Print

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

[Mike_IG]

Hi All,

I had an issue last month where my line would go down if the phone receiver was picked up (http://forum.kitz.co.uk/index.php?topic=14090.0), it turned out that the connection in the PCP cabinet was very loose and when the engineer moved it slightly it came out (apparently it's a very old style one which have to be screwed in at an angle), when he reconnected it the issue resolved itself.

Before I had the issue mentioned above I was getting 32/8 (fast path so to speak with no interleaving) but due to the loose connection and the connection being lost when the phone was used my line got butchered by the DLM and I went down to 28/6 (with high interleaving applied to DS of 1215 and the upsteam was banded to 6), now a week ago the upstream banding was moved upto 7.2 and now a couple of days ago it moved it again upto 8.5 (although I can only get around 8 ) so that has gone back to normal.
The interleaving on the downstream hasn't moved or reduced itself at all in this time frame and considering the US has sorted itself out I was hoping that it would of at least reduced itself or even removed it by now (like I said I didn't even have interleaving for months before this issue).

Now as I said before I didn't have interleaving on before so didn't have FEC errors but the current figure after 3 days 4 hours is 1177586 (CRC errors only 22 and HEC 36 with 8 ES) so it appears to be doing a good job, I was just wondering is that figure high?, and could it be the reason why interleaving hasn't reduced itself?. Looking at DSLstats attached pic I get spikes every now and again and unclipping that pic those 2 spikes at 10000 actually went to around 237500 (although HG612 modem stats is reporting half that although I've noticed that changing it to per sample instead of per min it tallys with HG612 modem stats), looking at the previous days on HG612 stats the spikes don't tend to go above 8000 (usually 6000max) but rarely it will be a lot higher (since the issue fixed on 26/7 the following are the higher spikes per minute) 12000 on the 28/6, 18000 on the 1/7, 11000 on 8/7 , 23000 on 10/7 and around 120000 today, all other times it's around 6000max spikes with most of the time it being 0-1000.

So I know that no-one knows the exact figures of what sort of errors are needed for DLM to kick in but does either the total FEC errors of 1177586 seem high in 3 days 4hrs or would that seem OK and the crazy high spikes every now and again be the reason why the interleaving has not been turned off like it was the 3 or so months before or could it be something else?.

Any help appreciated.

Thanks for looking.
Mike

[Mike_IG]

Just noticed at the time of that massive amount of errors today that my SNR shot up as well which has tended to mean that an engineer is at the cabinet, so I'm guessing thats why I got all those errors then.

[rob]

So I know that no-one knows the exact figures of what sort of errors are needed for DLM to kick in but does either the total FEC errors of 1177586 seem high in 3 days 4hrs or would that seem OK and the crazy high spikes every now and again be the reason why the interleaving has not been turned off like it was the 3 or so months before or could it be something else?.

My line was on fast path for many months until I replaced my AV200 home plugs with some AV500 ones which then caused lots of interference resulting in my downstream becoming interleaved (~1200 or so).  After replacing them with the original AV200 units I observed that my downstream FEC errors were in the region of about 10,000 every 15 minutes (the AV500 units were causing many, many millions in the same period).  Out of interest, the FEC count didn't really vary much with or without the AV200 adapters connected.

My line stayed interleaved for quite some time after that.  I rebooted the modem after about 30 days and saw the interleaving depth drop slightly to ~800 and then it finally removed itself after another 30 or so days and I've been back on fast path ever since.  Since the AV500 adapters were connected for a few days then I guess the DLM would've seen a long period of constant noise on my line rather than a short burst which may explain the length of time it took to return it back to fast path.

[Mike_IG]

Hi Rob,

thanks for the info, I did read about some home plugs being an issue with some lines, luckily I don't need to use them, your reply gives me hope that my line may drop the interleaving down some maybe even remove it (if this is the issue, as all the other stuff looks the same from before the problem I had), and all bar 1 of the big spikes I mentioned in the post had a big attainable increase as well (which as I said before seems to be when an engineer is at the cabinet, well the 3 times I've known one at the cab I've seen the increases anyway so it's at least linked to it).

thanks again.

[NewtronStar]

Hi Mike_IG having had a good look at your stats it would seem I have found someone with kind of the same line stats as myself, your FEC errors not that bad for a three day sync.

And I can see your on INP 8.00 with interleaving depth of 1215 ,give it five or so days and the INP should decrease to 3 and interleaving depth should also decrease, but it hard to know if you will get moved onto fastpath as some lines never do like my own.

Just for comparison look at my stats for 5 days sync

[boost]

I moved my old man's SR102 to the attic as an experiment (Sky 40/10).

Pings jumped up to 28ms from 11ms.

After I recovered consciousness, I put it back where it was originally and pings returned to 11ms in about two weeks ish.

I cut out about 5 metres of cable at the attic and used a new NTE with VDSL2 SSFP so I blame the latency blight on extra RFI from the attic (he lives on top of a hill as it is), although I don't have any stats to prove it.

[NewtronStar]

I moved my old man's SR102 to the attic as an experiment (Sky 40/10).

Pings jumped up to 28ms from 11ms.

After I recovered consciousness, I put it back where it was originally and pings returned to 11ms in about two weeks ish.

I cut out about 5 metres of cable at the attic and used a new NTE with VDSL2 SSFP so I blame the latency blight on extra RFI from the attic (he lives on top of a hill as it is), although I don't have any stats to prove it.

Yes Boost I also live on a hill it's great for the Ham DX'ers but the downside is your household wiring acts like a big antenna for RFI not good for Broadbanding 

[Mike_IG]

Hi Mike_IG having had a good look at your stats it would seem I have found someone with kind of the same line stats as myself, your FEC errors not that bad for a three day sync.

And I can see your on INP 8.00 with interleaving depth of 1215 ,give it five or so days and the INP should decrease to 3 and interleaving depth should also decrease, but it hard to know if you will get moved onto fastpath as some lines never do like my own.

Just for comparison look at my stats for 5 days sync

Hi NewtronStar,

Thanks for your reply, as you said our line stats are very similar, against your stats mine appear to be pretty good (error count wise) so hopefully the interleaving will go down, as I said before I've never been on interleaving (apart from 2weeks) for the 6 or so months I've had fibre, until I had the loose connection in the cabinet.
The worrying thing for me though is my line was fixed 3 and a bit weeks ago so I was hoping DLM would of done something by now for the downstream, especially since it's resynced at least twice since to remove the banding on the upstream which it's now fixed so I was hoping it would of at least lowered the interleaving in that time but it's stayed exactly the same.
Hopefully it will do something to the DS now that it's fixed the US issue.

Thanks again
Mike

[perry081064]

my line stats are pretty good really as NS would agree.
however my interleave has been high and continues to rise whenever theres bad weather etc.
be prepared for a wait , ive been waiting for mine to drop for the last 3 months !.

[boost]

Someone else will know better but I believe the DLM profiles are based on MTBE which is different depending which over riding profile is applied?

Something else tells me this maybe doesn't apply to FTTC though

Even if it doesn't, I suspect the MTBE will be different depending on whether you have an ECI or wahway cab so maybe the thresholds are higher/lower from one to the other.

Lots of guessing

[Mike_IG]

Someone else will know better but I believe the DLM profiles are based on MTBE which is different depending which over riding profile is applied?

Something else tells me this maybe doesn't apply to FTTC though

Even if it doesn't, I suspect the MTBE will be different depending on whether you have an ECI or wahway cab so maybe the thresholds are higher/lower from one to the other.

Lots of guessing

LOL yup I've just sacrificed a -insert animal or food here- to the relevent god to try to get it sorted.

I noticed an engineer working on the cab as I was on my way to the shops and thought I would have a look if it did anything when I got back, can you guess when he was at the cab lol.

[boost]

I was wondering why on earth the SNR increased and then it dawned on me... he must have lifted a noisy pair that's affecting your pair?

[Mike_IG]

I was wondering why on earth the SNR increased and then it dawned on me... he must have lifted a noisy pair that's affecting your pair?

It happens every time the engineer is at the cabinet without fail, (well the 5 or so times I've seen an engineer at the cabinet it's happened every time put it that way), others have said it's likely to be a loose connection at the cabinet and when the engineer is in there then a slight knock etc can cause a line change.
My connection was re-done recently though but since my cabinet is the old screw type then the connection is still shite so a knock here or there could still cause it to flutter a little, or maybe the same with a noisy line effecting mine I guess, either way like I said I get a temp boost everytime an engineer is at the cab, just wish the boost would stay lol.

[NewtronStar]

I noticed an engineer working on the cab as I was on my way to the shops

This as happened to me twice last year my SNRM shot up from 6dB to 10dB and the Attainable rate went upto 40000 kbps so I thought this was a great time to re-sync the HG612 and enjoyed seeing a sync of 36000 kbps, anyway had to leave my house and go back to work lo and behold there was an OpenReach Engineer working away at the Copper Cab.

I said to myself what ever he's doing please keep it like this, got home later and the SNRM must have nose dived and the HG612 did it's own re-sync.

So it does happen when the engineer is pawing around close to your pair, my question is if they can do this a few times why does the SNRM not stay at that level 

[kitz]

I was just wondering is that figure high?

afaik the DLM doesnt act on FEC's. At least certainly not on adsl.  FEC's are just a count of interleaving and error correction working how it should.  It will be looking at the more important errors such as ES & SES.

If interleaving was switched off, then there wouldnt be any FEC's - instead youd be getting CRC's... and if sufficient CRC's in a short time span they'd become ES.