Print

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

[plexy]

Thanks BE and NewtronStar

suggesting some sort of general & continuous US interferenece (despite US Interleaving depth being 1 - i.e. fastpath)

This is whats stumping me. The DS doesn't really have much in the way of issues, the problems seem to be on the US side, yet its the DS that gets interleaving and slower sync and not US.. weird!

possibly something being switched on/off, possibly even in a close neighbour's house?

would have to be a neighbours if its REIN. For some of those dips we were away on holiday and all electrics were isolated at the fusebox (apart from the downstairs ring main running only the router, modem and the pc for logging. Any suggestions for filtering interference?

Is there any evidence of anything untoward in the ROUTER's logs from around the times of the larger dips?

Nothing that I can see. It certainly doesn't interrupt PPPoE connectivity.

I prefer to plot the graphs using multiples of 1,2,3,4,6,8 days,

Thanks for the tip, ill do a few days in future

[plexy]

Just a quick update that plusnet are saying the line speed loss is likely because

more people have connected to the fibre service in your area and as such your line has dropped to a more realistic level or it could be down to the copper cabling in your area not being able to handle the full speeds.

I disagree with the second point there, as the copper was fine for the first 2.5 weeks. Plusnet also gave more diagnostics which showed no crosstalk, as per the earlier diagnostics. So either of the above is not really viable to me.

One thing that does stand out from the diagnostics they showed me is that they show no US CRC or ES at all, which is at a disagreement with the logging from HG stats which show significant CRC and Es on the US. Their stats do show a small number of 'ingress code violations' though.

The stance from plusnet again seems to be that the line is operating within 90% of the estimate so its fine. I would agree with that normally, but my ping is cut from 6ms to 20 due to DLM because there is some kind of issue on the line. From the plusnet responses, I can only go with the understanding of plusnets opinion that as the fault is 'tolerable' and is only degrading service, as opposed to interrupting service, that they believe its fine. They have also refused my request to remove interleaving from the line (after first saying it would code me £60 and an engineer visit to accomplish). Given that their sales blub states fibre is faster and less latency for gamers, part of the choice of choosing plusnet was lower ping times.

I have requested they move me to the speed DLM profile specified by BTW on the GEA FTTC factsheet. Lets see what they come back with then and how the line does there. If that fails, then there is no overall benefit between standard ADSL and FTTC in terms of latency for me,  which negates part of the offer of sale plusnet made under their 'gaming' section of the FTTc product

[c6em]

No ISP has access to interleaving status on FTTC.
It is all controlled by the DLM which is the only arbiter of whether it is on or off.
ISP cannot switch it on or off for you.
This is different from ADSL - I know.

[plexy]

No ISP has access to interleaving status on FTTC.
It is all controlled by the DLM which is the only arbiter of whether it is on or off.
ISP cannot switch it on or off for you.
This is different from ADSL - I know.

Ah that explains it then. However, switching to the Speed DLM profile should help, right? Its proving difficult to get plusnet to do that for me.

[plexy]

Additional question folks - Does my QLN look OK? I saw a post BE made on another forum (replying to a user who appears to have near enough identical line graphs to me) where he said that the QLN looked quite noisy up near 100 instead of 130-140. http://forums.thinkbroadband.com/fibre/4218198-interpreting-line-stats.html

I think ive also noticed an audible background hiss developed on the quiet line test. Almost like listening to the ocean? But quite quiet.

My US SNRM seems to have been a bit less funky lately, though now the large SNRM dips happen on hangup of a call, as opposed to the start (phonecall started around 19:00 in the graph below)

[plexy]

Well something a bit more positive today - bit of a breakthrough!

Found a plug socket in the front office which has a vodafone suresignal plugged into it. I forgot to disconnect that when I did the test with everything isolated, so as a long shot I unplugged it yesterday.

24 hours later, the only SNRM dips correspond completely to the phone being in use/off hook only. The mysterious random massive dips to 0SNRM are completely gone over the current 24h graph. It could be co-incidence so will keep monitoring.

Now I think about it, the suresignal did actually arrive around the same time as the DLM kicked in, so if this is the cause then hopefully DLM will notice and catch up at some point in the future. Will keep the thread posted if that happens (or the SNRM dips return)

[kitz]

Im only just getting around to catching up on posts over the past week  this interests me

I have requested they move me to the speed DLM profile specified by BTW on the GEA FTTC factsheet.

No we all know that the ISP has no control over interleaving, but the ISPs do have three profile options available to them.  All indications seem to imply that by default they are using 'Standard'.  Ive never seen a case yet where an ISP has moved anyone over to speed.

You have a valid point in that if an ISP advertises towards gaming then low latency is going to be important.

Have you had any joy with the suresignal?

[plexy]

Hi Kitz,
Plusnet reported back to me that all plusnet customers are automatically placed on the 'speed' DLM profile and you are only moved off from this by request. Therefore, my line is already on the 'speed' DLM profile apprently. With 1045+ interleaving this is surprising.

*edit* Plusnet also confirm the DLM policy here http://community.plus.net/forum/index.php/topic,116753.msg1009493.html#msg1009493

Re the suresignal, it did seem to stop the random drops in SNRM on the upstream that I could not attirbute to a phone call (meaning it did not stop the drops in US SNRM when the phone is ringing/off hook). I also checked the crimp connectors in the grey pipe on the outside of the house, one of the joints looked bloody horrible and appears full of (what I imagine is) insect poop or decaying insect juices (ew). The wires were also not twisted correctly up to the join, so i fixed that by twisting the pair properly. I also (as a bit of an experiment) tried a bit of faraday shielding with some tinfoil in the same section. It didn't do anything to the error rates, but what I do notice now is that when my phone is off hook, the US SNRM changes appear to only happen in U2 now. U1 and U0 have no SNRM change when off hook (though they used to), so im really beginning to think that the US SNRM changes when off hook may be a bad/corroded/insect juiced joint.

My FEC errors have been gathered over a period of 6 days constantly. Its clear there is a pattern corresponding to early morning (up to 8000 FEC errors per min between 07:00 and 09:00) falling sharply to a couple hundred per min through the day, peaking between 12 and 1 back up to 5k to 8k per min, then back down until around 15:30 (school finishing time) where it all shoots back up to 8000+ per min, before tailing off into the evening and bieng down to 'normal' by around 10.30 pm. This to me indicates either REIN or crosstalk (and given the times of plusnets crosstalk tests, they appear to always be ran in 'off peak' hours so no surprise they haven't seen anything). Ive confirmed that if its REIN, it is not from our house (two days we were out, with timed devices such as water heater switched off and sat boxes powered down) - the errors remain though.

Upstream is still messy - around 12 to 20 CRC errors per minute and 15 to 30 error seconds per hour. interleaving is still '1/fastpath' on the upstream. The errors are always there, irrespective of the time of day.

The line is stable though, no more decreases in speed, syncing at 66.5mbps and throughput is sold 60 meg. Upstream is 20 meg sync with speed tests showing consistent 16mbps throughput. My ping is not as good as it was, still mid 20's. I wish my 5ms ping would come back

So overall, yes the line is now stable. No answer from plusnet as to why it has deteriorated from 78 to 66, but as the deterioration in throughput is < 10% than the 'estimate' sync (72mbps), then they dont want to know. Its a shame as, as you noted, the product is sold as faster for gamers yet I would have had the same ping on ADSL2+

Ive added some ferrite chokes to power, ethernet and RJ11 lines going to the modem just to see if that helps but it does not. Based on the FEC patterns, im convinced I have a neighbor somewhere with a noisy TV feeding back up the phone lines. Will do the radio REIN test at some point when I find my old AM radio

[Chrysalis]

30 errored seconds per hour is ok. As long as none are SES.

My line gets more then that often with its normal behaviour although on the downstream.

30 per hour if we assume 1 per ES means on average an errored second once every 2 minutes which shouldnt be service affecting.

Of course I guess this depends on how your area is for noise etc. as some lines can get through an entire day only having a handful of errors.  If your line is like that this type of error rate is probably made you wary.

[plexy]

Thanks Chrysalis, good to know that the errors on the US are not a huge worry. They seem to be 24x7 so not at a particular time of day

Plusnet are now offering a phone line engineer to come and check the line in their latest reply. What do you good folks here make of this statement they have made though?

I have looked over the graphs that you have attached. I'm unable to advise on the Quiet Line test as this isn't something that we ever see any results. In regards to the SNR graph that you have attached this isn't something that I would be concerned over as on Fibre your connection isn't really affected by the SNR like it is on a copper connection.

[burakkucat]

Plusnet are now offering a phone line engineer to come and check the line in their latest reply. What do you good folks here make of this statement they have made though?

I have looked over the graphs that you have attached. I'm unable to advise on the Quiet Line test as this isn't something that we ever see any results. In regards to the SNR graph that you have attached this isn't something that I would be concerned over as on Fibre your connection isn't really affected by the SNR like it is on a copper connection.

I reads as if that statement has been made by a new employee who has not yet finished her/his training course! 

If I had a Plusnet service and that statement had been made to me, I would be referring it back to Bob Pullen or Chris Parr for remedial action.

From your description of the state of the external joints where (I assume) the service feed joins the internal cabling it reads as if they need to be remade (re-crimped). Flag up the 'manky-ness' of the joints with Plusnet and accept the Openreach engineering offer.

[plexy]

30 errored seconds per hour is ok. As long as none are SES.

#
Hi  Chrysalis,
Back h ome now so have had a check. Looks like some are SES. This is on an uptime of just over 48h.

ES:      13      1750
SES:      5      18
UAS:      29      29
AS:      247716

I reads as if that statement has been made by a new employee who has not yet finished her/his training course! 

If I had a Plusnet service and that statement had been made to me, I would be referring it back to Bob Pullen or Chris Parr for remedial action.

From your description of the state of the external joints where (I assume) the service feed joins the internal cabling it reads as if they need to be remade (re-crimped). Flag up the 'manky-ness' of the joints with Plusnet and accept the Openreach engineering offer.

Meow and thank you BC! I too thought it an odd statement to make Sadly chris and bob seem no longer active on this forum from their profile stats, but I will certainly try to raise them on the PN forums soon. I asked PN to send a SFI (fingers crossed) but lets see what they say in return.

[kitz]

Hi Kitz,
Plusnet reported back to me that all plusnet customers are automatically placed on the 'speed' DLM profile and you are only moved off from this by request. Therefore, my line is already on the 'speed' DLM profile apprently. With 1045+ interleaving this is surprising.

*edit* Plusnet also confirm the DLM policy here http://community.plus.net/forum/index.php/topic,116753.msg1009493.html#msg1009493

Hmmm now that is confusing.

Sorry should have clarified. "Standard" *is* the "Speed" profile, the only other two options we have are "Stable" or "Super Stable".

So, by default we do supply customers on the fastest profile available.

That seems to relate to BTw ADSL not  FTTC

BTw ADSL profiles =
Standard, Stable, Superstable

BTor FTTC profiles =
Speed, Standard, Stable.

It's this confusion between the naming that led Zen to believe that they were using the fastest FTTC profile, when in fact standard for fttc was the middle rate profile. 

http://forums.thinkbroadband.com/zen/t/4252187-re-ip-profile-reset-on-fttc.html

[burakkucat]

Meow and thank you BC! I too thought it an odd statement to make Sadly chris and bob seem no longer active on this forum from their profile stats, but I will certainly try to raise them on the PN forums soon. I asked PN to send a SFI (fingers crossed) but lets see what they say in return.

They (and other Plusnet staff) will 'look in' from time to time and will always respond to any PM from a Kitizen. 

[Ronski]

Just a thought Plexy, have you ruled out fridges and freezers, they can be electrically noisy and of course the compressors can switch on at any time.