Kitz ADSL Broadband Information
adsl spacer  
Support this site
Home Broadband ISPs Tech Routers Wiki Forum
 
     
   Compare ISP   Rate your ISP
   Glossary   Glossary
 
Please login or register.

Login with username, password and session length
Advanced search  

News:

Author Topic: Could i push a long line any further.  (Read 7022 times)

dizzy4528

  • Member
  • **
  • Posts: 54
Could i push a long line any further.
« on: November 18, 2008, 09:06:36 PM »


With the below stats ,do you think i could push the sync speed and up my profile without any instability.
The drop off of LOS occured when i flashed the netgear DG834GT with the DGTeam firmware and pushed the SNR to 50%
The Noise level runs about 9db constant through the day and never varies by more than 1db .
I am nearly 6Km from the exchange.


/usr/sbin/adslctl: ADSL driver and PHY status
Status: ShowtimeRetrain Reason:   0
Channel: INTR, Upstream rate = 448 Kbps, Downstream rate = 1792 Kbps
Link Power State: L0
Mode:         G.DMT
Channel:      Interleave
Trellis:      OFF
Line Status:      No Defect
Training Status:   Showtime
           Down              Up
SNR (dB):     8.8            15.0
Attn(dB):           63.5             31.5
Pwr(dBm):   16.4             11.9
Max(Kbps):   1952      824
Rate (Kbps):   1792      448


Previous 1 day time = 24 hours 0 sec
SF  = 5082378
CRC = 352
LOS = 0
LOF = 0
ES  = 174


Date            Profile   Speed   SF Errors   CRC LOS   LOF   ES
11/11/2008      1mb   1568   5082380   262   22   0     197
12/11/2008   1.2mb   1568   5082381   373   25   0     213
13/11/2008   1.2mb   1568   5082321   321   22   0     214
14/11/2008   1.2mb   1568   5082321   308   20   0     201
15/11/2008   1.2mb   1792   5075767   253     0   0     140
16/11/2008   1.2mb   1792   5075413   407     0   0     217
17/11/2008   1.2mb   1792   5076183   289     0   0     183
18/11/2008   1.2mb   1792   5082378   352     0   0     174
Logged

kitz

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 33879
  • Trinity: Most guys do.
    • http://www.kitz.co.uk
Re: Could i push a long line any further.
« Reply #1 on: November 19, 2008, 12:42:53 AM »

The error rate looks fine :)

If as you say the SNRM doesnt normally fluctuate more than 1dB, then you could always try a little bit more. 
Dont do a big jump though just in case... and do take it slow.
Your line is very long and highly attenuated, so drop off could be quite sudden. 

I suppose the worst that could happen is that you loose sync and resync at a lower speed which could affect the DLM, but since youre getting higher sync speed than previous then not too much harm done.
Logged
Please do not PM me with queries for broadband help as I may not be able to respond.
-----
How to get your router line stats :: ADSL Exchange Checker

dizzy4528

  • Member
  • **
  • Posts: 54
Re: Could i push a long line any further.
« Reply #2 on: November 19, 2008, 05:16:22 AM »

 :fingers:    Cheers, will give it a go   :fingers:
Logged

Weaver

  • Senior Kitizen
  • ******
  • Posts: 11459
  • Retd s/w dev; A&A; 4x7km ADSL2 lines; Firebrick
Re: Could i push a long line any further.
« Reply #3 on: January 14, 2009, 09:16:34 PM »

I would suggest that there _are_ quite a few things that are worth trying.
I myself have a similar very long line. I have put a lot of work into testing noise-lowering measures recently with good results.

A few points,

i) Is the -GT a Broadcom-based device?

ii) consider non-wireless routers (and simply pick up a better access point separately - I recommend the Netgear WN820T or better the newer 5GHz device [yet to be tested[) in case that allows you to get a quieter noise environment

ii) consider the DG834v3 (non-wireless) as I get the impression that this out-performs the DG834v4 which is Broadcom-based on very long lines, but this is yet to be confirmed as I need to do much more testing.

iii) the 63.5dB line attenuation reading is what I get too from a DG834v4; the figure varies between router models DG834v1 reports 59dB, v3=61dB, v4=63.5dB. If this is really the picture that the v4 Broadcom device 'sees' then it must be really suffering. Or maybe it's just an artifact of the way it calculates line attn.

iv) experiments suggest that it is well worth implementing a range of RFI and noise reducing measures; and of course you have got rid of all your extensions and had a faceplate-type microfilter fitted to your now one-and-only NTE5 master socket?

v) compare the Adsl Nation XTE-2005 with the 'Pressac' NTE-2005 as I'm of the opinion that sometimes the latter may outperform the former.

vi) use short wallsocket-to-router cables, which IMO should possibly be of the UNshielded twisted pair RJ11-RJ11 as my experience has suggested that the foil shielded type can sometimes be slightly worse, possibly increases attenuation? capacitance? (And what about the lack of grounding of the shield?

vii) Particularly if you have a cordless, phone retest _during a call_; my DECT phone worsened SNR by 2dB but only when a call was being made (compared with turning it off/unpluging it); the phone cable seemed to be acting as a transmitting aerial and so changing the cable to a shielded RJ11-to-BT cable helped enormously, possibly using a ferrite coke in it would do too, probably should put a ferrite on its power cable as well.

I did all of the above (and a lot of other anti-RFI and mains noise measures too) and managed to just about get from the 1500 IP profile sync range up to just about scraping a sync rate of 2272 at 5.5dB-ish which was held up for four days. Anyway, the sync range for the 1750 IP profile should be easily achievable even with the deterioration in the weather.
Logged

Weaver

  • Senior Kitizen
  • ******
  • Posts: 11459
  • Retd s/w dev; A&A; 4x7km ADSL2 lines; Firebrick
Re: Could i push a long line any further.
« Reply #4 on: January 19, 2009, 01:58:20 PM »

This was a while back now, but I've only just noticed that your from original stats it says

   Trellis: OFF

This is a disastrously bad thing is it not? (Especially given the ultra-high line attenuation.)

If its true that it really is off, is this a defect in the firmware you chose? Or something wrong at the DSLAM end?

Logged

kitz

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 33879
  • Trinity: Most guys do.
    • http://www.kitz.co.uk
Re: Could i push a long line any further.
« Reply #5 on: January 19, 2009, 04:31:18 PM »


   Trellis: OFF

This is a disastrously bad thing is it not? (Especially given the ultra-high line attenuation.)


Agree that Trellis should be ON.  Even non-interleaved lines will most often have trellis on.
There are actually 2 types of error correction - Convolutional coding and Block coding.  Both work differently.

Convolutional code (such as Trellis - TCM )  carries very little overhead and is good for correcting background 'white' noise.
Block coding (such as Reed Soloman - RS) carries a much larger overhead and is good for correcting burst type noise.

TCM occurs at a different layer to RS - TCM being an inner code, wrapped by RS.  Interleaving can also be applied to RS code.
It just wouldn't make little sense for RS code to be applied but not TCM.
 
Ive seen what you mention many many times in the past and some routers showing trellis being off despite interleaving being applied... 
and although I dont know for sure, its been my assumption that perhaps TCM isn't being detected by the router due to RS being used as the outer code.


For more info see http://www.kitz.co.uk/adsl/error_correction.htm
in particular

Quote

Convolutional coding is efficient at correcting random or white noise errors with little overhead, but doesn't fair well if a longer noise burst disrupts the sequence making the path impossible to recreate.
(See Trellis Coded Modulation)

Block coding is better suited for burst errors as it utilises more parity bits, but the downside of block coding is the redundancy increases bandwidth and delay. (See Reed Solomon encoding)

Both of these methods work independently but can be combined so that any burst errors resulting from convolutional coding may be caught by the block coding.
Concatenating these two different coding methods gives a high degree of correction. It's usual for the convolutional code such as TCM to be the "Inner code" whilst a block code such as RS is classed as the "Outer Code".


Also applying Interleaving to the block code, in order to scatter the errors from noise bursts, gives near optimum performance.


Logged
Please do not PM me with queries for broadband help as I may not be able to respond.
-----
How to get your router line stats :: ADSL Exchange Checker

dizzy4528

  • Member
  • **
  • Posts: 54
Re: Could i push a long line any further.
« Reply #6 on: January 25, 2009, 08:54:16 PM »

Just reading back through a few old posts and have noticed the comment about the trellis been off.
I have looked through the DGTeam firmware GUI and notice, i have the option to turn trellis on . Would this be ok to do at my end
and what effect would it have on my connection , speed and errors etc.
Logged

dizzy4528

  • Member
  • **
  • Posts: 54
Re: Could i push a long line any further.
« Reply #7 on: January 25, 2009, 09:35:27 PM »

Have turned Trellis code ON  from the default value
 (dont know what the default value was or is)
in the router menu but
it is still showing as OFF on the Netgear stats page .
Do you think that is in fact ON and as you said is not reporting
 the fact due to the RS error correction been applied.


Channel: INTR, Upstream rate = 448 Kbps, Downstream rate = 1824 Kbps
Link Power State: L0
Mode:         G.DMT
Channel:      Interleave
Trellis:      OFF
Line Status:      No Defect
Training Status:   Showtime
      Down      Up
SNR (dB):   8.7      15.0
Attn(dB):   63.0      31.5
Pwr(dBm):   16.3      11.9
Max(Kbps):   1984      836
Rate (Kbps):   1824      448
Logged

roseway

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 43467
  • Penguins CAN fly
    • DSLstats
Re: Could i push a long line any further.
« Reply #8 on: January 25, 2009, 10:35:24 PM »

Quote
Do you think that is in fact ON and as you said is not reporting
 the fact due to the RS error correction been applied.

I think that probably is the case. I have the same experience with DG834GT and the DGTeam firmware.
Logged
  Eric

kitz

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 33879
  • Trinity: Most guys do.
    • http://www.kitz.co.uk
Re: Could i push a long line any further.
« Reply #9 on: January 26, 2009, 01:30:48 PM »

TCM adds so little overhead  that you wouldnt even notice it - so it is certainly beneficial, and afaik its the default on all routers/dslams.

As regards to why some routers dont report it - I've never seen anything anywhere of why this happens.
I must stress that this is only my conclusion...  but the logical assumption is that some routers don't pick it up because Trellis encoding is done at a different layer and 'wrapped' by RS encoding...  and therefore just doesnt 'see it' because of the outer RS coding.

Thanks for actually trying it out - which tends adds further credence to the above theory.



Logged
Please do not PM me with queries for broadband help as I may not be able to respond.
-----
How to get your router line stats :: ADSL Exchange Checker

dizzy4528

  • Member
  • **
  • Posts: 54
Re: Could i push a long line any further.
« Reply #10 on: January 26, 2009, 06:40:09 PM »

Dont know if the info is of any use , but when i turned the Trellis code on via the
netgear GUI and  saved the change . The RCO% changed from showing a max of 2016kbps down to
a max value of 1984 kbps even though the sync speed stayed the same at 1856kbps and the router did not resync. :hmm:
Logged

kitz

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 33879
  • Trinity: Most guys do.
    • http://www.kitz.co.uk
Re: Could i push a long line any further.
« Reply #11 on: January 26, 2009, 06:48:00 PM »

Thanks for that...  I'm  not sure how exactly it would have changed things - perhaps with trellis the line is considered more stable it can sustain a slightly higher sync speed when calculating the BER. 
Unlike RS which adds a lot more overheads, therefore an allowance has to be made slightly in the opposite direction.

- guessing because afaik theres very little info about this particular subject.
Logged
Please do not PM me with queries for broadband help as I may not be able to respond.
-----
How to get your router line stats :: ADSL Exchange Checker
 

anything