Kitz Forum
Broadband Related => Broadband Technology => Topic started by: kitz on February 15, 2018, 04:02:18 PM
-
Admin note - I've split this off into it's own topic so as not to distract from Max's thread here (https://forum.kitz.co.uk/index.php/topic,21048.0.html)
My daughter is moving house this week - not far, just practically the other side of the main road.
Although I knew that area got g.fast last year, she didn't... but now the move is going through she rang me up all excited.
So it's going to be byebye Plusnet and hello to g.fast trial (https://blog.zen.co.uk/g-fast-make-broadband-better/).
I believe the order has been placed :fingers:
-
Hi kitz good luck to your daughter for new g.fast service with zen isp. Hope it went all well. Is she getting the maximum speed 330/50 if her cabinet are much closer as a stone throw. Don't think my area isn't getting any g.fast until late 2019/20. What about you kitz is there any g.fast available to you yet?
-
The order has only just been placed so not sure when it will go live. I'm a bit immobile atm so cant visit to get stats. :(
No sign of anything like that around here. Haven't even got g.inp yet ::)
-
Ha ha ...... isn't that how it goes, eh ....... your daughter gets it before you, bl00dy typical !! ::) ;D
PS - was heading up your way today, but never got further than Poulton. :)
-
haha within waving distance. ;D
Actually DD is now much nearer your home patch now just a tad further south.
-
haha within waving distance. ;D
Actually DD is now much nearer your home patch now just a tad further south.
He he, yeah ...... I personally love it when they ship me over to your area ...... some really lovely houses I get to wander around under the guise of 'Looking for other extension sockets'. lol. ;) ;D ;D
I'm guessing DD is in a place beginning with 'R' ?? :)
-
Is she getting the maximum speed 330/50 if her cabinet are much closer as a stone throw.
She doesn't have a phone no yet so can't check exactly, but just run the details through using the address. I've looked up the cab and its about 160m on google maps. So 237-269 Mbps according to BTw's checker. Despite her not moving far it's put her on a different exchange that has had Zen LLU for a while.
Attached
[1] Checker results
[2] PCP and cab from google maps
[3] PCP now
-
That's interested Kitz. Those stats are probably will be the same for my cabinet if the g.fast went live. I think it probably will be around 220/30 from 80/20 as my cabinet approx 248m away.
-
It'll be interesting to get some real world stats and at least check on its connection speed/stability.
-
Found out today that there may be a delay of a few weeks. Not sure of the exact problem other than a rather vague 'awaiting some hardware'.
Her exchange is definitely one of the listed pilot exchanges.
----
ETA. Just found this (https://www.ispreview.co.uk/index.php/2017/12/openreach-offers-carry-customer-kit-g-fast-broadband-isps.html).
We should caveat the above by saying that some ISPs are seeing a few delays with the G.fast work and so commercial product launches may be put back further into 2018.
:(
-
Thanks kitz. Is it normal for BTW to show up G. Fast on checker even thought the cabinet isn't fitted any G. Fast pod yet. Nothing yet. I thought btw wouldn't updated on the checker until G. Fast pod has installed first and ready for service as planned. Last time I spoken to openreach they say my cabinet will be ready around summer 2018 for G. Fast.
I think there's some delays even if the checker is showing as available. My daughter is on one of the trial exchanges and pod in place, yet there are still delays. Admittedly I'm trying to read between lines here [pun intended], but I think may also have to do with ISP provisioning as I believe they will be one of the first residential g.fast.
Despite Zen having a presence in that particular exchange, I think they may be waiting on BTw to provision in this case rather than using the standard FTTC GEA cable-link. I'm getting this info 3rd hand so it may not be strictly correct, other than in this instance the SP appears to be wanting to use using BTw rather than their usual GEA FTTC cable link. She's been in the new property now for a couple of weeks and they still don't have any broadband... but the latest info is that hopefully it should be live in the next few weeks. Haven't got an exact date yet but it is meant to be very soon {2 weeks}.. and will obviously depend on one of them being able to take time off work for installation.
It's a shame MDWS is no longer live as I was going to try get over and set up on of my modems.
Using the powers of charm I recently picked up a new VMG8924 for me in return for a charity donation.. which means when I get around to it*, my faithful VMG8324 will be my spare.. leaving me a VMG1312 which I was going to give to my daughter to set up monitoring. However that said, they are not the type of people who would leave a PC on all day just to monitor stats. I will try get over asap, but right now until I get pain management more under control, my fear is that I may not be able to get there and back in one day, yet I have a poorly cat who needs 2x daily meds and cant go without his meds, so I can't stay over.
As far as g.fast monitoring of live data that will be extremely difficult. The Openreach modems are locked and afaik there is currently only one consumer g.fast modem/router (ASUS) which aren't cheap. :/
---
* I've got so many things now hooked up to my network, that it would be no mean feat swapping all my prefs and configs over to a new modem
-
they are not the type of people who would leave a PC on all day just to monitor stats
Would they consider a Raspberry Pi? It could be powered from USB, if a free port is available.
-
In a word .. No. :(
My daughter has no geek tendencies other than uploading photography and her own site stuff. As long as she has her music and other hobbiest things she's happy and tends to rely on me when it comes to broadband and PC's.
Her OH is geeky to an extent but more concerned about actually using the connection rather than how it works.. and although you'd be surprised if you knew where he works, he just isn't into stats and the intricacies of dsl.
-
I can understand that completely. My OH has NO interest in technology!
The only reason I mentioned the Pi was that VNC server is baked into the software, so it would give access to DSLstats on the Pi to anyone with the right permissions, even remotely. You could then monitor their line yourself (if that was what was wanted by all concerned).
-
My daughter is in a pilot area and still not live. Beginning to suspect the issue is/was capacity/supply at the ISP/BTw side rather than Openreach.
I haven't looked into it in great depth but it appears the SP needs to purchase a separate GEA cablelink for g.fast (ie cant use the same FTTC cablelink). So the SP has 2 options - purchase their own cable link (GEA for 'LLU' SPs) or use BTw for WBMC/WBC and purchase a host link.
Bearing in mind she lives in a area that is well covered by the SP 'LLU' and will be their only g.fast customer on that exchange (in fact afaik they will be their first residential g.fast) then the SP appears to have opted to use BTw for g.fast rather than their own GEA cablelink like they do for FTTC.. so add on additional time whilst they get either a host-link or their own MSIL. What I don't know is if they will be using WBC or WBMC - I suppose it depends on what they project for future customer requirements - on reflection probably WBMC as they already use that in some areas.
Anyhow I was there yesterday when they found out the order could now be placed with Openreach (yay but nay!)... which is a bit of a bummer as her partner had a week of work last week and can't take time off over the next couple of weeks for the installation. As he said, if only he'd known a few days ago and it could have been booked for last week when he was off. So still waiting, but now until its convenient to get time off work. ???
-
The Cable Links. So the pods go to switches in the exchanges. In ECI areas these can't be the old ones, in Huawei areas it's technically possible so a matter of policy. What would certainly be wise is using 10G CableLinks rather than the 1G that almost everyone's FTTC goes through. Really not good for performance having a decent size cohort of 300Mb-capable users sharing a gig with lots of 38/76Mb. You can get a LOT of subscribers going through a single switch and hence a single CableLink.
-
Cheers. Its a Huawei area - see photos here (https://forum.kitz.co.uk/index.php/topic,21048.msg365300.html#msg365300). All I know is they appear to have elected to not install another cable link and despite having their own cable link for fttc are going to use BTw for g.fast. I have no idea what bandwidth they have on their cable links but as you say, it wouldn't perhaps be the best idea sharing it with 38/76Mb users
I suppose this may only be temporary until they get more EU's wanting g.fast from that exchange? If Leo still worked for Zen I'd ask him, but I don't like sticking my nose in too much as the g.fast connection will be in her partners name so not really my place to do so when he has his own contacts, so aside from asking him on occasion I keep out of it.
-
Cheers. Its a Huawei area - see photos here (https://forum.kitz.co.uk/index.php/topic,21048.msg365300.html#msg365300). All I know is they appear to have elected to not install another cable link and despite having their own cable link for fttc are going to use BTw for g.fast. I have no idea what bandwidth they have on their cable links but as you say, it wouldn't perhaps be the best idea sharing it with 38/76Mb users
I suppose this may only be temporary until they get more EU's wanting g.fast from that exchange? If Leo still worked for Zen I'd ask him, but I don't like sticking my nose in too much as the g.fast connection will be in her partners name so not really my place to do so when he has his own contacts, so aside from asking him on occasion I keep out of it.
Spot on.
-
Before I forget, my daughter's line went G.fast live the week before last.
I'm afraid I don't have any stats and not sure when I can get over there to check it out first hand. I believe speed tests are circa 230Mbps.
-
230Mbps is better than nothing. :)
-
It is truly a different world.
-
230Mbps is better than nothing. :)
I'm not sure what the actual sync speed is. Speedtests will obviously be slower due to overheads. From memory I think their estimated clean line rate was 240-270 Mbps.
My daughter has no interest in such things as long as it works and isnt slow, so a speedtest is about the most I will get from her. :D I need to catch her b/f to ask but main topic of convo this week has been several things occurring which have nothing to do with the Internet.
-
But it always great to download file lots quickly plus uploads too.
-
I'm not sure what the actual sync speed is. Speedtests will obviously be slower due to overheads. From memory I think their estimated clean line rate was 240-270 Mbps.
I would be very interested in any further details you can find out, the clean estimate for me was 262-295mbps but my current synch rate of 218Mbps doesn't even meet the handback threshold :(
-
Before I forget, my daughter's line went G.fast live the week before last.
I'm afraid I don't have any stats and not sure when I can get over there to check it out first hand. I believe speed tests are circa 230Mbps.
I'm not sure what the actual sync speed is. Speedtests will obviously be slower due to overheads. From memory I think their estimated clean line rate was 240-270 Mbps.
I think I mentioned in another thread somewhere that their speeds had started to decrease to ~200 Mbps.
For the past week or so they have been getting quite a lot of periods of no sync and speeds have dropped further. No stats due to them having an MT992, but I do have some screen caps of their down time (red) and quick resyncs (amber) together with chart showing number of resyncs provided by Zen I've asked him if he can try to get them to provide some line test results showing sync speed/SNRM etc but I guess that will have to wait until Monday now.
Zen have logged it as intermittent fault possibly REIN. Engineer booked.
DD's b/f was quite amused when their Faults Manager sent them a link to quote "A brilliant website which gives information about interference and REIN"
-
DD's b/f was quite amused when their Faults Manager sent them a link to quote "A brilliant website which gives information about interference and REIN"
I can just imagine your smile of pride at the compliment. :)
-
Seriously when are Openreach going to learn from this lesson? ADSL trouble, ADSL2+ trouble, FTTC trouble and now G.Fast trouble? Never gonna to get better ain't it?
-
DD's b/f was quite amused when their Faults Manager sent them a link to quote "A brilliant website which gives information about interference and REIN"
I don't think there are many folk around the North West/Manchester area that aren't aware of your site with regard to REIN/Interference.
I tend to show most my EU's by way of awareness, and they are for the most part genuinely interested, saving the link to their favourites.
We'll soon have you as 'Dame Kitz', on the next New Years honours list !!! ;)
-
:D
-
For those who are interested in stats, see attached which I asked if he'd get for me :D
Engineer is there atm, The message I got was they are trying to replace a stretch of cable but first engineer had to call out someone else with test rod and tubing as there was no rope left underground to pull it through.
-
Blimey. Line Attenuation 39.1dB (very long distance from the cabinet) but a good Sync Rate 241 Meg down with 33 Meg up.
Enjoy your new G. Fast Kitz.
My goodness me copper line length 3.55km to the exchange
-
Line Attenuation 39.1dB (very long distance from the cabinet)
No its not. :'( Its 120m from door to cab not sure about actual line length - perhaps 200m? It could get 80Mbps on FTTC. So FTTC attenuation would have been something like 8-9dB.
You have to remember how attenuation changes depending on the technology in use. Even swapping between adsl1 and adsl2+ increases the attenuation despite the copper still being exactly the same length. The higher the frequencies in use, then the more the line attenuates.
g.fast uses higher frequencies, less power, different tone spacing and TDD.. all of which can have an effect on the attenuation measurement.
copper line length 3.55km to the exchange
But its the copper between the cab and home that matters.
-
Ok but why is my FTTC line attenuation changed from 11.3dB to 11.9dB? It's been staying the same Line Attenuation for a very long time even the cabinet didn't move itself lol as increase Line Attenuation caused sync rate to drop.
-
Read this - What is Attenuation? (https://kitz.co.uk/adsl/linestats.htm#attenuation)
There's different attenuation measurements, the ones provided by your router are not always accurate and can change depending upon certain things - heat, weather, power, gauge of cable, number of tones in use etc.
True line attenuation - or Insertion Loss - can be measured at the DSLAM at the exchange via diagnostic tests and this figure should remain fairly static.
Only the Openreach engineer can get that figure.. and I just noticed its in the 2nd image - 8dB
-
PS also bear in mind that the line is faulting and there's apparently a length of cable that needs replacing which is why I showed the 2nd image.
Blacksheep may be able to explain some of those figures. Im not sure what the A-B, A-E, B-E points are, although there is probably a really simple explanation.
-
As I understand it, very basically, 'E' is ground/earth, 'B' is ring and 'A' is tip. I don't know however why one is called tip and the other is called ring, perhaps it's something to do with positive and negative? Beyond that I don't know.
EDIT - Wikipedia may clarify: https://en.wikipedia.org/wiki/Tip_and_ring
EDIT 2 - Looks like I might be right perhaps? http://www.tech-faq.com/tip-ring.html
"Tip is the nickname for the electrically positive wire and Ring is the nickname for the electrically negative wire. Tip is the ground side, which is positively charged, and Ring is the battery side, which is negatively charged, of a telephone circuit."
-
PS also bear in mind that the line is faulting and there's apparently a length of cable that needs replacing which is why I showed the 2nd image.
Blacksheep may be able to explain some of those figures. Im not sure what the A-B, A-E, B-E points are, although there is probably a really simple explanation.
"A-B" is A-wire to B-wire.
"A-E" is A-wire to Earth.
"B-E" is B-wire to Earth.
Looking at the second image, we can clearly see that the pair (in the section of cable being examined) is faulty. The HHT even helpfully highlights, in red, the out of specification AC balance, which is shown at 44.9 dB. (I believe that an active AC balance > 60 dB is what would be expected to be seen.)
The insulation resistance, both "A-E" and "B-E", also strongly hints at a sickly pair.
[Edited to correct a typo.]
-
Would I be right in thinking that using technology at a higher frequency like G.fast might expose any potential cable faults on the line?
The reason I'm asking is if they are having to replace a cable now for G.fast it seems very coincidental that the line could run well at full FTTC speed but nearly immediately things start faulting on G.fast. Or is this purely coincidence?
-
A more demanding service can highlight defects in the cable pair.
The important point to remember is that xDSL technology uses the cable pair as a radio frequency transmission line. So thinking about the circuit in terms of low voltage DC or AC (at 50 Hz) would be quite wrong.
Let's make the assumption that a typical pair, in a typical cable, has, over the last 17 - 18 years, carried each category of xDSL service as soon as it became available. The upper frequency that the pair has been required to carry has steadily increased. In the beginning was G.992.1 (fmax 1.1 MHz), then G.992.5 (fmax 2.2 MHz), G.993.2 Profile 8a (fmax 8.5 MHz), G.993.2 Profile 17a (fmax 17.6 MHz) and finally G.9700/G.9701 (fmax 106 MHz, with up to fmax 212 MHz proposed). As the frequency increases so does the "skin effect". With ever higher frequencies being propagated along the pair, lesser and lesser of the metallic pathway is being used -- the current is being constrained to an ever decreasing annular ring, hence the descriptive phrase "skin effect". A less than perfect pair may have been adequate for telephony and, say, G.992.5 (ADSL2+) services but could be significantly disruptive to a G.9700/G.9701 (G.Fast) service.
I include the use of aluminium or aluminium-alloy conductors within my phrase "less than perfect pair". Aluminium is a highly reactive element which passivates rapidly in the presence of moisture and oxygen to form an amphoteric oxide. Aluminium oxide is far less conductive than the pure metal. Hence a cable pair made of an aluminium-alloy will have an outer layer (a "skin") of less conductive aluminium oxide. Which would be a big "no" when attempting to pass a high radio frequency current over such a conductor.
-
Would I be right in thinking that using technology at a higher frequency like G.fast might expose any potential cable faults on the line?
The reason I'm asking is if they are having to replace a cable now for G.fast it seems very coincidental that the line could run well at full FTTC speed but nearly immediately things start faulting on G.fast. Or is this purely coincidence?
Yes, because each new technology pushes the line harder. We are now entering a period where lines that behaved well on FTTC could show up any less than perfect copper when on g.fast.
Very simple example using the SNRM graph below.
On a short FTTC connection they could have had 11dB SNRM but because the Margin is so good it can easily withstand such faults and remain in sync. Note on this graph SNRM can fractionally dip to 0dB but the line still remains in sync. If the line isn't monitored then no-one will know it does this. If its only a few times a day then DLM wont care and no-one is ever any the wiser.
Put that same line on g.fast and it no longer has 11dB SNRM buffer so the connection is going to start dropping.
-
Thanks Ixel and B*cat for the explanations.
I was obviously having a slight senior or FM moment. I was thinking about it when I was out this afternoon and it suddenly clicked what the A,B & E were. d'oh (fresh air rather than being in front of the PC must have been beneficial) :D
-
Thank you for the indepth explanations :)
This isn't something I thought about before. But reading about a couple of people who have G.fast and their speeds deteriorate over a short time makes me wonder what percentage of them are experiencing the problem, and is this something OR anticipated?
-
I thought one of the ideas of G.fast was that you've got such a large amount of bandwidth that losing a few tens of Mb isn't supposed to matter, and also it's much faster to reconnect.
-
it shows higher number weaver as its measured on a higher frequency than adsl.
-
There is theory . . . There is practice . . . And often the two are found not to agree.
When the Profile 8a G.993.2 service was first made available every end-user had to have an Openreach supplied modem. By implementing that rule, live performance data was obtained for every circuit. (Hence the existence of the VLAN tagged 301.) One advantage of the mass data collection was that, after suitable analysis, it became clear the average U.K. metallic pathway was capable of supporting a Profile 17a service. Hence my surprise to know that mandatory usage of an Openreach provided modem and mass performance data monitoring has not been implemented for the G.9700/G.9701 (G.Fast) service.
As to the secrets within the Openreach branded Huawei MT992 modem, I know not. No rumours have reached me; no written note has been pushed under the door of The Cattery during the early hours of the morning. :-\
-
I thought one of the ideas of G.fast was that you've got such a large amount of bandwidth that losing a few tens of Mb isn't supposed to matter, and also it's much faster to reconnect.
They noticed the slight decrease in speeds a couple of weeks ago, they weren't too concerned about that.
Much more of concern was the amount of noticeable disconnects and amount of time where they had no sync at all which could last a few hours. There's been 157 loss of connection over the past few weeks according to the Zen RADIUS report.
What I may not have mentioned is that DLM started taking action and capped at 95Mbps.. and then I believe 24Mbps a couple of days ago which is how I interpreted the message I got which was "Capping the speeds between 24 & 95 Mbps". So obviously one very poorly line.
The stats show the line after a DLM reset.
---
ETA. Just had a look at one of the other files I was sent and it would appear that the line appeared to have got worse from the 21st of July.
The more the DLM took action, it made no difference to the line stability.
-
When the Profile 8a G.993.2 service was first made available every end-user had to have an Openreach supplied modem. By implementing that rule, live performance data was obtained for every circuit. (Hence the existence of the VLAN tagged 301.) One advantage of the mass data collection was that, after suitable analysis, it became clear the average U.K. metallic pathway was capable of supporting a Profile 17a service. Hence my surprise to know that mandatory usage of an Openreach provided modem and mass performance data monitoring has not been implemented for the G.9700/G.9701 (G.Fast) service.
Wasn't that more for managing the modems themselves, such as updating their firmware? Surely all the data can be collected from the DSLAM anyway?
-
Wasn't that more for managing the modems themselves, such as updating their firmware?
I think so. IIRC a few years back, there were several myths surrounding the what the modem was doing when it was 'dialling home'. These were debunked by a few people - Adrian Kennard (AAISP) being one of them.
Surely all the data can be collected from the DSLAM anyway?
Yes. Whilst the DLM system is mostly only interested in the instability parameters, other data such as Sync Speed, SNRM is routinely recorded via the Element Manager.
The ISP's can also have access to this data if they wish via tools available to them such as BRAT/KBD/RRT eg this is one for sync speed
(https://kitz.co.uk/adsl/images/KBD_RRT.png)
-
Is G.fast still considered to be in a pilot phase or as it officially gone commercial?
-
IIRC a few years back, there were several myths surrounding the what the modem was doing when it was 'dialling home'. These were debunked by a few people . . .
The best description is in a short series of Wordpress posts, from June - July 2012, published under the title "Through Infinity; and beyond? (https://ukinfinity.wordpress.com/)" with the subtitle "A closer look at the UK's Infinity Broadband Network".
-
Is G.fast still considered to be in a pilot phase or as it officially gone commercial?
I think a lot of the problems may not so much to be with Openreach, but more the SP's installation of a backhaul link from the exchange - See my post here (https://forum.kitz.co.uk/index.php/topic,22133.msg373034.html#msg373034) and the following 2 posts for more info. Hardly any ISP's are yet offering it because it requires a separate GEA cablelink... and if you only have say one customer at an exchange wanting the product then it can work out mighty expensive to install a 10Gb cablelink.
If you look at the timescale and bear in mind my daughter was on a pilot exchange, it still took nearly 3 months from date of ordering until they were live.
So until the SP's start installing their cablelinks (or use a BTw WBC one if there's one installed) then getting finding an ISP to get connected with could take a while. :/
-
Update: Still not fixed.
they are trying to replace a stretch of cable but first engineer had to call out someone else with test rod and tubing as there was no rope left underground to pull it through.
Now waiting for planning permission from the council for the street to be dug up so the cable can be relaid.
Fault is being blamed on a gas explosion last month which was suspected to have ruptured a gas pipe. Apparently when the Gas board dug up the road for emergency repair it's caused damage to Openreach's ducting and cables.