To which type of noise are you referring? Audible, when using the telephone or EMI/RFI as shown by a well depressed SNRM? I assume you mean the latter . . . 

Whichever, have you performed the most fundamental test -- plugged a corded telephone direct into the test socket of the NTE5/A and performed a aural line quality test?

Has there been any change in your DS or US line attenuation? If yes, then that is a good indicator of a (developing) line fault. If either of the attenuation values is different (each time it is checked) and there is a hiss, crackle or other audible sounds when using the telephone then that is highly indicative of a (developing) HR (high resistance) fault. If that is the case, it should be reported to your telephony service provider as a telephone fault (do not make any mention of the broadband service) and they should book the services of an Openreach network (telephony) engineer to resolve the issue.

[snadge]

gawd I remember the Treadmill..lol

IN FACT, I use it as evidence when im helping others with similar problems...that and the one about a tumble dryer that knocked the whole streets broadband off when in use...

Ive just gone through 2 months of hell with sky and 7 engineer visits to get mine fixed...

B'Kat has offered up the advice anyway so I wont comment on that , but I will say that if PlusNet dont resolve to get someone out too look at it then perhaps you can somehow *ahem! "make it worse?" ...hehe so that an engineer is called out? - just a suggestion, probably a bad one but sometimes extreme measures need to be taken.. you say its overnight when it happens? I wonder if its street lights? or neighbours dryer hehe , or is it like that permanently?

best of luck anyway

[MartinGoose]

Just as I was about to do a quiet line test at the test socket, having done one with telecoms apparatus connected normally, I find that the SNR had improved overnight. Line stats changed to 7150 kbps | 12.1 dB | 40.3 dB at around 07:20 this morning and the router resynced to 11386 kbps | 3.0 dB | 39.3 dB. This is pretty much what I was getting before the problems started.  Some minor fluctuations in SNR since, but otherwise stable. I won't be taking an axe to the cable just yet

Will continue monitoring for the time being. Hourly updates on my Routerstats can be found here:
http://www.thegoosefamily.plus.com/Rx_Noise_margin_%28dB%29.jpg

[burakkucat]

Will continue monitoring for the time being. Hourly updates on my Routerstats can be found here:
http://www.thegoosefamily.plus.com/Rx_Noise_margin_%28dB%29.jpg

That graph does not show anything out of the ordinary, at the moment. 

Is there any possibility that you could also record and plot the corresponding attenuation value? In theory it should be a perfectly straight horizontal line . . . If it jumps up and down, etc, then that will be evidence of a fault.

For an example of what a plot of attenuation should not look like, please see one of Bald_Eagle1's epic threads. I'm sorry but I can't remember the link to the relevant posting. However I'm sure that Baldy_Bird will be passing by, later on, and he will provide a link. 

[Bald_Eagle1]

This one by any chance?

[burakkucat]

   Hmm . . . Got anything more jittery / alpine looking? 

[MartinGoose]

Added to my uploading script to include the graph of downstream attenuation.
http://www.thegoosefamily.plus.com/Downstream_Attenuation.jpg

Everything is good at present (Except for Polish goalkeeping that is!)

[MartinGoose]

After 3 good days the problem is back. It returned when I was away from the house and SWMBO was very inconvenienced!

I finally tried the quiet line test at the test socket behind the faceplate with a cheap wired phone. There is a very noticeable buzz on the line and Openreach are visiting on Monday am.

PS I currently have a faulty street light just outside the house. It fails to come on. Fault reported and due for a fix (eventually). Perhaps the problems are linked? The timescales are certainly comparable.

[kitz]

>>> Perhaps the problems are linked? The timescales are certainly comparable.

It wouldnt be the first time, although as you say 'eventually' depends on how quick your local council are...  and I wouldnt be surprised if the OR visit is first.   Good luck

[guest]

Its quite possible.

Ballasts really don't like having a faulty lamp attached.

Low-pressure sodium lamps don't generally cause the ballast to cycle but they're not that easy to dim and have terrible colour rendering.

The easy way to tell is if the lamp starts off a dull red colour and takes a few minutes to brighten then its low-pressure. If not then its high-pressure and they have a known failure mode at end of life which causes the ballast to cycle - if the lamp was flickering in colour for a while before failure then its probably high-pressure and there will have been significant electrical noise radiated from it. If you don't prevent the ballast from repeatedly cycling then it becomes damaged.

[MartinGoose]

Its quite possible.

Ballasts really don't like having a faulty lamp attached.

Thanks for the info on street lamps and the types of faults.

Does the electrical noise appear at 612kHz as previously discussed here? If so I may go off and survey the lamp posts!

[guest]

Thanks for the info on street lamps and the types of faults.

Does the electrical noise appear at 612kHz as previously discussed here? If so I may go off and survey the lamp posts!

It will vary but it is MF noise and harmonics of that usually. It's dependent on a variety of things, not least of which is the type of lamp - there are a variety of spectrums, all of which arc at slightly different frequencies. The lamp temperature will affect the frequency for example.

For a magnetic ballast it'll be low end of the MF band.

If its an electronic ballast (generally called "digital ballasts") then the arc frequency will vary depending on ballast and whether/how its being dimmed. The early revisions of electronic ballasts also had a tendency to radiate lots of RF noise from the cable between the ballast and the lamp. This has been largely fixed in later revisions but it does require the screen to be connected at both ends (which frequently doesn't happen) otherwise it can actually make things worse. I would expect an electronic ballast on full power to radiate at around 1MHz and if the lamp is being dimmed then anything from 500kHz - 1MHz is likely. There are lots of harmonics present too - usually visible on a spectrum analyser up to around 8MHz although they are very low-level - unless the cable isn't screened.

If your line was perfectly balanced then the noise shouldn't be an issue as it'll be present and in-phase on A & B wires so should be cancelled out at the ADSL modem input (which is a differential amplifier but I digress). A perfect line balance is unlikely and causes the noise on A/B to be out of phase, this in turn means the diff amp won't reject the noise signals as well as it ideally could.

tl;dr HID lamps work by striking an arc to form plasma. They are inherently noisy when brand new. Things only get worse from there

[MartinGoose]

The faulty lamp was replaced replaced on Monday. I managed to speak to the technician that was dealing with it and persuaded him to replace all the components with new. I was worried that Openreach would then arrive, not find any noise on the line, and charge me as it must have been my equipment. Fortunately Openreach turned up before the he left and they spoke to to exchange information.

Since then the voice line has remained quiet and the broadband performance stable but degraded compared with its previous excellent state.  My ISP tells me that its current performance is within the expected performance band, so I will not be pushing it any further.

[coolsnakeman]

Interesting fact on the street lamp. My ex has a street lamp outside her house which tends to know on and off so i need to ask her if it is still doing this as she does encounter some speed problems.

Gary