cancel
Showing results for 
Search instead for 
Did you mean: 

Downstream SNR varying wildly

summers
Aspiring Pro
Posts: 275
Thanks: 50
Fixes: 1
Registered: ‎01-06-2014

Re: Downstream SNR varying wildly

@ejs: did you find a way of monitoring ppp status? Don't think I've quite got a way to pick up ppp drops, well
Quote
ip route

Probably gives most info, eg:
Quote
195.166.128.195 dev ppp0  proto kernel  scope link  src 146.90.63.166
192.168.2.0/24 dev br0  proto kernel  scope link  src 192.168.2.1
default via 195.166.128.195 dev ppp0

so I get the default gateway, as well as ppp being used for routing. But don't get to see LCP messages. I'd hoped that sending -USR1 to pppd would give more messages, but no cookies yet ...
Think the crux is that LCP messages withh only happen in the pppd process - so need a way of accessing those. In a read only file system it is hard to set up somehing like pppd, however the standard TP-Link http interface does give LCP messages, so there must be a way of accessing this info ...
ejs
Aspiring Hero
Posts: 5,442
Thanks: 631
Fixes: 25
Registered: ‎10-06-2010

Re: Downstream SNR varying wildly

I don't think it's really necessary to monitor or log the PPP messages. You could configure the TP-Link to send all its system log messages over the network to a something running a syslog server, then they'd be saved somewhere at least. Other than that, I don't know of anything else. I would have used the ifconfig command to record the amount of data transferred, but it was a bit odd in that the ifconfig command only ever showed a tiny amount of data, far less than how much was actually downloaded.
summers
Aspiring Pro
Posts: 275
Thanks: 50
Fixes: 1
Registered: ‎01-06-2014

Re: Downstream SNR varying wildly

Yes - ifconfig would just show data from the router itself, e.g. originated for finished in the router. IIRC it doesn't show forwarded traffic, e.g. most of the traffic from your home. Now forwarding happens in the kernel, but how to follow - I guess in one of the /proc or /sys files. Its probably written in the iptables documentation somewhere.
[Just checked, reading documentation, the command you want is:
Quote
iptables -L -v -n

I havn't checked this on the router yet - and as logging is currently active, I can't log on to check]
I run out of time to write the code to process the raw data from the router, so instead just dump everything, every 10s - then compress the log. Means I should now have a record of most things.
@CRT, can you reset the SNR margin on my line to 3db or 6db? I expect the line to during training to increase the SNR margin, but with logging switched on I should be able to dig into what is causing the margin to increase, e.g. if it is an Error Seconds thing or something ...
ejs
Aspiring Hero
Posts: 5,442
Thanks: 631
Fixes: 25
Registered: ‎10-06-2010

Re: Downstream SNR varying wildly

(Other routers will show the traffic totals in ifconfig, I think this one doesn't because it apparently has a Protocol Processor Engine to which it offloads some network tasks from the CPU. For iptables, I'd need to also check "iptables -t nat ...".)
summers
Aspiring Pro
Posts: 275
Thanks: 50
Fixes: 1
Registered: ‎01-06-2014

Re: Downstream SNR varying wildly

Interesting. Just trying to remember back when through work had to set up openwrt. If I remember correctly, things set up under linux, and forwarding packets were hidden. So many routers use linux (any why not L4), and it should be the same on all routers uning linux + iptables.
summers
Aspiring Pro
Posts: 275
Thanks: 50
Fixes: 1
Registered: ‎01-06-2014

Re: Downstream SNR varying wildly

Line hasn't yet had a SNR margin reset (hint to CRT), but number of error seconds in last few days:
24/3 17
25/3 17
26/3 17
So seem to be getting a regular 17 error seconds a day (which seems a bit of a coincidence). So is 17 ES a day an issue?
Anotherone
Champion
Posts: 19,107
Thanks: 457
Fixes: 21
Registered: ‎31-08-2007

Re: Downstream SNR varying wildly

There's no way 17 ES a day is an issue Smiley   but do expect the rate to be higher of course, with a lower SNRM ............ unless that count isn't ES - I'd be inclined to try and discover why that seems a "co-incidence". It could just be that they are being caused by regular timed events that repeat at the same times each day.
james1944
Newbie
Posts: 3
Registered: ‎16-01-2015

Re: Downstream SNR varying wildly

First of all, you are wasting your precious time posting statistics to this forum. Anotherone is a Government HACK and will be no support whatsoever. [Removed] believes in REIN/SHINE and tries to convince people that it exists. All that you will find using the REIN information is that every single electrical device produces noise, some louder than others. This NOISE is no more than the EMF produced by EVERY electrical appliance. Even if you are in a high RFI area all electrical devices are bound by strict regulation so no one device causes problems with another. All power supplies have ferrite core chokes to filter out undesirable NOISE to 50-60Hz. Does anyone understand!???. If not then you are a FOOL.
   ADSL Down stream frequencies are the same used for LW/MW Radio and are the ONLY problem in REALITY that most people have with long distance broadband. I have been fighting these people for THREE YEARS now. I pin pointed the WIRING FAULT three years ago and it was VERY VERY SIMPLE to find using portable radio. At the end of last year a representative of BT kept on phoning till I answered. She claimed to be a RF EXPERT previously working with the RAF. Her and two other engineers came to property around 8pm. I showed them the statistics that I had been collecting for multiple months, and they were able to witness the SNRM problem with their own eyes. I told them where the problem was and how I found it, most of what I said went in one ear and out the other. I told them the pole DP35 had the fault, they went to check came back and said there is a lot of noise at DP36 the next pole after. I questioned them about DP35, she grudgingly replied that there was a lot of noise there too. I tried to explain that it was obvious that if there was noise at 35 then of course there would be noise at 36. She left saying this, I will do everything in my power to fix the fault but there is POLITICS INVOLVED!!???.
   She emailed statistics that CLEARLY showed there was indeed a fault. The pair of wires were different lengths by about 100 metres!. They should know that this is an unbalanced pair!. I responded to email and within hours she turned up with an engineer who was trying to convince me the problem was between Rx and Tx, yes LW Atlantic 252KHz is a problem but the worst noise is MW Radio, put aside the local stations, you are not going to track down the FAULT tuning into them. LW can be heard clearly through daytime but at night it becomes fuzzy. MW characteristics are different especially at NIGHT late at night, when its DARK!. Multiple foreign radio stations can be tuned into with portable radio at the pole where the problem is. From 612KHz tune radio above and below this frequency to fine tune. Make sure radio antenna is close to pole with WIRES running down it to underground. Signal should disappear just a few feet away.
   The same engineer turned up at door an evening shortly after, again trying to convince me the problem was between Rx and Tx AGAIN!. He then went on to say the lady from RAF was only trying to make a name for herself as she was new to the job!. A sad excuse and more so and pig headed insult to the fairer sex. He then told a story of a customer who stuck a shovel through 100 pair of wires after openreach once again failed to help. He went on to say that this customer got away with this destruction as though he was trying to convince me to do the same. What does he take me for a moron!. He again promised that he would be back to fix the FAULT, never seen him again.
   The RAF RF expert then booked an appointment with an REIN specialist. He turned up a day late I personally missed visit but all he did was poke about with a spectrum analyser and antenna. He once again cleared the property from being source of noise, then visited neighbouring properties where he found that I wasn’t the only one suffering from broadband drop-outs. It took 2 MONTHS for a reply to his findings, and this is what he said. 1.1MHz = 1100KHz was the source frequency of REIN and that he couldn’t track it down!!. 1089KHz is Talksport, 1109KHz is MFR 2 and a French  station at 1098KHz all are in the mix. She replied saying that they had completed the work and once again avoided fixing the FAULT!. What I have learned over the last 3 years is, Openreach are 100% controlled by BT. BT walk hand in hand skipping down the street with WESTMINSTER POLITICANS and is run by them too. Its about time these morons, who are a cankerous growth on society, were radically reformed...  
[Moderator's note by Mike (Mav) Derogatory term used for forum member removed as per forum rule.
summers
Aspiring Pro
Posts: 275
Thanks: 50
Fixes: 1
Registered: ‎01-06-2014

Re: Downstream SNR varying wildly

Woah james,
Thats a bit of a rant and a half.
I work in an engeering company, and have a long histoty in physics/engineering/ and computers. So I have the background to have some knoweldge in this.
Plus.net on this formum keep having to defend, that the connection between the home and the DRM, is in in openreach/bt wholesales hands. Yes its the same company as plusnet, but due to regulation there are walls between the different parts. Now the DRM sets up the speed of the link in the last mile, it does this automatically, but can be tweaked by human hand.
In my case, its clear there is something in that last mile, that during training pushes the SNR margin to high levels. WHat that is hasn't been found yet, but plusnet are not in a possition to find it. Now I and openreach, probably are able to find the problem, given time. However I at least need a bit of help from plus.net - as they can tweak the DRM, in particular reseting the noise margin.
Now I on my part, have taken time and effort, to set up hardware on my end, to monitor the line. Specifically to monitor low level SNR against frequency, and bits per tone. This should mean, that with a fair bit of processing, should be able to track what frequencies, and what times, are giving me a problem. The yes, its walking the street  with a MW radio, to find the source. Then negotiation with BT to moan at the noise source. Its not an easy route.
But can you expect plus.net to do this? Lets face it, I pay them just a few pounds a month for the line - that necessarily means limited resources. So can I expect them to do more than is under there control? For me, I'd just be happy if the did the occasional reset of the SNR margin, and the prodding of bt openreach/wholesale - when fault found - this afterall is where my contract is with them for.
And thats the crux, whats in the contract? You can't expect plus.net to do more than the contract.
So stand back, take a deep breath; and relax. Then look at ways that are productive, what you can do, what will help, what is postive, and what is the best way forward.
Anotherone
Champion
Posts: 19,107
Thanks: 457
Fixes: 21
Registered: ‎31-08-2007

Re: Downstream SNR varying wildly

@summers
You are wasting your breath. As you know this forum is full of Plusnet customers who try and help each other out and get Plusnet to provide the support that is required to resolve problems AND to get the incompetently managed organisation called Openreach to fix some of the line issues that cause problems like the one james1944 is suffering. The last time james1944 posted on this forum - over a year ago - he just ranted on someone else's thread (eventually moved by a moderator - another Plusnet customer) saying his problem was the same as that poster's - I wasn't sure as I remarked his post wasn't well structured it was difficult to follow. He seemed to think that nobody else here understands REIN or RF Interference etc. and just chose to have a go at me when I tried to get some more detail. I'm certainly not a Government HACK as he puts it and certainly don't appreciate his Libel. He'd been sent a deadlock letter by his previous supplier and because of that Plusnet refused to address the problem. He did nothing but rant again and have a go at me seemingly thinking I was a BT engineer when he made a second post, and made no effort to post in a manner that would elicit help from other forum members to try and get his issue resolved. He obviously knows enough to solve his own problems, so good luck to him.
summers
Aspiring Pro
Posts: 275
Thanks: 50
Fixes: 1
Registered: ‎01-06-2014

Re: Downstream SNR varying wildly

Find attached error seconds accumaing vs days since the router was rebooted. Doesn't seem any structure to them. WiIl look at better ways to plot.
summers
Aspiring Pro
Posts: 275
Thanks: 50
Fixes: 1
Registered: ‎01-06-2014

Re: Downstream SNR varying wildly

Thanks Chris for resetting the SNRM. Now to follow carefully how the line copes.
Whats initially intersting, the the nES counter has gone a bit random:
nES=4294967289
Worked this out - nES reset to -7 in a 32bit number. Question is why did it reset to -7? Anyway after a handful of errors, it wrappen and went positive again ....
However is incrementing at a fairly slow rate.
Am logging everything right now, probably won't get a chance to look at the detail until the weekend.
summers
Aspiring Pro
Posts: 275
Thanks: 50
Fixes: 1
Registered: ‎01-06-2014

Re: Downstream SNR varying wildly

After reset at 8:22 GMT to reset the SNRM; line also decided to reconnect at 8:46. So thats intersting, only about 30 minutes up at 6db SNRM before the line went down. I'll dig into the logs later as to what took the line down ...... its looking like it was just a ppp reset - think the adsl stayed up, but reading raw logs is hard work ....
summers
Aspiring Pro
Posts: 275
Thanks: 50
Fixes: 1
Registered: ‎01-06-2014

Re: Downstream SNR varying wildly

So with a SNRM of 6db here is the SNR per tone, and bits *3.
Nice to see getting data across up to 350 tone. So I do benefir from ADSL2+ doubling the spectrum. Some signals in there are clear, Radio 4 at ~50. Need to identify the others, see if they are medium wave satations.
162 = 698kHz
212 = 914kHz
283 = 1220kHz = 1.22 MHz
Now its a case of waiting for the line to drop, then doing the graph again.
And not unexpectly the ES are up, 41 so far today.
ian007jen
Rising Star
Posts: 392
Thanks: 4
Fixes: 2
Registered: ‎06-09-2007

Re: Downstream SNR varying wildly

I wonder if the first drop was fastpath to interleave.......normally DLM's first action.
One way to check the above would be no FEC's between 8:22 and 8:46.
41 Error seconds is fine; my line averages 200/day with no problems.