Print

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

[stumpy 101]

Hi to all

After some advise on my VDSL connection, does this look acceptable for a 900 meter line. or is there a underlying problem. sync and ip profile did drop to 9 meg and was told that was all the line could handle due to the length. An accidental mistake by the engineer reporting the line as under preforming triggered a drop and swop, surprise surprise current figures.  My concern is with that amount off errors and the interleaving being so high will this cause DLM to throttle my connection once again. I was estimate 27 meg to start, the engineer said that must of been an error.
 Regards 

[stumpy 101]

Hi all

 thought i would add this as it may help.

[waltergmw]

Hi Stumpy,

It's certainly not a brilliant line for 900 m if that is the actual distance.
You'll also note that you have had the standard 20 Mbps cap applied.
I'll leave B_E1 and BKK to add their thoughts too.

Kind regards,
Walter

[burakkucat]

Thank you, Walter. 

The only comment I can really make at this time is that the performance is definitely limited by the D-side metallic path length. The VDSL2 service on that pair does not seem to be afflicted with environmental RFI and the like. Overall, a nice but longish D-side pair.

I think Bald_Eagle1 will be able to give you a more thorough analysis, as if anyone can read the graphs 'at a glance' it is the originator of the code that creates them! At 900 metres in length, your D-side is fairly close to that which serves the Eagle's Nest -- so some comparisons will no doubt be forthcoming. 

[stumpy 101]

Thanks for your feed back so far.
My fear have come true DLM has reduced my ip profile down to 12 meg, when will BT listen and believe that there is an issue with the line 

[Bald_Eagle1]

Hi to all

After some advise on my VDSL connection, does this look acceptable for a 900 meter line. or is there a underlying problem. sync and ip profile did drop to 9 meg and was told that was all the line could handle due to the length.

When my similar length connection was physically faulty, I was also told not to expect any more than 15Mb from my connection, due to distance from the cabinet.

It took almost 11 months to get it finally fixed. Without any graphs/stats I'm convinced it still wouldn't have been fixed.

An accidental mistake by the engineer reporting the line as under preforming triggered a drop and swop, surprise surprise current figures.  My concern is with that amount off errors and the interleaving being so high will this cause DLM to throttle my connection once again. I was estimate 27 meg to start, the engineer said that must of been an error.
 

My estimate started of at 14.6Mb on the old 8c profile & is now 30Mb using the 17a profile.
I did achieve around 35Mb sync speed for the first month & now it is around 25Mb to 30Mb depending on what time of day I force a resync.

Things that have lowered my sync speed over time seem to be increased connection take-up from my cabinet, causing some crosstalk (I believe I was possibly the first to be connected to my cabinet via Plusnet, as at the time BT were reporting that it wouldn't even go live for FTTC within the next 6 months.

Additionally, as we simply cannot max out a 40 Mb connection, very slight things can have an effect as we have no spare margin to play with.

I experimented with Plusnet using different upstream profiles & we proved that downstream sync speed was always higher with a 2Mb upstream cap than a 10Mb upstream cap of which I can only achieve around 5Mb.


It should be noted that the FEC & CRC error graphs are incorrect.
For the harvesting/graphing version of the scripts that you are using, that data was taken from the modem's GUI.
There is a now known bug in the modem's GUI where it gets those stats wrong.

You should concentrate on the RSCorr, RSUnCorr, HEC & OHFErr graphs.

Some of those look quite high (in bursts), particularly RSCorr.

Your Interleaving depth is also quite high.
That may be as a result of previous problems where you achieved around 9Mb speed & your SNRM was decidedly flaky.

With Interleaving completely OFF, you wouldn't see any RSCorr errors at all.
Interleaving depth may settle down in time, but if your connection continues to see reasonably high error counts it is applied at varying depths to induce some stability & correct errors as they occur (RSCorr & FEC are really basicallythe same thing), rather than having to re-transmit the data.

My Interleaving depth was around 1700 before the physical external cabling/joint problems were finally fixed.
These days it settles at around 400 to 450.

I have attached the latest 6 days of graphs from my connection for comparison against yours.

Note: the additional graphs with the FEC/CRC modem bug eliminated will be available for all quite soon.
They are now generated by compiled programs rather than batch file scripts, are more robust & much quicker to harvest the stats & generate the graphs.

I would keep an eye on your connection for the next couple of weeks, keeping any forced MODEM resyncs to an absolute minimum to allow DLM (hopefully) to see some sustained improved stability & MAYBE allowing the 20MB banded cap to be increased.

Overall though, your graphs suggest there wouldn't be a great speed improvement.

Finally, we can see from your pbParams band plans that you are connected to an ECI cabinet DSLAM.
My connection is to a Huawei DSLAM.

We have occasionally seen higher RSCorr errors when Huawei HG612 modems are connected to ECI DSLAMS, but probably not enough to actually cause any problems.

As the ECI modem is harder to unlock & obtain stats from, I suggest you keep the HG612 connected (even if you actually have an ECI modem), in order to monitor longer term for any improvement/deterioration.

Who is your ISP?
We have seen that BT seem to be rather slow to deal with FTTC problems/are in denial that Infinity connections ever under-perform, whereas ISPs such as Plusnet & Zen do seem to be able to problem chase more efficiently with rather better informed customer service staff.

[Bald_Eagle1]

Thanks for your feed back so far.
My fear have come true DLM has reduced my ip profile down to 12 meg, when will BT listen and believe that there is an issue with the line

Can we see some more ongoing stats?
It may have been a huge error count/physical problem that has caused a lowered IP Profile.

Has DLM ever been reset back to a wide open profile on your connection, via an engineer's request?

[stumpy 101]

More stats from my line before engineer visit this morning.
 the spike in my hlog graph is this any think to be concerned about as it does change if you look at previous hlog. Stats a bit limited due to PC being switched off had visitors staying.
Will let you know how 4th openreach visit goes. BT  forum modulator has been pushing for a TPM figures crossed

Regards
   

[Bald_Eagle1]

Hlog shows attenuation over frequency.

It SHOULD be a smooth curve with no upward or downward spikes.
So, it does suggest a potential physical "issue" - maybe a bad external cable joint or a bridged tap due to the way things are wired up (possibly internally).

Are these stats obtained with any internal wiring disconnected?

I would let the engineer investigate, but if he says everything is O.K. I would them show him the graphs & stats, explaining that SNRM is high enough for much higher sync speeds and that the Hlog graph suggests a physical problem.

One way for an engineer to spot "faults" that may not be apparent using BT's standard tests would be to conduct a TDR test.
That should identify any poor joints etc. right back to the cabinet, with an indication of the distance where the faults occur, either from your home or from the cabinet end.

I believe a TDR test can be conducted using a typical engineer's JDSU or he may even possess a 'Hawk' or 'Mole' TDR tester.

If any faults are found & rectified, ask/insist that the enginerr requests a DLM reset or you may end up with a repaired connection, but still capped/banded at low speed by DLM.

Seeing that your comments suggest you are a BT Infinity user, you may struggle to get DLM reset.
They seem to point-blank refuse for some connections even though connection stats & stability have been proven (as WalterGMW would testify).

Good luck for today.

BTW, what is a TPM that you mention?

[Black Sheep]

Hlog shows attenuation over frequency.

It SHOULD be a smooth curve with no upward or downward spikes.
So, it does suggest a potential physical "issue" - maybe a bad external cable joint or a bridged tap due to the way things are wired up (possibly internally).

Are these stats obtained with any internal wiring disconnected?

I would let the engineer investigate, but if he says everything is O.K. I would them show him the graphs & stats, explaining that SNRM is high enough for much higher sync speeds and that the Hlog graph suggests a physical problem.

One way for an engineer to spot "faults" that may not be apparent using BT's standard tests would be to conduct a TDR test.
That should identify any poor joints etc. right back to the cabinet, with an indication of the distance where the faults occur, either from your home or from the cabinet end.

I believe a TDR test can be conducted using a typical engineer's JDSU or he may even possess a 'Hawk' or 'Mole' TDR tester.

If any faults are found & rectified, ask/insist that the enginerr requests a DLM reset or you may end up with a repaired connection, but still capped/banded at low speed by DLM.

Seeing that your comments suggest you are a BT Infinity user, you may struggle to get DLM reset.
They seem to point-blank refuse for some connections even though connection stats & stability have been proven (as WalterGMW would testify).

Good luck for today.

BTW, what is a TPM that you mention?

Just for info ........... there would not be a problem with the engineer requesting a reset, as we go directly through to the NGA Helpdesk who monitor the circuits. Usually, they would want to hear a reason as to why the DLM should be rest, such as an actual fault being found and rectified, but it should never pose a problem in getting the necessary carried out.
 

[stumpy 101]

Hi to all

An update on today events,
 Very helpful engineer came around and agreed that the 11 meg was a bit slow. he carried out all the usual test and there were a few fails but he said their were so close to being a pass he didn't think it was a problem. I believe he then did the TDR test as he showed me a graph on his unit that had a big spike within 5 meters which went when he removed the filter he said that was normal, there was a small one around 120 meter which when he look on his laptop was between my house and the pillar. He confirmed my line length as 900 meter and decided to swop my pair for another stats of which are below.
Very happy so far.

@ bald eagle 1
 A TPM is a tie pair modificatioin.

Slight downside the engineer told me that his manager had told him to fob me off will the long line spill 

[Bald_Eagle1]

Sync speed looks a shade on the low side for 900m, but stil quite a large improvement.
Did the engineer mention aluminium cabling at all?

I have a similar length line that was syncing at just below 30Mb for months, right up to 31st December when it took an almost 4Mb hit, where it has remained since.
There are very few users between my home & the cabinet, so I probably don't suffer too much from crosstalk that would end up lowering speeds.
Do you live in a densely populated area?

I see that DS & US interleaving depths both now have a value of 1:-

D:      1         1

That means OFF (i.e. Fastpath), so it does look as though DLM was also reset back to a wide open profile.

Interleaving may end up being re-applied over the next few days if sufficient errors are seen, but touch wood, the pair swap may have removed the issue that caused the lower speeds & high(ish) Interleaving depth.

There is still a slight blip at the L/H side of the Hlog graph, but overall, all the stats currently look much better.

If possible, I suggest monitoring your ongoing stats 24/7, supplemented with a few snapshot graphs for a couple of weeks, just to see what (if anything) changes.
Fingers crossed that nothing does.

[stumpy 101]

Hi all

 After been hit with the dreaded DLM again I took matters into my own hands , here is some before and after stats. lets hope DLM see the improvement and removes cap!

[Bald_Eagle1]

How long has it been capped at 15Mb & do you have any idea what caused the cap?

Interleaving depth looks quite high at 1027.

[waltergmw]

Hi Stumpy,

Perhaps BS can recommend a procedure to get the cap removed?

I believe that he does not favour DLM resets without there being a good reason after a JDSU examination has been completed.

We have had significant difficulties, firstly even getting an acknowledgement that a cap exists*, and then finding the only way (sometimes) is to arrange a SFI visit after which - if the engineer has been reminded to call his control, will the DLM be reset allowing the modem to renegotiate the optimum speeds.

* I have had a lengthy discussion with Level 2 in India who state the BT does not cap a line, but a restriction is applied if the line is of poor quality.

I have also had one incident where the cap was raised from 20 Mbps to 30 Mbps and the SFI was cancelled without warning, even though the line showed a 40 Mbps rate was possible.
I very much hope that this was an isolated incident as we could all get into a significant pickle if arbitrary caps are adjusted on a whim.

Kind regards,
Walter