Kitz Forum

Broadband Related => Router Monitoring Software => Topic started by: AlexS on December 05, 2010, 04:54:50 PM

Title: Analysing DMT data
Post by: AlexS on December 05, 2010, 04:54:50 PM
Hi, Could someone familiar with the DMT tool have a look at a screendump please ? ( attached file )
I believe I should be getting around 4mbps downsync on my line for the attenuation, ( 51db )
but I can't work out where I'm losing speed.
I've tested my line using the test socket,several different plugin filters and Speedtouch 546v6 router with similar results.
I normally use a Sky DG934G ( Netgear DG834Gv3 ) router with Sky LLU in gmt mode connected to NTE5 via Pressac filtered faceplate.
Phone line tested via NTE5 test socket with different wired phones / quiet line test, no interference heard.
Many thanks
Alex



[attachment deleted by admin]
Title: Re: Analysing DMT data
Post by: HPsauce on December 05, 2010, 06:04:15 PM
What sort of connection are you (supposed to be) on with which ISP? 800kbps upstream is a bit odd.....
Title: Re: Analysing DMT data
Post by: GunJack on December 05, 2010, 08:34:18 PM
sky LLU is normally ADSL2+, can you connect in that mode ??
Title: Re: Analysing DMT data
Post by: AlexS on December 05, 2010, 10:32:17 PM
Hi, I'm on Sky unlimited LLU which is up to 20mbps. Sky set the line to gmt mode as ADSL2+ was unstable at night.
Title: Re: Analysing DMT data
Post by: AlexS on December 08, 2010, 08:49:57 PM
Hi, I'm on Sky unlimited LLU which is up to 20mbps. Sky set the line to gmt mode as ADSL2+ was unstable at night.

Er, Sorry I meant G.DMT mode, not gmt ( that's Greenwich Mean Time  :-[ )
I seem to lacking "tones" at frequencies over about 500khz, but I don't know why.
A BT Openreach engineer checked the speed at the street cabinet last year and said it was about 4mbps, so I seem
to be losing almost 40% of the speed over a few hundred metres...
Is there anything obvious I'm missing from the DMT graphs ?
thanks
Alex
Title: Re: Analysing DMT data
Post by: GunJack on December 10, 2010, 08:16:15 PM
the bitloading looks well poor, do you have another router you could try ??
Title: Re: Analysing DMT data
Post by: AlexS on December 10, 2010, 09:34:39 PM
the bitloading looks well poor, do you have another router you could try ??

I was expecting a D-Link DSL-2640B being delivered today so I could try the connection with DMT Tool v8
but it looks like it will be Monday now...
I wonder if there is a similar utility for TI/Infineon chipsets so I could use my existing Sky router...
Title: Re: Analysing DMT data
Post by: GunJack on December 11, 2010, 08:54:45 AM
I use 2460B's - I've got three of them :)  

When you get the router, you'll need to check what firmware version it's on. DMT v8.07 works directly with v3.06 fw, but not with v4.01. If the router is on v4.01 then there is a tweaked v8 about that will work with it that gets over the logon issue. Routerstats Lite is the same, works with router fw v3.06, but not with v4.01. 4.01 works better with ADSL2+.

Gis a shout as to how you get on once you've got the nice D-Link in your mitts :)
Title: Re: Analysing DMT data
Post by: AlexS on December 11, 2010, 09:20:13 AM
I will do. Thanks.
Title: Re: Analysing DMT data
Post by: GunJack on December 13, 2010, 12:23:10 AM
Just noticed you've asked the same on TBB forums - good idea, more ideas, etc.

Anyhow, for your reference, your DMT "should" look similar to mine, similar attenuation etc, on ADSL2+.....see what happens with your 2640 when it gets there :)  note:- this shot was taken just past midnight....

[attachment deleted by admin]
Title: Re: Analysing DMT data
Post by: HPsauce on December 13, 2010, 09:15:03 AM
but I can't work out where I'm losing speed.
Well, the simple answer is everywhere, but especially on lower frequencies.

Not sure that helps a lot though as it doesn't suggest a cause....
Possibly interference from another ADSL line?

A longer-term monitoring session (days/weeks) with Routerstats may help, and of course try another router just in case.
Title: Re: Analysing DMT data
Post by: AlexS on December 13, 2010, 12:16:02 PM
Hopefully the DSL-2640B will arrive today and I can try it out soon.
It appears routers with different chipsets measure line attenuation differently.
A Billion 5200 (Trend) reports 53db.
The Speedtouch ST546v6 (Broadcom) reports it as 51db,
and the Sky DG934G (TI-AR7) is currently reporting 47db !
Title: Re: Analysing DMT data
Post by: waltergmw on December 13, 2010, 12:45:52 PM
Hi Alex,

The attenuation measurement is some sort of average over the tones (i.e. frequencies) available to the individual modems.
The actual attenuation differs at different frequencies.
E.g. a 2Wire modem diagnostics give an average (on a lousy line I'm monitoring) of 73.7 dB downstream and a specific attenuation @ 300 KHz of 75.4 dB

Kind regards,
Walter
Title: Re: Analysing DMT data
Post by: AlexS on December 13, 2010, 07:32:38 PM
I always thought line attentuation stayed pretty much static and was measured at a set frequency like
the 300Khz you mention, only the noise margin changed...
I was surprised the value changed so much between routers !
I've since learned the attenuation can change when changing modes ADSL > ADSL2+ for example
for the reason you mention.
Looks like my D-Link router is going to be delayed as it's coming from a company in Scotland with a
delivery backlog due to the recent weather.  :(
If I hadn't bricked my Netgear DG834v4 (Broadcom) during a firmware upgrade last year I'd already
have plenty of line data to work on...Grr.
Title: Re: Analysing DMT data
Post by: GunJack on December 14, 2010, 05:19:25 PM
Just to add to walter's attenuation observations...

on G.DMT with my 2640B, my attenuation was 46-47dB. On ADSL2+ it's 49dB.

Using a netgear dg834gt, G.DMT was 44dB, ADSL2+ was 47dB

..and they use the same chipset....
Title: Re: Analysing DMT data
Post by: kitz on December 17, 2010, 12:23:08 PM
>>> I was surprised the value changed so much between routers !

It is also important to note that different routers may load the frequency bins in slightly different ways, and on top of that some routers report the average across the frequencies actually in use, whilst some may report the average across the frequencies available - which is why sometimes using a different router may report a slightly different attenuation figures.

Read more: http://www.kitz.co.uk/adsl/linestats.htm#attenuation#ixzz18N3mTZEo
Title: Re: Analysing DMT data
Post by: AlexS on December 22, 2010, 01:49:56 PM
Well my Dlink 2640 arrived this morning. 2640B's appear to be as rare as hen's teeth ( all Ralink version in shops )
so I ordered the Dlink DKT-710 starter kit.
On Dlink's website it clearly states the DKT-710 kit contains the 2640B router.
I received the DKT-710 starter kit containing...    ...the 2640R version !!! You couldn't make it up !
As it happens I'm expecting another DKT-710 kit from another supplier ( ordered as backup ) next week.
If that doesn't have a Broadcom chip I'll have to think of something else.
I have heard from neighbours at opposite ends of my road that they are also having speed problems
and were're all connected to the same telephone pole.
Would a corroded connection somewhere effect lower frequencies or just higher ?
If it's a form of REIN interference causing it, it's definately on 24/7...
Any suggestions welcome...  :(
Title: Re: Analysing DMT data
Post by: GunJack on December 22, 2010, 05:11:51 PM
2640R is carp - won't work with DMT or routerstats lite either, send it back  >:(

2640B can usually be had on ebay, if you're really lucky from only a few quid ;)

...and if still no joy I could sell you one of mine....
Title: Re: Analysing DMT data
Post by: AlexS on December 22, 2010, 06:37:15 PM
2640R is carp - won't work with DMT or routerstats lite either, send it back  >:(

2640B can usually be had on ebay, if you're really lucky from only a few quid ;)

...and if still no joy I could sell you one of mine....

Thanks for the offer. They'll take the Ralink version back as it said 2640R on the box label
 so I didn't have to open it to find out.  ;D
I'll see what the replacement has and take things from there.
Do you know of any utils like the DMT tool for other chipsets ? It does appear to be more
comprehensive than most.
Title: Re: Analysing DMT data
Post by: GunJack on December 22, 2010, 09:01:05 PM
dunno of any other tools other than DMT and RS, anyone got any ideas on this one  :no:
Title: Re: Analysing DMT data
Post by: AlexS on December 24, 2010, 02:52:20 PM
2nd router arrived today. This one is a 2640B, firmware v3.06 .
Ran DMT tool v8 for about 10mins and I've attached the output, but I'm none the wiser.  ???
thanks
Alex


[attachment deleted by admin]
Title: Re: Analysing DMT data
Post by: GunJack on December 26, 2010, 11:30:49 AM
the bitloading looks like it's suffering from sinusoidal interference, i.e. noise, but not ramdom bursts....
Title: Re: Analysing DMT data
Post by: AlexS on December 27, 2010, 02:41:12 PM
Yes, it's happening 24/7 every day.
I'm not getting noise using a corded phone or an AM radio tuned to 612khz so it's going to be a pain to track down.
If I can locate a frequency or range where the bits are being lost, that would assist in an attempt to locate the source.
It's probably something like an old freezer as the cable from my phone pole attaches to the next very close to a couple of houses
on the next road.
Could a "split pair" on the cable cause such interference ?
Title: Re: Analysing DMT data
Post by: GunJack on December 27, 2010, 03:05:49 PM
I'm beginning to think it's a constant source of REIN :(