Kitz Forum

Broadband Related => Broadband Technology => Topic started by: Oranged on January 08, 2009, 11:17:02 AM

Title: Line attenuation suddenly lower
Post by: Oranged on January 08, 2009, 11:17:02 AM
I've had broadband since 2004 and the line attenuation has always been either 37.5 or 38. This was the case with Orange and also now that I'm with O2.

Yesterday evening at 8.47pm, the router unexpectedly re-booted while at SNRM 6.5 and the line attenuation is now 34.5.
There have been no variations to any of the other data eg  the sync, SNRM or Output Power data.

Currently stats are :

Bandwidth (Up/Down) [kbps/kbps]: 1,321 / 8,187 
 
Output Power (Up/Down) [dBm]: 12.0 / 19.0
 
Line Attenuation (Up/Down) [dB]: 19.0 / 34.5
 
SN Margin (Up/Down) [dB]: 7.0 / 8.5

Any suggestions why this should be ?
Title: Re: Line attenuation suddenly lower
Post by: roseway on January 08, 2009, 11:30:07 AM
Hmm.. that is rather a large change. Of course, the reported attenuation isn't a simple single figure, but a sort of weighted average of the various attenuations of the different bins, so it can change depending on how the bins are used when the connection is negotiated. But I don't think that would explain what has happened here. Apart from that waffling, I really don't know.
Title: Re: Line attenuation suddenly lower
Post by: Oranged on January 08, 2009, 12:05:06 PM
But that's a reassuring response because I'm baffled as well  ???

I've asked O2 if they can check connection data from their end as I've had a small number of unexplained reboots recently but the connection always resumes immediately with no strange stats......except for last evening.
Title: Re: Line attenuation suddenly lower
Post by: kitz on January 08, 2009, 12:21:31 PM
As you already know attenuation should more or less always remain static.
As eric has already stated, it can be affected by different routers due to the way they record the frequency bins, or it can increase if you get higher speeds and therefore use the higher frequencies which are more attenuated (eg if you go to adsl2+).

A decrease in attenuation is normally different equipment on the line (router), improvement of internal wiring, a repair on the physical phone line or.. a new twisted pair.

TBH although O2 may perform a woosh equivalent, I wouldnt expect them to find anything much since it will be your router that responds to the tests and will likely give similar figures.

If it wasnt for the time at which this occurred, I would have said most likely some work has been carried out on your line which has improved the attenuation such as a new twisted pair or linecard.
Title: Re: Line attenuation suddenly lower
Post by: Oranged on January 09, 2009, 10:47:37 AM
A further unexplained reboot occurred at 3:35am this morning after which the attenuation reverted to 37.5dBm.

These reboots are occurring regularly even though SNRM when it happens never seems to be below 5dB.

Current stats are :

Bandwidth (Up/Down) [kbps/kbps]: 1,325 / 8,187
 
Output Power (Up/Down) [dBm]: 12.0 / 18.5
 
Line Attenuation (Up/Down) [dB]: 19.0 / 37.5
 
SN Margin (Up/Down) [dB]: 7.0 / 9.5

So it does seem likely that there is a line problem of some description.

By the way, this is an O2 Standard ADSL2+ LLU service plugged into the main BT linebox using an O2 Wireless Box II TG585v7. I'm going to try my Belkin 7633 router and monitor if that reboots.
Title: Re: Line attenuation suddenly lower
Post by: kitz on January 09, 2009, 12:33:59 PM
>> A further unexplained reboot occurred at 3:35am this morning after which the attenuation reverted to 37.5dBm.

:shrug2: maybe a false 34.5dB reading?

>> using an O2 Wireless Box II TG585v7.

There have been instances of some of these boxes doing unexplained reboots, so it may be worth giving the belkin a shot.

Title: Re: Line attenuation suddenly lower
Post by: Ezzer on January 09, 2009, 01:17:42 PM
I would say its a false one off reading, was your router particuarly cold during the reboot perhaps.

If it was a change that stuck then I'll say some work has occured somewhere along you line between you and the exchange, but given the figures gone back to normal I'lld rule this out
Title: Re: Line attenuation suddenly lower
Post by: Oranged on January 09, 2009, 01:40:50 PM
I would say its a false one off reading, was your router particuarly cold during the reboot perhaps.

O2 router is in an unheated room within the house......would this make such a difference ?

I thought TG585v7's ran warm.....I have mine stood on it's front edge, with the aerial on the back edge uppermost so heat is dissipated.

The reboots that have happened have been at varying times of the day/night, not necessarily only at night when it's coldest. Having said that, the reboots have only started since the sub-zero external temperatures have been with us.
Title: Re: Line attenuation suddenly lower
Post by: Ezzer on January 09, 2009, 01:56:53 PM
I woder if you might be getting condensation inside the router, What made me think about cold temerature is if electronics are cold then the same energy from the current used has a more distinct/ crisper effect. Possibly (although only a theory and I WOULD NOT RECOMEND USING THIS AS A POSSIBLE MEANS OF TRYING TO CHANGE ROUTER PERFORMANCE) if the router was particuarly cold could have the dsl signal appear to be relitively stronger hence the lower attenution reading.

If this was the case the input signal is still the same as before, just the way the router has interprited the signal, this should naot make any difference to the sync speed which is one half of the reason I put the above bit in capitals.
In this respect its different to the effect of having a cooler cpu chip in a pc where a cooler temeratur does have a change in performance.

If the area ia particuarly cold with the router running at its normal temperature then maby condensation could be building up rather like dew and if theres any mineral content in the condensation (which will occur after some time) this could be causing a hr fault in the router and so the reboots.
Title: Re: Line attenuation suddenly lower
Post by: mr_chris on January 09, 2009, 02:51:47 PM
could have the dsl signal appear to be relitively stronger hence the lower attenution reading.

I wouldn't imagine it would change by as large an amount as 3dB though.

Condensation sounds like a plausible theory though - and as it's a TG585v7 anyway, I have heard these aren't particularly fantastic routers to begin with - so perhaps a bad combination.

So I'd better cancel my custom Botch-Jobs-R-Us router liquid nitrogen cooling kit then, should I? :P
Title: Re: Line attenuation suddenly lower
Post by: Oranged on January 09, 2009, 02:57:20 PM
That's an interesting and logical theory.....before I dust off the Belkin, I'll move the TG585v7 to a warmer position  ;)

Unfortunately that's an extension (hard-wired from NTE5 minus ring wire) but I had no problems using for BB previously.

Re. B-J-R-Us, you could always first try putting one of those small freezer packs on top of the router  :lol:
Title: Re: Line attenuation suddenly lower
Post by: Oranged on January 09, 2009, 03:22:37 PM
I've moved the router to a position in a bedroom on the floor behind the bed......cosy now  :D

In this location I'm only connecting at 24Mbps of the 54Mbps to the router but signal quality is "Very Good" and stats are still OK :

Bandwidth (Up/Down) [kbps/kbps]: 1,325 / 8,187
 
Output Power (Up/Down) [dBm]: 12.0 / 18.5
 
Line Attenuation (Up/Down) [dB]: 19.0 / 37.5
 
SN Margin (Up/Down) [dB]: 7.0 / 7.0

and a speed test shows 7100/1100......so that's as good as before connected @ 54Mbps.....so now I just need to monitor the reboots with RouterstatsLite.
 
Title: Re: Line attenuation suddenly lower
Post by: roseway on January 09, 2009, 03:28:43 PM
Quote
and a speed test shows 7100/1100......so that's as good as before connected @ 54Mbps

The 54 Mbps (now 24 Mbps) is just your wireless speed, and has nothing to do with your ADSL connection speed. As it's way above the ADSL speed it won't affect your actual download speed either.
Title: Re: Line attenuation suddenly lower
Post by: Oranged on January 09, 2009, 03:53:59 PM
I appreciate that this is wireless speed to the router but I was inferring that an ethernet connection at 100Mbps sometimes produces a better result than a wireless connection on speed test sites.......I phrased it poorly  :doh:
Title: Re: Line attenuation suddenly lower
Post by: Ezzer on January 09, 2009, 06:31:43 PM
Don't go overboard with keeping the router warm as they ALL with become tetchy to loosing connection altogether if too hot. had 2 dsl fault on one day last week both due to overheating routers.

Rule of thumb when you place your palm on the top or bottom of the router and if it were a cup of tea/coffee could you drink it, if yes, too hot. (Voyagers tend to run warmer than most routers but should only be a bit warm near to where the power lead plugs in.

As far as cold routers are concerned. I'm guessing on this one. Cold electronics can start to misbehave. some processors though get a big improvement in performance when cooled down hence some hard core pc nuts actualy refrigirate their cpu's in their pc's. Cooler complex intergrated circuits get as best can be described a better contrast with the digital signals running within them, hence my guess.

The last thing I want to do is start some silly urban myth and then have people going round refrigerating their routers in the belief it will improve performance. I'm guessing a cold temperature may cause an sligh change in the stats given but the actual signal loss on the line will remain the same.
Title: Re: Line attenuation suddenly lower
Post by: orainsear on January 10, 2009, 12:42:31 PM
You could be seeing the attenuation change due to the effect of the weather.  The electrical conduction of copper improves as temperature lowers, hence you may see a subequent drop in the attenation.  I've seen the attenuation of my own line drop by around 8% with a 20 C change in the temperature.
Title: Re: Line attenuation suddenly lower
Post by: roseway on January 10, 2009, 01:10:29 PM
I'm a bit dubious about that theory. The attenuation of a line to ADSL signals is predominantly due to its high frequency AC losses, not the resistance of the copper. That's why the downstream attenuation is higher than the upstream - it uses higher frequencies.
Title: Re: Line attenuation suddenly lower
Post by: orainsear on January 10, 2009, 02:31:56 PM
It's a natural physical phenomenon that designers have to make allowances for.  Telecoms Tech Reference (http://books.google.co.uk/books?id=i4nwxZjpOa8C&pg=PA436&dq=dsl+attenuation+temperature&lr), DSL Tech (http://books.google.co.uk/books?id=Jjkd74jY47oC&pg=PA223&dq=dsl+attenuation+temperature&lr=), DSL Advances (http://books.google.co.uk/books?id=VWaMdIq2w8QC&pg=PA69&vq=temperature&dq=dsl+attenuation+temperature&lr=&source=gbs_search_s&cad=0)

Edit: fixed hyperlinks
Title: Re: Line attenuation suddenly lower
Post by: Oranged on January 10, 2009, 03:23:37 PM
Well my attenuation has returned to "normal" i.e. 37.5dBm and it's still zero outside or thereabouts so there's always one example that bucks the theory.  :-\

Since moving the router from the unheated room to a floor location in a bedroom I've had no reboots in 24 hours.

Current stats are :

Uptime: 1 day, 0:04:38
 
DSL Type: G.992.5 annex A
 
Bandwidth (Up/Down) [kbps/kbps]: 1,325 / 8,187
 
Output Power (Up/Down) [dBm]: 12.0 / 19.0
 
Line Attenuation (Up/Down) [dB]: 19.0 / 37.5
 
SN Margin (Up/Down) [dB]: 7.0 / 6.5
 
Vendor ID (Local/Remote): TMMB / BDCM
 
Loss of Framing (Local/Remote): 0 / 0
 
Loss of Signal (Local/Remote): 0 / 0
 
Loss of Power (Local/Remote): 0 / 0
 
Loss of Link (Remote): 0
 
Error Seconds (Local/Remote): 0 / 0

What is strange is that even though the router is now using a hard-wired extension box (ring wire removed) I am seeing no Loss of Signal or Error Seconds which I was seeing using the main BT linebox even when the room temperature was 18c+.

There are FEC, CRC and HEC errors but nothing to concern me for a 24 hour period.
Title: Re: Line attenuation suddenly lower
Post by: orainsear on January 10, 2009, 04:05:04 PM
Well my attenuation has returned to "normal" i.e. 37.5dBm and it's still zero outside or thereabouts so there's always one example that bucks the theory.  :-\

 ;D Doesn't surprise me at all.  For all the theory and laws of physics a large part of understanding ADSL seems to be what I would describe as an art, and every line seems to have distinct characteristics.

Out of interest how long have you been using the TG585v7?  As has already been mentioned, quite a few people have had problems with this device. 



Title: Re: Line attenuation suddenly lower
Post by: Oranged on January 10, 2009, 04:58:50 PM
ADSL seems to be what I would describe as an art

Or rather a "black art"  ???

Since moving over from Orange to O2 on 16/9/08 I've used the O2 router with no real complaints until the problem in this thread which would seem to have been of my own making by expecting the router to live in an unheated room during the coldest early winter since who knows when  ::).....so it doesn't like getting hot and it doesn't like getting cold  :baby:

I still have my trusty Belkin 7633 (used since May 2005) which I set up and tested on the first day with O2 then put it away in its box.

I think routers are like any other modern technical device.......you get perfectly good ones and you get absolute lemons.....it's the luck of the draw.

The only thing I don't like is that the full-blown version of Routerstats won't work with it.....I try each new version but still no luck so I stick with Lite.
Title: Re: Line attenuation suddenly lower
Post by: roseway on January 10, 2009, 11:19:55 PM
It's not so much an art as a very complicated science. Electronics engineers will understand all the factors, but there are so many variables that every line is different. In principle it would be possible to make lots of individual measurements on a line and so determine how to get the best result for the user, but the cost of doing this would be totally prohibitive. So we have to resort to practical experience and a fair bit of 'suck it and see'.

Of course an extra complication in the mix is that there's no universal agreement on what is meant by 'best'. For some people it will mean fastest, for others it will mean most stable, and for others again it will mean lowest latency. Or some compromise between the three. So in a practical sense, yes, it's a black art. :)
Title: Re: Line attenuation suddenly lower
Post by: Oranged on January 12, 2009, 04:06:49 PM
Still no reboots since plugged into the bedroom extension, current stats are :

Uptime: 3 days, 0:55:11
 
DSL Type: G.992.5 annex A
 
Bandwidth (Up/Down) [kbps/kbps]: 1,325 / 8,187
 
Output Power (Up/Down) [dBm]: 12.0 / 19.0
 
Line Attenuation (Up/Down) [dB]: 19.0 / 37.5
 
SN Margin (Up/Down) [dB]: 7.0 / 8.0
 
Vendor ID (Local/Remote): TMMB / BDCM
 
Loss of Framing (Local/Remote): 0 / 0
 
Loss of Signal (Local/Remote): 0 / 0
 
Loss of Power (Local/Remote): 0 / 0
 
Loss of Link (Remote): 0
 
Error Seconds (Local/Remote): 1 / 0
 
FEC Errors (Up/Down): 458 / 5,189,926
 
CRC Errors (Up/Down): 458 / 382
 
HEC Errors (Up/Down): 4,037 / 260

So  it would seem that the cause was probably the router being in an unheated room and possibly being affected by condensation.
Title: Re: Line attenuation suddenly lower
Post by: jeffbb on January 12, 2009, 05:55:01 PM
Hi
Hard to believe as after awhile the average router would be warm enough to get rid of any condensation . But hey don't matter its working  ok now :clap:

regards Jeff
Title: Re: Line attenuation suddenly lower
Post by: Oranged on January 12, 2009, 07:36:26 PM
Hard to believe as after a while the average router would be warm enough to get rid of any condensation.

Well that was my thinking but as that's the only change I've made, I can't put it down to anything else. I know the BT linebox is OK as I ran a check using the test socket and got the same results as when using the linebox normally. So even though I'm now using an extension, albeit hardwired from the main linebox, this is giving better stability.

Not knowing anything about dew point and relative humidity to be able to explain the results, I guess I'll just have to be thankful that the reboots have stopped  :D
Title: Re: Line attenuation suddenly lower
Post by: Oranged on January 13, 2009, 05:54:43 PM
Just one more item of data before I finish with this thread.

Would anyone like to comment on the DMT graph shaping and the gaps......are these results anything to do with the reboots ?
Title: Re: Line attenuation suddenly lower
Post by: kitz on January 13, 2009, 07:48:03 PM
Hard to say - the graph displays your bitloading only... unfortunately when Thompson b@stardised their firmware it stopped DMT displaying the SNR in some versions.
The gaps indicate that your router has deemed the SNR at that particular frequency too low to be able to transmit any data at some point. Normally you would expect the SNR and bit allocation to more or less mirror (with a caveat that it doesnt always - depends on how long the router has been up for).

Based on the assumption that your SNR and bit loading is mirrored, the the bit swapping process should have sufficient head room to keep your router alive.  The gaps do indicate though the frequencies where your connection suffers interference the worst from.
One or 2 gaps are sometimes unavoidable on some lines and the tail off at the end doesnt look too unusual so Im ignoring those. 
Tones around 200 are possibly the ones that are loosing you the most from your sync speed.. but its still a smallish amount. - off the top of my head based on your graph and very approx guess  - circa 200-250 kbps across all the gaps.

The gaps also show where to keep an eye on and where problems may most likely be occuring but for it to be more useful youd really need to see active monitoring of the SNR

eg Ive just fired DMT up now to quickly grab a screenie to show you how and where my own SNR(M) is currently falling.  DMT marks it with the little orange arrows.

If youre interested in understanding DMT and bit loading then theres more info (and graphs) on this page
http://www.kitz.co.uk/adsl/adsl_technology.htm
Title: Re: Line attenuation suddenly lower
Post by: Oranged on January 13, 2009, 08:11:27 PM
Thanks for that analysis Kitz.

I'd already read through that link and it's why I decided to post the graph and ask the question. It was the high end gaps that I was unsure of, as regards their importance. Also fortunately the number of gaps do not seem to increase at night.

I've also seen references to overhead phone lines (my BT service is overhead from a pole across the road) picking up interference. There are no local radio transmitters but there is a major hospital <1Km and a significant MoD establishment <4Km.
Title: Re: Line attenuation suddenly lower
Post by: kitz on January 13, 2009, 10:16:32 PM
>> It was the high end gaps that I was unsure of, as regards their importance.

Something you are most likely to see as the line gets longer.  The frequencies are higher and more susceptible.. and because of the way DMT works and because the SNR is naturally lower, these bins are most likely the ones that your router marks as unusable and bitswap to other subchannels.
Its not at all unusual on a line that is syncing at its maximum capabilities,  to see the spike-gap-spike-gap effect on a line thats been up for a while at the tail end frequencies.
A  total resync may cause some of these bins to be used again.. or it may not..depends on the SNR at time of initialisation... and the fairly complicated calculation that decides how many bits to allocate depending on the BER (now Im beginning to get out of my depth)..  but at those frequencies on the tail end you're only looking at a couple of bits per bin (which bit swap will have already allocated else where), so youre not likely to gain much at all, if anything.