Kitz Forum
Broadband Related => FTTC and FTTP Issues => Topic started by: can2002 on February 21, 2014, 07:34:37 AM
-
I switched over from ADSL2+ to FTTC at the beginning of January. I live just outside Chertsey in Surrey in a small sem-rural village that is half-served by the Chertsey exchange with the other half served by the exchange in Ottershaw. I'm the last house on my side of the road served by Chertsey and frustratingly, my cabinet is 2.09KM from my house (my neighbour is 0.6KM from his cabinet).
On ADSL2+ I was achieving rates of 3.5Mbps down and a fraction under 1Mbps up. After nearly two months, my FTTC stats are 2.6Mbps down and not much over 0.4Mbps up. While I'd not expected miracles, I'd (perhaps naively) expected to at least match my ADSL speeds with the hope that future developments of VDSL might gradually bring some improvements.
Openreach have been out a couple of times now and carried out a whole bunch of tests on my line and on each occassion they tell me the line is absolutely fine...
I've attached various screenshots from dslstats; I hope I've included everything important. My main observation (compared to postings from other people) is that there's a big hole in the bitloading graph. The FEC errors are noticeable, but I'm not sure whether they're particularly high?
I would appreciate any advice or feedback people can provide!
Regards,
Chris
-
Hi Can 2002,
If your line really is 2 km (presumably plus routing lengths as well) you're probably doing quite well for VDSL !
My first question is did Openreach examine your entire internal network as well as just from their test socket within the NTE5 master socket ?
Next I suggest you install an unlocked Huawei and run the modem monitoring firmware from baldeagle1 and that lovely Black pussy.
The Hlog graph can point to problems quite often.
Kind regards,
Walter
-
Thanks Walter,
I guess it could be slightly closer - the 2K is measured by road from my position on the road to the cabinet - two Openreach engineers have confirmed where my cabinet is.
Regarding the firmware, I'm running the following:
bcm96368MVWG_fs_kernel_HG612V100R001C01B030SP06_Wolfy_Unlocked
I presume this won't allow the monitoring you mentioned?
Regards,
Chris
-
You can display the HLog graph in DSLstats. Look on the "Items to monitor" section, under Configuration.
-
Regarding the firmware, I'm running the following:
bcm96368MVWG_fs_kernel_HG612V100R001C01B030SP06_Wolfy_Unlocked
That firmware will be fine for line monitoring. As you are already using DSLstats, perhaps you would configure the utility to produce the Hlog graph, please, as explained by Eric.
(Just outside Chertsey, Surrey. Almost within trundling distance of Walter's Wheelbarrow! ;) )
-
Brilliant, thanks both and will do!!
Sent from my GT-i9100 using Tapatalk
-
@ BKK,
That thought had also crossed my mind. Let's see what develops (and whether a canoe is still required). Incidentally I doubt whether all BT u/g joints can withstand submarine pressures either !
Kind regards,
Walter
-
Hi again,
Sorry for the delay in responding, I've attached the hlog graph.
Chris
-
@ Can 2002,
The Hlog confirms that all is far from well with your line and / or internal phone wiring.
I would start by checking you are on the first and only master socket with an integral filter faceplate direct off the incoming cable**, that no ring-wires are connected anywhere within your house - i.e wires are only connected to pins 2 and 5. ** Check both inside and outside of the house for connector boxes without any star wiring. Once you are absolutely certain that your wiring is perfect, then check Hlog again and if it's still similar openreach need to look for bridged taps and crossed pairs.
Kind regards,
Walter
-
Hi Walter,
Thanks for the comprehensive reply!
I don't have any extension lines in my house at all and the modem definitely connects directly into the newly fitted filter plate on the master socket (installed when I was migrated to fttc in early Jan).
The feed to my house is via an underground cable, which terminates in a small junction box outside the house, which in turn feeds directly to my master socket (I actually have two phone lines fed from the same 3 pair cable up to the master sockets, which are next to eachother).
My isp is going to send openreach out again, so I'll mention exactly what you suggested.
Sent from my GT-i9100 using Tapatalk
-
Hi Can 2002,
The reason I'm suggesting all is not well are illustrated in the thread below where an excellent Openreach engineer discovered we had crossed pairs.
Although the scales are different you'll see that the post-discovery curve is very much smoother.
http://forum.kitz.co.uk/index.php?topic=13505.0
If your phone provider supports 17070 option 2 quiet line test and option 1 ring back test you could try those tests on both lines.
Otherwise you might try ringing the VDSL line whilst you watch the modem's lights when there's practically no other traffic.
Perhaps the real Gurus might like to comment too ?
Kind regards,
Walter
-
I am at a loss as to what to type regarding the Hlog graph. I have never seen one that exhibits "straight-sided wells", such as the above. :o
This is a FTTC service, isn't it? (Rhetorical question, no answer required!) There should be 4096 tones accessible . . . :(
-
Hi and thanks both.
It is an fttc line, yes. Funnily enough when I first bought the Huawei modem I had an odd problem where it wouldn't sync at all unless I put the very latest hacked firmware (which burakkucat gave some advice on), which now makes me wonder if there is a shared root cause...
I'll try the last couple of suggestions from Walter (thanks).
Chris
Sent from my GT-i9100 using Tapatalk
-
Hi again,
I have an openreach engineer booked for this Wednesday and would like to be as prepared as possible for the visit.
Does anyone believe they will be interested or influenced by the graphs I've collected to date? Also, if there are specific questions I can ask (Walter suggested asking about crossed pairs or bridged taps)?
Lastly and regarding the 17070, the line is with BT at the moment so this should be fine. Sorry if it's a silly question, but should I be doing anything with dslstats while running the test, or is the goal just to run the test and listen to the results?
Thanks,
Chris
p.s. I hadn't enabled the QLN graph in Dslstats before - in case it's of use/relevant, I've attached a screenshot.
-
I would really like to see a re-plot of the Hlog graph, using the same x-axis scale as the QLN graph and showing all potential 4096 tones.
Perhaps Walter would care to don his waders and trundle the wheelbarrow to a village just outside Chertsey? :-\
-
Hi Burakkucat, thanks again!
Graph attached, although it shows nothing more (I'd cropped it before to show just the tones with data).
There was an option on the QLN tab "Show extended tone range", which I checked. There was no equivalent option on the Hlog tab, might I have missed something obvious?
Cheers,
Chris
-
No, I don't think you've missed anything in the DSLstats configuration. (I'm sure Eric would mention it, if you had . . . )
Thank you for the un-cropped version of the Hlog graph. It is nothing like what I have ever previously seen. :(
-
Thanks (I think ;-) )...
One thing I realise I haven't actually said is that I have an ADSL2+ on my other phone line. I'd assumed Openreach would have flagged this as a problem if it had been, but I feel silly not having mentioned this before...
Sent from my GT-i9100 using Tapatalk
-
Thank you for the un-cropped version of the Hlog graph. It is nothing like what I have ever previously seen. :(
Neither have I.
Hlog's maximum attenuation/minimum value for VDSL2 connections is around 96dB (i.e. no data, bitloading etc. can be obtained from the affected tones).
Hence the perfectly flat shape of the graph, in places.
Maximum attenuation for VDSL2 connections seems to be somehwre between 63dB & to dB.
It's most unusual to see such Hlog graphs from a VDSL2 connection.
To me, it indicates quite severe physical line problem, unlike the usual 'V' shape we see when bridged taps are present.
Unless the ADSL2+ connection is having a negative effect (not seen previously).
The issue does seem to be within the the ADSL2+ tone range.
-
Thanks again for all the feedback - I shall keep my fingers crossed for a good engineer tomorrow (the last guy suggested the solution was to drop me from a 80/20 profile to 40/10...) and have my graphs ready!
Chris
-
with vdsl when the only frequences available are those shared with adsl then typically adsl will yield a better sync speed as that wont have the power cutback.
-
Hi,
The visit happened without much of a conclusion. He was a fairly young chap who didn't really seem to be interested in the Hlog graph, he was just surprised my rate was so low and suggested he would not have been able to commission the job if I'd got less then 15Mbps (??)...
He plugged his tester in and got the same set of green ticks the previous two guys got and then "reset the fibre ports" before going to the cabinet and checking performance at that end.
He said he would log a job for an 'underground engineer' to check the connections between me and the cabinet.
The bearer 0 speed jumped up after the reset yesterday, but has been nudging back down and will undoubtedly return to where it was after a few days (graphs attached), fingers crossed for a result from the 'underground engineer'...
Cheers,
Chris
-
interesting your reply confirmed engineers in my area talking rubbish, the ones who visited me claimed when they get green ticks they cant justify 'any' work.
good luck with the undergound checks lets hope they find and fix a problem.
-
So I had an 'underground engineer' out on Saturday (1st March) and following various feedback her as well as a suggestion from an electronic engineer a bit more was done this time round. I explained the background of my issues to him and also showed him the HLOG graph and explained the feedback I'd received (here).
Firstly, the chap did the same 'copper' line test the other chaps had done previously with the same positive results as below. He made particular reference to the 'AC balance' being good (65%) given the distance of my line. He then attempted to do a broadband test, but his tester would not synchronise. It then transpired that each of the previous engineers had been unable to synchronise their testers either. To rule out something obvious he went to the cabinet and was able to sync his tester there.
One of the questions asked by my colleague was whether previous engineers had done a reflection test, so when he returned I asked the chap about it and after we agreed on terminology (he referred to it as a HR fault test) he took a look. He explained that they wouldn't normally run this manually if the automated test came back without error (which of course mine did not) but did so. In the process he noted a potential HR fault approximately 200 metres from my house and to his credit he got his laptop out and identified a potential junction box. He paced it out and the distance matched.
After re-terminating my connection (noting that there was no sign of a bad connection when he looked at it) he repeated the manual HR test and we both saw that the blip had gone. I then reset dslstats to see whether any difference was obvious on the HLOG graph, but there was no change. By this point he'd pretty much run out of options and said that he felt confident that re-terminating the connection would resolve the problem and estimated I shouldn't get below (of the sync speed progressively dropping from 6Mbps down to just over 2Mbps). While I didn't feel particularly confident, to be fair on the guy he made a lot more effort than any of the previous engineers.
I'd deliberately left it a few days before posting to see what happened to the speed and the initial signs are not great. Speed has been dropping each day (see attached graphs).
Chris