Print

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

[michty_me]

Good Afternoon all,

I'm guessing this may be down to cross talk but over the past month or so I've lost approx 6-7mbps sync speed. One day my downstream started reporting between 4.9-5.3db but remained having sync at just shy of 74mbps. I updated the firmware of my D7000 and re-setup my connection where my sync rate dropped to 67mbps and SNR raised to 6.5db which I expected due to the low SNR reported prior to the update.
I've also noticed that I am now getting 8 times the amount of FEC's recorded. I gather this is G.Inp doing its job? At first I thought it could have been down to some neighbours and my own Christmas decor but now that is removed, I'm unsure.

Also, If someone would care to explain the hlog and any other valuable graphs on MDWS to aid in determining line stability and condition, I would be forever greatful. I've been asked to look at others, Seen people commenting on other threads about them but I have not much idea on how to understand these correctly, So some pointers would be great.

[ktz392837]

This is very similar to me.  I had a 80/20 connection with db to spare but now I am down to 60/20.  Thankfully I still have my 20 upload with still a bit of db to spare.  Saying that I hope my download doesn't drop any further.  I am probably running at 40% loss of download due to crosstalk - hopefully that is it now.  On an eci cab so no hope of ginp/vectoring so cross fingers I do not have any interference issues.  I think a lower target of 6db perhaps 3db would be advantagious as apart from crosstalk I seem to hold the dbs between 5.5db and 6.5db.

[michty_me]

Its very strange. I haven't noticed the BT vans out to my cab in a while either. Until today that is but no further change to stats. I was thinking that maybe over the christmas period the errors had increased so DLM had dropped my speed to maintain a suitable dB level and cut errors. I would have thought my G.INP level would have increased from its level of 8 first though. Who knows. Thankfully I'm on a Huawei cab so I have the advantage of the G.INP feature.
It is out of my control anyway but I would still be interested in knowing more about fault finding/pin pointing issues through using MDWS graphs.

[NewtronStar]

The Kitz forum is full of information regarding how to interpret your line stats the graphs are different but once you compare like for like which is difficult as no two lines are the same you start to see similarities and from here then you can start to judge how well this line is performing.

As I said above no two line will be the same there is many variables but you can if you look hard and have previous stats/graphs start to see evidence of line deaggregation over time.

Maybe the SNRM has fallen and after a re-sync you notice a lower Attainable sync that sounds like crosstalk

You may see more spikes on certain tones on your QLN graphs this sounds like extra RFI being picked up on your line and could be the start of a small HR fault or just more interference from other devices in the home.

Or say an Hlog Graph shows up a possible bridge tap you can't learn this stuff in a week or even after 3 years but will have a better idea what to look for when you see others graphs.

[michty_me]

I guess like you say, Just get stuck into the available information on here. Thanks for some pointers too NS.
I'm guessing it is down to cross talk but time will tell I guess. It may increase unexpectedly again like what happened in the past with my upstream. It was sitting at 16mbps with a SNR of 6.5db then one day it decided to reboot with the same SNR with a full sync.

[ktz392837]

You may see more spikes on certain tones on your QLN graphs this sounds like extra RFI being picked up on your line and could be the start of a small HR fault or just more interference from other devices in the home.

Are there any examples of this?  A good graph and bad graph?

Or say an Hlog Graph shows up a possible bridge tap you can't learn this stuff in a week or even after 3 years but will have a better idea what to look for when you see others graphs.

[michty_me]

Yeah some examples would be good to see if you have any at hand or can point to a thread NS.

[burakkucat]

For the effect of a bridging tap of moderate length and the visual signs in an Hlog plot, please have a read through gazzi's forum thread from last December (2015), "About my possible bridge tap issue".

[michty_me]

Thanks burakkucat. Will give it a read tonight.

[burakkucat]

Thanks burakkucat. Will give it a read tonight.

When you do so, please be aware that someone was caterwauling nonsense during the first attempt to quantify the tap's length.    With WWW's assistance (and re-reading what I had assumed I knew but had actually mis-remembered), we managed to calculate a good approximation of the length. 

[michty_me]

Will do. Thanks. Never got a chance last night but seeing as the weather is miserable again, I'll get stuck in over the weekend.

[burakkucat]

As for a QLN plot, consider it to be a snapshot of the entire frequency band in use by the xDSL service.

Most circuits in the UK will show a peak corresponding to the 198 kHz signal broadcast from the Droitwich transmitter, whilst circuits far north of the border into Scotland will show an identical peak originating from the Burghead and the Westerglen transmitters.

In the past, some of us have made attempts to geo-locate fellow Kitizens by cross-referencing the peaks originating from licensed broadcast transmissions. In my own case, it can be quite clearly seen that The Cattery is located in East Anglia.

Radio frequency ingress is inversely proportional to the AC balance of an xDSL circuit, amongst many other things.

An ideal QLN plot would be fairly flat, with a noise floor around -140 dBm/Hz. However that ideal is very rarely seen . . .

As examples of QLN plots that are "real world" but away from the ideal -- with the xDSL circuits operating quite normally -- take a look at the graphs from N*Star and tommy45, attached below.

[michty_me]

It would seem I still have a lot to learn anyway. On MDWS and Hlog, Is the green and red lines showing D/S and U/S I presume?

My QLN graphs appear to have gaps in them compared to others. What does that mean? Its on MDWS under michty

[burakkucat]

DS and US. (There should be a legend, somewhere.)

If you are seeing missing blocks in the plots it will be for the US sub-carriers (tones or frequencies), as not all devices provide all of the required information. (Tony does display a note, regarding that, on MDWS.)

[tbailey2]

DS and US. (There should be a legend, somewhere.)

If you are seeing missing blocks in the plots it will be for the [color]US[/color] sub-carriers (tones or frequencies), as not all devices provide all of the required information. (Tony does display a note, regarding that, on MDWS.)

The legends for SNR/Tone, QLN and HLog were specifically turned off for reasons that escape me - they are now on but will not display for graphs smaller than medium