|
|
|
Hi folks,
Had BTOR out all day trying to sort out a fibre problem, I have 2x VDSL 80/20 lines, one from Talktalk and one from BT. BT line uses a homehub5 as modem, TT uses their HG635 super router.
Each line works fine on its own but if both modems are on the speeds plummet.
The TT line syncs at 79987 and shows a max rate of 100M
The BT line syncs at 79987 and shows a mas rate of about 87.5M
I called BTOR out because with both modems on we were getting full 79987 from the TT line but only 57M from BT.
The BTOR engineer found a minor fault on the BT line so re-terminated the junctions, it cleared his fault but made the fibre speed worse!
Sync speeds are identical with only one modem on, but with both on now the TT is down to 56M and the BT 53M
BTOR is coming back Tuesday but don't know what to do, one suggestion was to install a separate drop wire from the pole which I don't mind.
Anyone got any suggestions on this? I presume it is a crosstalk problem, I don't know whether noise vectoring is in use yet or whether these modems have that ability, any suggestions welcome, I did think about buying a Bipac 8800NL as an experiment to try a different modem although the speed drop is similar when BTOR had their test kit on the other line.
TIA
|
|
|
No vectoring around yet. One presumes the drop wire to the home is twisted pair and not very very old untwisted.
Changing modems may help, but its an experiment to see, have two fttc lines here but are much slower (18-21 Meg) and no ill effects. So needs some feedback from others with near full speed lines.
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
|
I shall avoid 2 x FTTC then! Not worth to lose speed risky!
|
|
Register (or login) on our website and you will not see this ad.
|
|
|
|
What type of drop wire do you currently have? Are there two twisted pairs?
|
|
|
|
The drop wire is circa 5 years old from memory, it has 2 pairs in it, I presume they are twisted pairs.
The BTOR eng hung the tester on a separate test drop wire dangling from the pole and the max rate drop was a lot better, circa 20M better, up from 55 to 75.
|
|
|
|
Will order a modem to test.
Can't understand how we lost about 25M from the TT line though, all he did was re-terminate a couple of joints on the BT line
|
|
|
|
If that joint fix lets more power through the BT line, you might see an improvement to the BT line but it would/could produce more crosstalk onto the TT line ... so reducing the speed.
Of course, if there has been a lot of errors or disconnections on the TT line in all this messing around, you might have also triggered DLM to intervene; while it is a single line, there is enough spare overhead (20Mbps) to absorb FEC/interleaving and stay full speed ... but perhaps with crosstalk from the second line present, that spare overhead has disappeared, and FEC/interleaving combines to give the reduced speed.
|
|
|
|
Both lines got worse, the BT line went from 57 sync, 69 max to 53 sync 56 max.
He spent the whole day on the job and everything got worse, even though he didn't touch the TT line.
He got the DLM removed from the BT line.
DLM doesn't seem to be on either line as both will sync on their own to 79987.
I've pulled the BT line out at the moment as I don't want the TT DLM to kick in, assuming they have one.
|
|
|
|
DLM on FTTC is provided by Openreach, not the ISP. After a reset, DLM will not kick in until the next day, assuming no major instability is found.
|
|
|
|
Thanks, what I don't want is the DLM to drop speed on the TT line, I've pulled the plug on the BT line for now, I'm going to get a new modem to try in the morning, do you recommend I leave the BT line disconnected?
My TT stats at the moment with the BT modem off are:-
Upstream line rate (kbit/s):
19999
Downstream line rate (kbit/s):
79987
Maximum upstream rate (kbit/s):
31056
Maximum downstream rate (kbit/s):
99260
|
|
|
|
It doesn't really matter. If you find a cure, DLM will automatically raise the line profile within a few days. Are you load-balancing these 2 lines, or are they separate?
|
|
|
The DLM on FTTC is Openreach and cannot be removed. If it has imposed restrictions due to a fault it can be reset by the engineer after they (think they) have fixed it.
TalkTalk do not have control of the DLM, unlike their ADSL2+ LLU which is controlled by their own kit in the exchange.
My broadband basic info/help site - www.robertos.me.uk | Domains,site and mail hosting - Tsohost.
Connection - Plusnet UnLim Fibre (FTTC). Sync ~ 56.6/14.1Mbps @ 600m. - BQM
"Where talent is a dwarf, self-esteem is a giant." - Jean-Antoine Petit-Senn.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Allergy information: This post was manufactured in an environment where nuts are present. It may include traces of understatement, litotes and humour.
|
|
|
|
The 2 lines go into a Prosafe SRX5308 Quad WAN VPN Firewall which load balances.
I should probably change both the supplied ISP modem/routers as neither can be put into modem bridge mode, so won't pass the WAN IPs through.
Before on ADSL I used to use 2 bipac modems in half bridge mode which worked a treat.
|
|
|
One presumes the drop wire to the home is twisted pair and not very very old untwisted.
The old non-twisted stuff was only ever single pair, so both lines cannot be through the same dropwire.
|
|
|
|
Another thought on DLM, for a few weeks the BT modem was reporting 57M sync but 69M max line rate, shouldn't DLM have raised the sync to nearer the max?
Before todays visit, the TT modem was synced at 78 with a max of 79 when the BT modem was on as well.
|
|
|
|
Sounds like your visitor has broken something. You'll probably need another visit.
The Openreach modem, plenty available on Ebay, will pass the WAN IP straight through. 2 of those perhaps? If you get the HG612 v3B, you can unlock it and configure it greatly.
|
|
|
Yes but I don't really know what he could have broken, he says he only re-terminated a couple of the junctions on the BT line and didn't touch the other.
I can only assume there is a chronic crosstalk problem in the drop wire but I can't work out how he's made it worse.
He is coming back Tuesday, think he said he would bring or talk to his "coach", think he was quite new to fibre but was doing his best.
At the start of the day we had syncs of 80 + 57, after the *fix* we had syncs of 56 + 53 
|
|
|
|
Maybe he caused some instability during the fix and it's just DLM applying interleaving or capping? Can you get any stats from either modem?
|
|
|
|
Can't get any real stats, no telnet on the HG635 it seems well locked down, these are the basic stats with the Homehub switched off on the other line:-
DSL synchronization status:
Up
Connection status:
Showtime
Upstream line rate (kbit/s):
19999
Downstream line rate (kbit/s):
79987
Maximum upstream rate (kbit/s):
31084
Maximum downstream rate (kbit/s):
99391
Upstream noise safety coefficient (dB):
15.7
Downstream noise safety coefficient (dB):
11.5
Upstream interleave depth:
0
Downstream interleave depth:
0
Line standard:
VDSL
Upstream line attenuation (dB):
20.2
Downstream line attenuation (dB):
10.9
Upstream output power (dBm):
6.8
Downstream output power (dBm):
13.7
Channel type:
None
I've ordered a couple BiPac 8800NLs an a morning delivery tomorrow, should be able to get full stats from them.
|
|
|
|
No issues here - I have 2x 80/20 lines, one from Sky, and one via my work - both use the proper OR modem to terminate, both have no issue running at full 80/20.
I'd suggest the "all in one" routers those providers like aren't great... and aren't coping particularly well with the interference/crosstalk.
|
|
|
|
Those stats for sync and noise margin are really good. So what happens when the Homehub is switched on? The noise margin plummets? The throughput dies?
|
|
|
|
Thanks for the info, good to know it should work ok.
Couple of questions, are both you lines in the same drop wire or 2 separate drop wires?
Which OR modems, HG612 or ECI
I've ordered 2 modems for tomorrow but don't mind investing in a couple of OR ones to try as well, just want to get it sorted, lost almost a whole day today baby sitting the BTOR engineer.
|
|
|
|
Yes, before todays *fix* the TT modem would still sync at 78M and show a max rate of 79M, so circa 20M was lost from the max rate but no real impact on sync. SNR dropped to 6dB.
Now if I switch on the BT HH5, the TT sync drops to about 56M 6dB SNR and max about the same, so 40+M lost from the max rate.
Also if I switch on the BT HH now the initial effect as it syncs is a complete TT disconnect, SNR goes to 0 and even negative followed shortly after by a resync to the slower speed.
The BTOR test kit also had the same effect so I can't blame it on the HH5, however when we simulated a second drop wire hanging off the pole the sync still dropped but to 75M not 56M, still worse than when he started though.
Somehow he seems to have increased the high frequency coupling/crosstalk between the 2 pairs but says he done nothing except re-terminate some joints on the BT line between the house and the cabinet.
|
|
|
|
He's probably got a blob of solder bridging a gap or something. Definitely another fix required.
|
|
|
|
I think they only use horrible crimps
|
|
|
DLM doesn't seem to be on either line as both will sync on their own to 79987.
DLM can intervene on a line, and it can still result in a sync speed of ~80000. For that to happen, the pre-intervention maximum speed would probably have to be around 88Mbps or more.
Another thought on DLM, for a few weeks the BT modem was reporting 57M sync but 69M max line rate, shouldn't DLM have raised the sync to nearer the max?
If DLM has intervened, then the actual speed will always be less than the maximum.
A quirk in the algorithm (of the HG612 modem) that estimates the maximum means that it over-estimates the maximum value when DLM has intervened. Perhaps your equipment does the same... If so, then when DLM removes the intervention in the case you give, the sync speed is likely to increase to around 63Mbps, while the max would drop to around 63Mbps.
My line (on an HG612) synced at around 78Mbps just prior to DLM intervention, and a matching maximum. It now comes up with an actual speed of about 72Mbps but the maximum increased to 84Mbps.
|
|
|
|
I see what you mean, repeated throughput tests on a wired connection are only giving 55M on the TT line which says it synced at 79987. Throughput before was always well over 70M, I wonder if the sync rate shown is even correct?
Maybe DLM is limiting throughput after todays playing? Its a right old mess now, wish I'd left it alone but BT said it must be a fault so decided to send out BTOR.
|
|
|
|
Hi,
I had my two lines (same drop-wire) upgrade to Zen Fibre two weeks ago, both lines sync at 79999 19999 and both have a throughput of 74 Mbps using two Openreach HG612's (Huawei Cab) connected to a Draytek Vigor2925n+ dual wan router
I am very pleased with the performance as prior to Fibre I had 2 Meg on one line and 2.8 on the other!
Paul.
|
|
|
This is typical of crosstalk in my experience. Especially if the distance to the cabinet is only a little more than good enough for 80mbit/s .
I get the exact same thing with 2 neighbours 1 on the same phone pole. if the one on the same pole turn of their modem, my speed goes up to 74-80Mbps, from 56mbps, another further up the street if their's is off to I get 80Mbps with a 8-9db noise margin.
If your lucky enough to have 2 80Mbps line you must be pretty close to the cab, or have 2 lines in different cable clusters.
_________________________________________
now on BE Pro
|
|
|
I dont know why people are going on about DLM.
I dont think your lines are interleaved, to the point the FEE is wiping out 40mbit.
They definitely not banded as banding doesnt affect attainable rate.
Looks like a clear case of crosstalk to me.
Now normally openreach will say "tough luck" on crosstalk but because you own both lines and can prove it to a visiting engineer I guess is why they trying to fix it.
Edited by Chrysalis (Sat 23-Aug-14 07:05:08)
|
|
|
Both on the same drop, both sockets next to each other. Both are ECI.
If you were just seeing a drop in performance generally over time, I'd put that down to crosstalk if your line was marginal that it could only just achieve the 80/20 speed (mine has sufficient margin that it could go significantly faster if it was an option).
My hunch is that the piles of brown stuff like the Home Hub are introducing lots of interference to cause such a major drop. I have nothing good to say about the kit BT chuck out - the wireless is pathetic, the sync capability poor, the functionality laughable. Total garbage but sadly BT continue to get the majority of the broadband business.
I've yet to see any of our customers suffer significant drops in speed.
Edited by therioman (Sat 23-Aug-14 08:33:57)
|
|
|
|
I only went with BT because the kids wanted BT Sport, after factoring in the cost of that per month vs free with BT over 18 months, plus I got £120 quidco cash back and £60 Sainsburys card, it seemed a good deal, Sky weren't interested in giving me a good deal, preferred to lose me!
I thought BTOR would install their own modems but alas not. First thing I had to do was disable all the poop on the HH5 and HG635, all I need is modems, the wifi and firewalls are all disabled.
My sockets are separated by about 13m, drop wire is about 33m to split then 1m to BT socket and 12m to TT socket.
|
|
|
|
Hi
It is crosstalk and nothing you can do about it unless you can physically separate the pairs. Your pairs are very likely running together from your home all the way to the cabinet, and so is worse case scenario for cross talk.
Any changes, and yes improvements to the lines, could see speed reduce a little, for example the DLM backing off the power because the line has less errors and it is detecting crosstalk which it is trying to mitigate. Or it could just be while the work was being carried out some one else connected up to VDSL, causing more crosstalk.
It is an industry known problem using telephone cable that was never designed to carry these frequencies, you will go mad trying to fix it. Best hope is vectoring, even then the improvement might only be slight in worse case scenarios like this.
Regards
Phil
|
|
|
|
Bipac 8800NL Hooked up on TT, think DLM is limiting as attainable shows 102084 but sync is 53980 19.8dB SNR Stats below, first number download, second upload, will now hook another 8800NL onto the BT line and see if it drops the same.
Line Coding (Trellis) On On
SNR Margin (dB) 19.8 11.9
Attenuation (dB) 15.9 0.0
Output Power (dBm) 13.9 6.8
Attainable Rate (Kbps) 102084 32678
Rate (Kbps) 53980 19999
B (# of bytes in Mux Data Frame) 239 47
M (# of Mux Data Frames in an RS codeword) 1 1
T (# of Mux Data Frames in an OH sub-frame) 64 52
R (# of redundancy bytes in the RS codeword) 0 16
S (# of data symbols over which the RS code word spans) 0.1415 0.0763
L (# of bits transmitted in each data symbol) 13568 6714
D (interleaver depth) 1 421
I (interleaver block size in bytes) 240 64
N (RS codeword size) 240 64
Delay (msec) 0 8
INP (DMT symbol) 0.00 4.00
OH Frames 0 0
OH Frame Errors 0 0
RS Words 0 3858496
RS Correctable Errors 0 43
RS Uncorrectable Errors 0 0
HEC Errors 0 0
OCD Errors 0 0
LCD Errors 0 0
Total Cells 296577803 0
Data Cells 4120974 0
Bit Errors 0 0
Total ES 0 0
Total SES 0 0
Total UAS 33 33
|
|
|
|
Yes, DLM it is.
|
|
|
|
Both lines have now got 8800NL modems on, things are similar as expected but at least I can see some stats.
The BT line max rate tops out at 87M, TT 102M even though TT is 13m longer.
The whole tone spectrum is about 5dB worse SNR until you hit tone 3400 where the BT line rapidly rolls off to 15dB SNR at the top end, the TT line still has about 28dB SNR at the highest tones.
The high frequency performance of the BT line is quite a bit worse, should 2 lines differ in performance this much?
Regarding the crosstalk, nothing I can do, BTOR have made it much worse, max rate drops 44M on the TT line 102M to 58M and the BT line drops 31M from 87M to 56M
BTOR coming back Tuesday, what should I ask them to do as I don't think they know how they have made it worse. I am thinking insist on a second drop wire as that seemed to lose 20M ish off the speed??
|
|
|
|
Sounds like a plan. He must have done something wrong...
|
|
|
I've done some more analysis on the 2 lines, this is with only one modem on at a time, both lines using bipac 8800NLs with latest firmware, should 2 phone lines differ as much as shown in the graph in the link below, the Talktalk line is longer but better across the whole spectrum, even more so towards the higher frequencies, graph here:-
SNR Per Tone
|
|
|
Two lines may share the same drop wire to the home, but be in different bundles and even different routes back to the cabinet. Though with FTTC the chance of this is lower than it was with ADSL services.
The trick is the to get your VDSL up and running and rebooted and check the same graphs ASAP after a power cut or during a power cut if you can. i.e. when cross talk from other lines is likely to be at its lowest.
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
I'm in a very small village and from what BTOR told me there's not many fibre conns yet, I had connections 1 & 2 and they had done 2 others I think.
I hooked a spectrum analyser onto the BT line, image at link below, blue trace is how the line looks from 100kHz to 18MHz with both modems off, when I switch the Talktalk line modem on you can see the level of coupling especially in the block between about 8.5 and 12MHz
VDSL Crosstalk
|
|
|
Yes, DLM it is.
Actually, it looks like DLM has intervened upstream only (with INP=4.0 and delay=8). Downstream, both parameters look to be zero.
The modems have negotiated parameter R to be zero downstream, so none of the bandwidth has been stolen in that direction, whereas upstream, R/N = 16/64, showing that 25% is being used as the overhead for FEC in that direction. Luckily there is plenty of room.
Interleaving depth D is also set to 1 downstream, so it is turned off there.
Sometimes the modems negotiate upstream FEC to be turned on alone (ie without interleaving), even when DLM has not asked for intervention (ie delay=0 & INP=0.0);
|
|
|
|
DLM has capped the speed to raise the SNR margin.
|
|
|
Bipac 8800NL Hooked up on TT, think DLM is limiting as attainable shows 102084 but sync is 53980 19.8dB SNR Stats below,
Those numbers suggest that, on this occasion, the TT modem was originally synced while the BT modem was turned on (ie crosstalk was present at the time of sync).
Then the BT modem was turned off, but the TT modem was not re-synced, when the stats were read.
The result would be that the SNRM values jump a lot (19.8 is pretty high), and the calculated maximum speed would increase (as is 100Mbps+) as a consequence.
If this is NOT what happened - if the TT modem were synced while the BT modem was turned off - the it suggests DLM has intervened with banding. Your top speed has been limited "forever".
|
|
|
Maybe. Depends on circumstances, as described on other post.
But it is rare for DLM to intervene with banding alone.
Edit: and rare to use such a non-rounded number for the speed.
Edited by deleted (Sun 24-Aug-14 18:15:14)
|
|
|
|
The talktalk line will only sync to 53980 even when syncing with the BT modem off, the modems have been on and off so many times I guess the DLM has decided to join in, although the BT modem on its own will still sync at 79 but I think he reset the DLM on that line last Friday.
I'll be asking BTOR to reset both lines DLM when they've done what ever they plan next, hence I am keen to take all measurements and do whatever tests I can do before Tuesday.
|
|
|
Your top speed has been limited "forever". Or 10 days, which can seem like "forever" 
|
|
|
Edit: and rare to use such a non-rounded number for the speed.
I'm wrong on this, perhaps...
Looking in this NICC testing document, there are banded profiles with an upper bound of 54Mbps. The format of a profile is described on page 37; earlier on page 29 they happen to include one with an upper bound of 54Mbps (it is a re-transmission profile, so I guess it isn't active yet)
However, the profiles in that document suggest that banding of upstream happens at the same time as the downstream. If that suggestion is true, rather than being a small snapshot of a huge set of profiles, then 54/20 wouldn't be possible.
BT's own SIN 498 for testing show the same description for profiles, but don't happen to include one with 54Mbps.
|
|
|
|
Then it is indeed DLM banding ... and a new combination to chalk up to experience.
Perhaps this is indeed caused by the disconnection counter rather than the error trigger!
|
|
|
The simplest way to find out what is going on is to get TT/BT to do a GEA Circuit test (like here http://community.plus.net/forum/index.php?action=dla... which will show if there is any detected crosstalk or potential wiring issues.
The thresholds for DLM kicking in for retrains (disconnections/reconnections) alone are quite high, but it depends on which DLM profiles TT/BT are using:
Speed - 20 retrains in 24 hrs
Standard - 10 retrains in 24 hrs
Stable - 5 retrains in 24 hrs
If you're line is erroring also (beyond the thresholds), then the retrain count is lower:
Speed - 12.3 retrains in 24 hrs
Standard - 7.1 retrains in 24 hrs
Stable - 4.6 retrains in 24 hrs
|
|
|
|
TT line DLM seems to have rebanded overnight, now 73971/19999 SNR 13.9/15.7
BT modem has been left powered off.
|
|
|
TT line DLM seems to have rebanded overnight, now 73971/19999 SNR 13.9/15.7
BT modem has been left powered off.
Maybe interleaved is ON!
|
|
|
|
Interleaver Depth showing 1 on up and down
|
|
|
Interleaver Depth showing 1 on up and down
must be crosstalk then!
|
|
|
|
I will take some stats from before one of my two lines is deactivated and then check the remaining one after, see what impact it has, and will post here.
|
|
|
8.5 to 12 MHz is used for UPSTREAM and as you are measuring close to the upstream source that level does not surprise me.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
M H C
taurus excreta cerebrum vincit
|
|
|
is it possible engineer made a report then banding was applied to stop the customer stating speed changes with only one line?
this wouldnt surprise me, because a customer showing an engineer a line speed up when the other is disconnected would make the engineer uncomfortable knowing its proven.
|
|
|
|
I don't think the engineer got any banding applied, he got the banding removed from the BT line as that was showing 69M attainable with 57M sync. This line on its own will now sync 79987 with attainable 82.9M, the attainable is 5M down today, its wet, wonder if there is a water issue on this line.
He did admit to me he was noob at fibre but was very helpful and I spent I good few hours going through it with him.
I just don't know what they are going to expect me to accept now, he found a line fault, claims to have fixed it but doesn't know how....
We started out at syncs of 79M and 57M, I don't want to accept anything less.
|
|
|
the attainable is 5M down today, its wet, wonder if there is a water issue on this line.
It could be. Or perhaps another customer has ordered FTTC, and their line is adding to the crosstalk picture.
I guess you'll find out when things have dried out.
I just don't know what they are going to expect me to accept now
...
We started out at syncs of 79M and 57M, I don't want to accept anything less.
The problem is that 3rd parties will order FTTC, and some of those will cause some level of crosstalk. Unknown amounts, and random - it depends how coupled the wires are. BT know this, so never go chasing faults reporting small drops in speed. And by small, their threshold seems to be 20% (or is it 25%?)
The cabinet was quite empty at first, right? Was it recently installed? If so, I guess you can expect take-up to rise reasonably soon. In such a dynamic environment, you might *never* be able to get these speeds back, even if your line gets "fixed."
|
|
|
|
BTOR installed a second drop wire which helped quite a bit, my crosstalk problem is between my 2 connections rather than anyone elses line, both sync at 79997 on their own.
I guess my only hope for both at full rate is that vectoring is used in the future.
Currently the BT line shows 63M sync and attainable.
The TT line now shows 79984 attainable but only synced at 64052, hoping DLM will come back up on this line.
The TT line has now turned on interleaving, not sure why.
|
|
|
BT know this, so never go chasing faults reporting small drops in speed. And by small, their threshold seems to be 20% (or is it 25%?)
If my experience is anything to go by it they aren't interested if there's a 30% drop. Dropped from 72 Mbps to 50 Mbps in one go back in April and is now 47 Mbps and still falling.
|
Pages in this thread: 
Print Thread
deleted
