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'm wondering if anyone could help with some issues I've been having recently and if a graph looks right.

OK so since the dawn of time I always had an issue with my ADSL connection (have since upgraded to FTTC), basically I work from home and so ordered the ADSL Max premium product which gave me a faster upload, the trouble was 99 times out of 100 whenever the phone rang the line would go down. We had so many engineers come out (close to 15 I would say) and no-one could fix it, you name it they did it, I had D and E side changes, at least 5 lifts and shifts etc and still the same thing happened.
Long story short(ish) it turns out that on the lower upload speed my line stayed up a lot more (prob went down 50% of the time instead of 99%).

Anyway at the end of last year thanks to the BDUK funding my cabinet had a fibre cabinet linked to it and I upgraded to fibre (first one on the cabinet to do so), changing to fibre seemed to fix the issue and I was happy with my connection which had gone from 2meg to 53 meg download and 0.4meg to 15meg upload.
Anyway roll on a month and my speed dropped to 30meg download and 8meg upload, engineer was called out to look at the speed drop issue but basically as I was 700m or so away from the cabinet this was the speed I could expect and that I shouldn't of even been able to get 50+meg (he didn't even believe me until I showed him my line stats).
Fast forward another few months and my speed had dropped under 30meg, when that happened I decided to get myself a HG612 modem and unlock it so I could view what was going on, that worked well but I wasn't on the latest firmware so I decided to upgrade to the latest firmware and within a week my line went from 28meg up to 32meg and interleaving was turned off, great I thought and all was fine.

In the last few days though my line went down when the phone rang (so like the issue I had when I was on ADSL), and now if anything it's worse as when we just pickup the phone to dial out the upstreams noise margin plummets (sometimes causing a disconnect), to me this again seems to point to a HR fault issue like I had before if not the same HR fault that I had before that no one seemed to be able to find but wanted to get other peoples opinion on the matter.

Also when it went down I decided to have a detailed look at the other graphs that were recorded and noticed that my bitloading graph seems to look different to everyone elses that I've seen and just wondered if it looks right to everyone else (specifically near the beginning as everyone elses I've seen has a gradual drop where as mine has a V shape from a drop then rise and then the gradual fall).

So to summarise

1) Does the first issue with the noise margin drop look like a HR fault issue or could it be something else???.
2) Does that bit loading graph look OK to everyone else or is that V shape an issue (I could only find 1 other person with that and it turned out to be an issue)

Thanks to all that made it this far and even more so to anyone that replies

Mike

[burakkucat]

Hello Mike, welcome to the Kitz forum.

Yes, after reading through your line's past history, it does appear to indicate a HR or semiconducting joint somewhere in the circuit between your home and the cabinet. Have you:

• carefully checked any internal wiring that may exist on your side of the NTE5/A?
• performed a QLT and listened for any audible abnormalities?

Regarding the shape of your bit loading graph, no there is nothing wrong with it. That V-shape wedge is a result of power cut-back, which has implemented to minimise any D-side cross-talk with other ADSL signals, signals which have originated from within the exchange building and have been attenuated by the distance over which they have travelled before reaching the PCP.

[Mike_IG]

Hi burakkucat,

thanks for the welcome.

I've plugged the phone directly into the Mk2 filter and the same issue still occured.
I only had a dect phone at the time so couldn't really do a QLT due to the noise you get on them to begin with, I've got a corded phone now but.............

for whatever reason the issue has gone (well almost) now if someone picks up the receiver if anything the noise margin actually goes up (while on the phone earlier the noise margin was over 7 on U0 band, although it's now down to 6.6 now it's not being used).

just out of interest whats the difference between SNR Margin and SNR margin per band?, on both the downstream is a straight line and doesn't change and neither does the upstream SNR Margin but when I look at the SNRM per band the upstream one jumps up and down all the time (often by at least 0.5 although it seems to stay between 6 and 7 at the mo).

The reason I asked about the bit loading graph is someone said that the V shape could of been caused by a bridge tap issue, but if it's the result of a power cut-back that's fine, I've just not noticed it on my friends graphs I've only ever seen the small gradual drops or the big spikey drops on 1 tone that come straight back up on the next, not the big V shape like mine over the number of tones mine was.

I forgot to put at least part of my D-side is aluminium as well so obviously that's not good, I think they should just remove the aluminium and replace it with copper, what do you think my chances of that are .
actually scrap that they should do it all fibre as this part was down to be upgraded to FTTP not FTTC but apparently they found an issue somewhere and it would of cost to much money to do so it was scaled back to FTTC.

thanks again

Mike

[custard]

The reason I asked about the bit loading graph is someone said that the V shape could of been caused by a bridge tap issue, but if it's the result of a power cut-back that's fine, I've just not noticed it on my friends graphs I've only ever seen the small gradual drops or the big spikey drops on 1 tone that come straight back up on the next, not the big V shape like mine over the number of tones mine was.

Mike

The bridge tap would show up on your hlog graph. The decline would not be smooth.

[Mike_IG]

Hi Custard,

thanks for that.

looking at the attached hlog graphs it looks pretty smooth, so I guess there is no bridge tap on the line.

thanks again
Mike

[burakkucat]

Joints in aluminium cables . . . 

It's just the fact that aluminium is such a reactive metal (it auto-passivates itself by forming a coating of aluminium oxide) that any joint could be suspect and the potential for a HR or semi-conducting joint is . . . 

For an authoritative "run through" of what the various graphs tell for your VDSL2 circuit you really need the attention of the Baldy Bird (a.k.a. Bald Eagle1). Let's see if I can "whistle up" an eagle for you . . . 

Attached below are two Hlog graphs obtained from the same circuit. The first graph shows when the circuit had a bridging tap in situ. The second shows when the defect was removed from the circuit.

[Bald_Eagle1]

I have attached the montages for the same connection - with & without the bridging tap.

@ Mike_IG,

What would you like to know about the various graphs?

[burakkucat]

   Hmm . . . Looks familiar, to me.

[Mike_IG]

I have attached the montages for the same connection - with & without the bridging tap.

@ Mike_IG,

What would you like to know about the various graphs?

Hi Bald Eagle,

I think I've got most of it from what people have said now thanks, I guess you can't really tell as I don't have the graphs before the speed drop but as I mentioned at the start I was at 53meg download and 15 upload and am now on 32/8 ish, is it possible to tell what caused the large drop, I realise cross talk would be involved but can you tell from the graph how much speed I've lost because of the power drop enforced (as in the bits missing from that big V shape on the bitloading graph, is it a set amount of speed per bit or not, i've no idea about bits if you couldn't tell  ).
 I also noticed that my D3 band is not being used, Attenuation to high i'm guessing, could maybe that of been used before but the crosstalk took it over a threshold and I lost speed from that??, and if so do we know what the threshold is or is it not set?, the reason I ask is that my upstream U2 band is at 58 so wondered if I'm about to lose that.

1 thing I did notice that when I was on an earlier firmware version for the router that often my bits per carrier on the bitloading graph would plummet every few hours and then go back up to then normal levels, any idea what would of caused this to happen??  (I should have the rest of the graphs saved somewhere if they are needed).
I noticed that when I updated to the latest firmware ....038m that the issue stopped, I also noticed the all the shared bits/tones all became just download ones and not upload was that changed in the latest firmware do you know??.

Thanks
Mike

[Bald_Eagle1]

I think I've got most of it from what people have said now thanks, I guess you can't really tell as I don't have the graphs before the speed drop but as I mentioned at the start I was at 53meg download and 15 upload and am now on 32/8 ish, is it possible to tell what caused the large drop, I realise cross talk would be involved but can you tell from the graph how much speed I've lost because of the power drop enforced (as in the bits missing from that big V shape on the bitloading graph, is it a set amount of speed per bit or not, i've no idea about bits if you couldn't tell  ).

Looking at your band plans, you are connected to an ECI cabinet DSLAM.

It's not at all unusual to see such a V shape in the bitloading graph.

We think it's due to power cut back at those lower frequencies, to avoid swamping any ADSL connections that may have significantly weakened signals due to the original PCP cabinet being a long way from the exchange.

I'm connected to a Huawei cabinet DSLAM & I see a similar V shape.
I am over 5 km from the exchange & around 1.1 km from the cabinet.

I could only achieve around 1 Mbps sync speed on ADSL & at first, I achieved up to 35 Mbps on VDSL2.
Since then, I have been hit by increased crosstalk & I'm now lucky to get around 20 Mbps sync speed (currently 18.73 Mbps).

Over time (24/7 logging) I have seen a number of occasions where my DS SNRM would drop quite suddenly by 1 or 2 dB & various errors would increase for a few days before the connection would resync.
At the resync, SNRM would increase back to the target 6 dB, but sync speed would always be 1 or 2 Mbps lower.

QLN on my connection has also worsened at a similar rate.

I also noticed that my D3 band is not being used, Attenuation to high i'm guessing, could maybe that of been used before but the crosstalk took it over a threshold and I lost speed from that??, and if so do we know what the threshold is or is it not set?, the reason I ask is that my upstream U2 band is at 58 so wondered if I'm about to lose that.

Maybe the D3 band was used previously, but as you suggest, attenuation is now too high.
Attenuation is around 67.8 dB for my D2 band & I only get a small amount of bitloading from that band, so the cut off point is possibly somewhere around 70dB.

Line attenuation shouldn't really vary much (slightly increased during warmer weather), so unless there has been a permanent deterioration of line conditions, it probably hasn't changed much.

I suspect that crosstalk is the main factor in your reduced sync speeds, worsening QLN & SNR, and/or reduced power.

I think BT could use power of up to 14.5 dBm for profile 17a VDSL2 connections, but choose to restrict it to something like 12 dBm anyway & when crosstalk is likely to be caused is a factor, DLM reduces it even further.

Do you have any older graphs or even raw telnet data at all from when you achieved higher sync speeds for comparison purposes?

1 thing I did notice that when I was on an earlier firmware version for the router that often my bits per carrier on the bitloading graph would plummet every few hours and then go back up to then normal levels, any idea what would of caused this to happen??  (I should have the rest of the graphs saved somewhere if they are needed).

I think that was simply down to a 'glitch' in the modem's previous firmware version(s).

I noticed that when I updated to the latest firmware ....038m that the issue stopped, I also noticed the all the shared bits/tones all became just download ones and not upload was that changed in the latest firmware do you know??.

Yes. 'Shared' tones disappeared when firstly the DSLAM band plans changed, shortly followed by another slight change via the remotely applied modem firmware updates.



You may/may not have also noticed that DS Line & Signal attenuation values no longer match exactly for the majority of connections, unlike with the older firmware.

We have no idea how accurate the reported values are, but they do at least now show a difference.

It is perfectly feasible to have low Line attenuation, but due to the effects of crosstalk or other 'noise' interference, worsened QLN & worsened Signal attenuation as a result.


FWIW, I have attached a recent snapshot montage from my connection for comparison against yours.

[Mike_IG]

Thanks for the info.

Very helpful.

[Mike_IG]

Hi all,

Little update with graphs (inc 3 connection resyncs) after my connection decided to go crazy earlier.

The U1 green line spikes down from 8 is myself picking up the phone receiver and the spike back up afterwards is me hanging up, the crazy part at the beginning I have no idea as the phone was not being used at that point, whatever it was though it caused at least 3 resyncs. (the U0 band which is the most shaky doesn't seem to be effected by the phone pickup at all, yet U1 and U2 which the noise margin is pretty darn constant are effect a lot, especially U1).

I've noticed that the Upstream has been battered into submission and the power from that has taken a big hit, also the attenuation has gone up (possible new person/cross talk issue??), the power had recovered a little though.
The Downstream has actually gone up a little though (the downstreams SNR also went up close to 8 when the upstreams SNR plummeted at the disconnects).

My Downstream seems a lot more stable than my Upstream and my Upstream was always an issue when the phone rang when I had ADSL (well my router always reported the upload going down and like I mentioned in the first post, lowering the speed down to the standard 448k seemed to fix this, someone a few roads away has issue with upload as well on same cabinet), just wondering why it always seems to be the upstream that's the issue?, if it was a HR fault wouldn't it effect the download as well?, or is it just the way the graphs report it?

Mike

[Bald_Eagle1]

Do you have the ongoing stats FULL__MONTY montages for us to have a look?

The snapshot graphs you have been posting are usful, but the ongoing graphs show any changes each minute, which can be very useful.

[Mike_IG]

Hi Bald Eagle,

Is this it??, the PC is currently in the bedroom so it gets turned off every night/morning so that's why the graphs tend to show no changes 11pm -10am or so each day.

Regards
Mike

[Mike_IG]

Just noticed that for whatever reason the noise margin has shot up at the moment as well, (obviously the attainable has shot up as well). I notice this seems to happen a few times a month but doesn't last long and I've never actually caught it live so to speak so will be interesting to see how long it lasts. 

*update*

latest around 1hr before dropping back to normal.

*update2*

was normal for 40mins or so and then shot backup for 10mins or so before again dropping back down to normal.