Print

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

[DeanB]

Orainsear - found that TI document after I'd posted really good....  Not LOw but L0!  The document has some interesting comments on xtalk....

After discussing the L0/1/2 states, it goes on to say that this is not necessarily the answer to controlling power consumption as it causes time varying cross talk.  Then goes on to say power cut back "In some cases, the service provider may prefer to manually set the value of power cutback. For example, the service provider may have performed an analysis through DELT and determined the best transmitted PSD level for the given link. This programming can be made independently for the downstream and the upstream transmission directions.

Does anyone know what DELT is?  Sounds like the service provider (Sky) can manually set the power cutback, which as Kitz says will mean, less power being sent down the line to me, leading to lower signal to the noise and hence bit loading.  As my DMT graph shows, I have the tones low from the start of the downstream and essentially flat.

Result - Noise across the spectrum is reducing bit loading
*Proposed cause 1 = Xtalk on the way or in the exchange, but 5Mbps seems excessive if it's part of acceptable design
*Proposed cause 2 = powercut back - by the DSLAM automatically (based on attenuation?) or manual and needs adjusting.....

Next steps
Get a modem that can show me if power cutback has been applied to the downstream - which I think means I need to borrow a Broadcom based unit

Also realised I can see what the DSLAM venor is ATUC Vendor Code:       4753504E  == GSPN  == Lucent Globespan/Conexant

Kitz & HP -really useful DMT graphs thanks, explanation all makes sense. 

[orainsear]

Does anyone know what DELT is?

Dual End Loop Test - a test with a modem installed at the end of the loop (your property), it's more accurate and refined than:

Single End Loop Test - test with no equipment needed to be connected at the property.

Explanation here.

[DeanB]

Now using the DG834GT with DGTeam firmware to provide more data.

Line sync and noise identical to DG834v3
This one taken at 22:37, 22:38, 22:39 - notice the DS power level fluctuates 0, 1.6 8.0 and finally routerstats clearer version.

I'd appreciate views on the power levels and any further information that might be useful that I can copy up!


[attachment deleted by admin]

[attachment deleted by admin]

[GunJack]

Just noticed something on the DMT graphs, your interleaving depth is being reported as 160 !! Is that normal for ADSL2+ ??

[jeffbb]

Hi
quote  : just noticed something on the DMT graphs, your interleaving depth is being reported as 160 !! Is that normal for ADSL2+ ??

I have seen this before on another forum where a user was having trouble  . Then it seemed odd .

see link below
http://forum.kitz.co.uk/index.php?topic=5154.0

Regards Jeff

[DeanB]

Jeff & Gunjack - re-sync today, brought downstream interleave to 96 from 160 marginal ping time improvement.  bbc.co.uk = 23mS

Does anyone else have such low downstream power ? maximum 8dBm, does this indicate power cut back or a fault....

[jeffbb]

HI
It seems as interleave depth  is something like 8,16,32,64,96,160 
Regards Jeff

[DeanB]

... make that and 288.

A few more investigations carried out.

Go back to the offical GT firmware to see if this changes anything
DS-Sync 13311, Interleave 96 RCO at 100%
I don't think this is much more than Re-sync changes

Set target SNR 4.5dB
-expected result higher sync rate
-actual result, same sync rate, higher interleave 288! and poor error seconds, RCO=16200kbps
What does this say? that something is holding the sync rate back

Meaure ds-Power out over an hour
This graph is all over the place and I can' believe it as noise doesn't follow it.  I'm inclined to believe that there is a bug here between DSLAM and router, where the DS value is not being sent or read back correctly.  Does anyone have this combination that gives a sensible value? ie DG834GT and Sky-GSPN DSLAM

Finally did try this analyser from Charles Fryett http://adsl.blueether.net/stats.php  which is very smart, but I don't think gets me any further.
You are on ADSL2+
Down sync is low (68% of expected max.).
SNR is good.
Attenuation looks normal.
The power values are out.
There are a few uncorrected FEC errors.

It looks like you are using a an unknown modem with adslctl to connect to (Unknown)'s network via Cheltenham (GB).
Overall your line looks OK. Line has a score of: 2
A lower score is better.

Your expected max sync depends on many things, mostly the quality of the line and the target Signal to Noise margin at the time of sync.
Expected max syncs are:
3db Fastpath: 22,057
6db Fastpath or 3db Interleaved: 20,436
6db Interleaved: 19,549

You appear to be on the standard Signal to Noise margin and with Interleaved set.
Your sync is 68% of the estimated maximun for a line with Interleaved and standard SNR at your line length (1,665 meters).
As your sync is not above 85% of the estimated maximun you might wish to investigate a potential problem.

The output power for either the DSLAM or your modem indicate that there is an issue somewhere.
This is usually a hardware problem rather than a line issue. Ignore this if you are synced close to the max sync rates as less power is needed if you are close to the exchange.

If you set the power in the text to 19dBm it reports power OK and just reports the sync as the issue as expected

What's the next step?  Relax and leave it, or is there anything else I can do or ask?


[attachment deleted by admin]

[attachment deleted by admin]