Print

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

[Berrick]

Kitz, is your line still totally stable?

As there are quite a few similarities between your fault and mine I wonder if the cause of my fault may be the same?

If I get time over the weekend I might try bypassing the NTE??

NE hose I'm glad your line issue seems sorted

[burakkucat]

If this was the original NTE (which I now suspect it likely was)...  and remembering the state of the BT66 which it directly backed on to, then it wouldnt surprise me if the circuit board had been exposed to the same damp conditions which had totally corroded the terminals inside the BT66.

Ah, then that is highly likely. And in that case, I would be tempted to get a tube of silicone sealant and fill the hole, via the BT66, to stop any future moisture penetration affecting the extension socket that is now located on t'other side of wall.

[kitz]

>>> get a tube of silicone sealant and fill the hole

Good suggestion - thank you.

>>> Kitz, is your line still totally stable?

I know this may sound silly...  but it 'feels right' this time.  The voice line is crystal clear now.   My SNRm doesnt budge when I use the phone, nor when the phone rings.

I have lost some upstream SNRm (it previously could reach >18 dB) which has knocked about 1.5Mbps off my upstream attainable rate which happened when the engineer tested at the cab last week, so perhaps he disturbed something there, but since that isn't affecting day to day running Im not going to bitch about it.

Im getting hardly any errors - no FECs and only a couple of CRCs & HECs.
My Bit loading now looks right..  HLog looks good, QLN is a nice shape.

I'll attach a screen grab of my SNRm for the past 24hrs..  how boring is this (fault was fixed at about 10.30)

I'll also run a HG612_modem_stats for the past 24hrs, because I want Bald Eagle to see what my Upstream SNRm now looks like...  all of those weird crossovers have now totally gone.

[kitz]

>>> If I get time over the weekend I might try bypassing the NTE??

This may not be so easy.  Of course I have to publicly state that by totally removing the NTE you are encroaching on BT's equipment on which we the general public are not allowed to touch as its the other side of the demarcation point.

However I dont mind stating what BToR did.  Using the test socket in the NTE showed that the line still had a problem, but when he removed the NTE he grabbed hold of the bare black (dropwire) cable which came into my house.   Using alligator clips he attached his JDSU to the dropwire and then watched what happened when he rang my number.

[Berrick]

Thanks Kitz,

I don't really want to cross to the "dark side" of the demarc but I may have too 

[kitz]

Okay...  Im winding down my ticket now with Plusnet so I'll add a couple of comments here too for the sake of closure.

Ive been continuing to monitor with DSLstats & HG612 modem stats for the past 4 days and I can say things have been rather boring.

Voice line is crystal clear and no more SNRm fluctuations when the phone is in use/rings.   No more disconnects.
SNRM completely flat lines... its not even budged by 1dB since the fault was found.   
All that weird cross-over on the SNRm has gone and all 3 bands are equal.

Plusnet asked me to send some stats up.  Because I can only add one attachment per ticket I did a before & after montage,  which I'll add here.  I think you'll notice quite a difference.

Id like to say thank you to everyone who added input into this thread, theres too many of you so forgive me for not naming each one individually but there are certain people whose posts have been extremely helpful.   I cant believe that I was ready to throw the towel in and it only took one engineer to actually listen to what I was saying for the fault to be traced...  I'm so glad that Plusnet automatically rejected and booked another appt for me.

---
Bald Eagle just look at my upstream SNRm now..   see how all 3 bands are at the same level.   Unfort I was too much of an FTTC n00b to not realise that on my line... the SNRm bands should all be level.   Its almost like the fault on the circuit board was causing noise in the lower frequency bands and the higher tones were making up for it.   I wasnt able to monitor my power output for that band because it was too much of a negative figure and never graphed.   However, looking at a now and then on pbParams,  since the fault has been fixed, my tx power has gone down another -5dBm

[burakkucat]

        

[Bald_Eagle1]

Plusnet asked me to send some stats up.  Because I can only add one attachment per ticket I did a before & after montage,  which I'll add here.  I think you'll notice quite a difference.

Indeed that does look just a little better.


The shame is though, that more 'normal' users would simply have no idea how poorly their connections were performing against what they should be capable of.


Even with such undeniable evidence, my own experiences tell me that not all engineers would be at all interested in graphs provided by a 'geeky' user, especially if his/her 'standard' tests all came up with LTOK, as we unfortunaltely know they often will.


At least you are with an ISP that does understand & does take responsibility for pursuing matters on users' behalfs, albeit rather slowly at times.

[kitz]

I think the hardest part was being aware that I had that 18dB margin, and therefore the line only dropped when the phone was in use and things got really bad.

 

If you will recall though, this fault orginally developed when I was on adsl2+ with a margin of just 3dB.   It had held rock steady for years, but all of a sudden I was losing sync several times per hour.    On less of a margin on adsl2+ it would have been much easier to replicate the fault.
If you look back at some of my early graphs you can see that the line was dropping out many times per hour regardless or not if the phone was in use.

Unfortunately the fault reared its head just after I'd decided to move and asked for my MAC, so there wasnt any point me raising it with BE* I was half hoping it would vanish during a change of MSAN.   If I'd been on a longer line with less of a margin it also would have had way more of an impact.   This is why I was nearing throwing the towel in and wonder if I should just accept that most of the time it worked ok...  until I used the phone.

Having had a perfect 24/2.6 for years to suddenly start seeing frequent disconnects I knew something was up.


>>> my own experiences tell me that not all engineers would be at all interested in graphs provided by a 'geeky' user

I totally agree they are not:(  It was only the guy this w/e that took note...  in fact he even said he was going to have a go with his own HG612.

>>> At least you are with an ISP that does understand & does take responsibility for pursuing matters on users' behalfs, albeit rather slowly at times

I guess I cant fault the service as such, but like you say the ticket answer time is often slow hence always a week between something happening...  could be a hell of a lot worse though I guess.

[Bald_Eagle1]

>>> my own experiences tell me that not all engineers would be at all interested in graphs provided by a 'geeky' user

I totally agree they are not:(  It was only the guy this w/e that took note...  in fact he even said he was going to have a go with his own HG612.

Wey hey!!!

I hope you pointed him in the direction of where he could get hold of a stats harvesting/graphing program or two 

It might not be too long before we see engineers firing up DSLStats & HG612 Modem Stats on their laptops in preference to their JDSU's when troubleshooting poor connections.

[kitz]

I did indeed BE   I think this guy was a bit on the ball 'cause he noticed himself the 2 LAN cables and asked why.

[Chrysalis]

Eurika!

I think today's engineer has found the problem.    Engineer this morning understood and could see the large fluctuations in my upstream whilst the phone was in use...  and said he was off to change the d-side.   A few mins later he called me to 'come look at this' as he noticed that the d-side to the BT66 was fine.    He showed me the tests on his JDSU and sure enough the line was rock steady even with the phone in use and phone ringing.   

So he comes back inside, and plugs his JDSU again from the test socket...  line starts fluctuating and upstream dipping, along with the max sync dropping.  So its not the fttc i-plate.    Removes the NTE completely and tests from the bare wire.    Line is fine.
Repeats the test, puts the NTE back on, trys the test socket..  line plays up. 
Gets a brand new socket, installs it...  and line is perfectly fine.

Ive no idea where the NTE came from, but according to todays engineer its not a new one, as he can tell by the badging.  It's possibly the same one thats been moved from the old NTE location.   I think the engineer was a bit surprised as he said its not an obvious failure point and not something they normally look for.  I repeat this is not the face plate, nor the i-plate that failed, but the back part of the NTE

For the first time in several weeks my SNRm doesn't budge when the phone is in use nor when the phone rings.  I really really hope that this is what it was.   

you know what I will say as I am a jealous mardy sod

you had a great adsl line, a fault appeared, fault early on with your vdsl.

but instead of the responses I typiclaly get "deal with it", you got a good engineer.

maybe my mistake was acting dumb (when my engineer came I hid away my unlocked hg as I was scared it would be taken away for unlocking the thing.  My engineer and his boss made a comment that my 60mbit sync was extremely good, 80meg sync impossible and that the error count I had at the time was excellent (much worse than your error count).  So yes my moan is about the inconsistencies of treatment different people get on faults.

My own personal line is still about 35mbit down on the attainable from install day.  Clearly massively underperforming and my error rate is getting higher and higher which has me worried I will be interleaved soon.  No point me getting an engineer I will be told its all normal.

Also my snrm drops about 4-5 db during calls. Engineer ignored that also.

My theory is engineers and their area managers will know the state of any given area, if an area has a batch of bad lines (meaning everyone asking for pair swap) they get told no cant be done, else they be pair swapping all day long all year.  But if an area the lines are good and a fault stands out then they look into it and try to fix.  Its the only rationale I can put on the different levels of treatment people get in different areas.

Glad you got yours fixed tho.

When I read your adsl post at start I did chuckle I admit as I put up with that kind of issue for years on adsl.

[Chrysalis]

kitz I forgot to ask do you have an overhead or underground line outside your property?

[kitz]

UG cabling.   Theres a photo somewhere in the thread showing the BT66 and how corroded the terminals were.


>> I hid away my unlocked hg as I was scared it would be taken away for unlocking the thing.

Plusnet know full well that some users unlock and they are very well aware of Bald Eagles Graphing scripts.  In fact if they know you have an unlocked modem they will often ask you to upload BE's full monty graphs.

In my particular case it was PN that encouraged me to show the graphs to the engineer so he could see too. 

The problem is that not all engineers will understand them..  and I did have one engineer who was entirely dismissive telling me that all the CRCs he saw racking up on his JDSU werent important and my line was fine cause I had no FECs (on an un-interleaved line d'oh).  I did politely say that perhaps he may have that the wrong way round, but he insisted he was right and CRCs corrected themselves and he wasnt going to argue.  On two of my appts the Test Head was down at the exchange anyhow and full tests couldnt be completed.

If you notice I was also having voice issues with some very noisy voice calls and had resorted to using to using my mobile more as I was getting tired of having to ring people back when they couldnt hear me on the landline.   In fact Plusnet rang me and the guy I spoke to was witness to call where he couldnt hear me at all about 3 times due to noise bursts, so they were aware first hand that there was an intermittant voice fault too.

You will also note if you read through the thread,  that I too was just about the throw the towel in.  However as it turned out I didnt have to because PN straight off rejected BT's FNF and had already escalated it & arranged another appt (with an SFI engineer this time) without me saying anything.


This is one area PN's BOT team do seem to be pretty good at, and if they can see your graphs have something glaringly obvious,  then they do appear to fight your corner as much as they can.

>>> When I read your adsl post at start I did chuckle I admit as I put up with that kind of issue for years on adsl.

An adsl line that had sync'd at full 24Mb and 2.6 up on annexM...  and sat stable with an unwavering SNRm of 3bd for 6 years without issue totally flatlining. 
Then practically overnight started losing sync about 20 times per hour.   Downstream dropped to 16Mbps and at times it lost the ability to sync at Annex_M  because of the upstream issues.
The only reason I didnt pass it over to BE is cause Id already requested my MAC...  If I hadnt already started the FTTC process then I would have been screaming at BE to get it sorted.  That is some fairly serious degradation in my books.

[Chrysalis]

thanks, I have long suspected UG cabling has a huge impact on error rates (as well as snrm variation) so your answer is added to my database.

My line is UG but only up to the pole, my dropwire is from a pole exposed to radio signals etc.

Also I agree PN do seem good at dealing with these type of faults.

Also I do agree on your adsl experience I wouldnt be happy in the same boat, my chuckle was more in reference in that we were both used to different standards on our adsl lines as yours was of much higher standard.