Do I need to raise a fault?

Sorry in advance for the long post.

I seem to be on a banded profile that is steadily reducing my sync speed because of occasional losses of sync. I raised a ticket to ask for the banding to be removed but have been pointed towards the fault reporting route but I don't think there really is a fault (at least not directly on the line) apart from the reaction of DLM to the losses of sync.

I strongly suspect it is due to interference from one (or more) of the electric fence units in the field my line passes over on its way to the green cabinet. between losses of sync my line is generally pretty stable (see info below - total time should probably be 4 days not 1, I believe it is a router bug that never show more than 1 day), very low ES figures. The problem is the LOS figures - 6 shown below - as these will have been in one, or possibly 2, bursts as the ADSL signal gets wiped out for a few minutes (even an SNR margin of 25-30dB doesn't stop the router losing sync - if I manage to catch it happening there are usually at least 3 or 4 attempts to resync, presumably til the interference dies away) which is what I believe is driving down the banded profiles. I was syncing at 3070 for quite a while (not ideal but not too bad) but it has recently sunk to the current 1050 (via a short spell at around 2268 on the way).

As the reduced sync/raised margin (SNR currently running at 29-30 at night, up to 30-31 in the day) doesn't stop the loss of sync I would prefer to have the banding removed - permanently if possible. Can this be done?

Thanks for your help.

> adsl info --stats
adsl: ADSL driver and PHY status
Status: Showtime  Channel: FAST, Upstream rate = 440 Kbps, Downstream rate = 115
0 Kbps
Link Power State: L0
Mode:                   ADSL2+
Channel:                Fast
Trellis:                ON
Line Status:            No Defect
Training Status:        Showtime
                Down            Up
SNR (dB):       29.6            26.9
Attn(dB):       41.0            22.1
Pwr(dBm):       19.4            12.3
Max(Kbps):      7540            888
Rate (Kbps):    1150            440
                        G.dmt framing
K:              16(0)           56
R:              16              8
S:              4               1
D:              16              8
                        ADSL2 framing
MSGc:           69              12
B:              15              55
M:              4               1
T:              2               1
R:              16              8
S:              1.7250          4.0000
L:              371             128
D:              16              8
                        Counters
SF:             9943254         9943252
SFErr:          0               0
RS:             372872060               676141136
RSCorr:         0               0
RSUnCorr:       0               0

HEC:            0               0
OCD:            0               0
LCD:            0               0
Total Cells:    436190175               694493
Data Cells:     13864645                8345
Drop Cells:     0
Bit Errors:     0               0

ES:             8               0
SES:            6               0
UAS:            475             0
Total time = 1 days 4 hours 6 min 45 sec
SF  = 22195918
CRC = 1
LOS = 6
LOF = 0
ES  = 8
Latest 1 day time = 4 hours 6 min 45 sec
SF  = 915504
CRC = 0
LOS = 0
LOF = 0
ES  = 0
Latest 15 minutes time = 6 min 45 sec
SF  = 25099
CRC = 0
LOS = 0
LOF = 0
ES  = 0
Previous 15 minutes time = 15 min 0 sec
SF  = 55639
CRC = 0
LOS = 0
LOF = 0
ES  = 0
Previous 1 day time = 24 hours 0 sec
SF  = 5342367
CRC = 0
LOS = 0
LOF = 0
ES  = 0
15 minutes interval [-30 min to -15 min] time = 15 min 0 sec
SF  = 55638
CRC = 0
LOS = 0
LOF = 0
ES  = 0
15 minutes interval [-45 min to -30 min] time = 15 min 0 sec
SF  = 55639
CRC = 0
LOS = 0
LOF = 0
ES  = 0
15 minutes interval [-60 min to -45 min] time = 15 min 0 sec
SF  = 55700
CRC = 0
LOS = 0
LOF = 0
ES  = 0

yes i would raise a fault. i think you will find it is not the fence.

you can check this yourself if you walk the route the fence runs on. firstly check that there is nothing touching the conductor ie branches etc etc. also check any joins are in good order and importantly check the earths of the unit itself, it should have 3 earth rods in parrallel driven into the ground. google it as i cant mind exactly.

if it was the fence its likely you would see issues with every pulse.


more likely you may have an issue somewhere with the line.

BT don't work over xmas. Service resume 3rd January 2011.

Edited by adslmax (Tue 28-Dec-10 10:38:41)

Thanks.

I realised that was probably the case.

One reason I asked was that I started down the faults path until I got to the try a different router. (I've got a spare one but it isn't set up at the moment so I need to find time to do that without disconnecting the rest of the family from the t'Internet. ) That was the point at which I started to wonder if the fault route was worth it & if anything useful was likely to come out of it - I only seem to get less than 5 minutes of problems perhaps every other day. The problem is that it seems to be enough for BT's DLM to be pushing my sync speed lower & lower.

Perhaps it is time to escape?

I suspect it is when one of the units (I can see at least 3 close to the - fenced - footpath that runs along under the route of the phone line) gets moved or put into action/ tested possibly combined with the weather - it seems to be worse in very dry weather. The reason I started to suspect the fences was one time when I briefly lost both ADSL & dialtone and instead got a ~1-2Hz ticking sound. The weather had been very dry but about the same time started raining & my problem went away so my suspicion was a badly earthed electric fence because of very dry ground (or it got switched on before it was properly installed).

The biggest problem in trying to track down the source is that it doesn't seem to last more than a few minutes and doesn't seem to happen at a regular time, not even every day. That's really the main reason for wondering if it is worth going down the fault route or if anything can be done to stop DLM managing my line to a crawl without raising a fault (other than migrating to an LLU ISP).

In reply to a post by adslmax:

BT don't work over xmas. Service resume 3rd January 2011.

The response to this rhymes with "Rowlocks". Many BT help desks are open 24/7, although it can be difficult to get an engineer visit on a Bank Holiday, and would certainly be expensive if no BT fault is found.

Hi bluetroll,

Even if it is very intermittent it's still something we'd be happy to take a look at for you - whether as a fault or if we can try to get you on the DLM trial (assuming ADSL2 is available) which would let us set a profile for your line and stop the DLM changes happening.

Hi

If you want to do or suggest anything that may help then please do.

I've responded to my original ticket (#38311020) as I tried a different router (I didn't have one immediately to hand) and that seems to have held a connection since 29 Dec (although I can't get much info out of the router - it doesn't show anything that suggests a reconnect in the "system log"*), although with vastly different reported stats - particularly attenuation showing 51dB now v 41.5dB with old router.

Thanks.
===============
*log shows following lines after reboot but none have been repeated since

12/29/2010 11:17:28 	If(ATM1) PPP connection ok !
12/29/2010 11:17:27 	ATM1 get IP:81.174.xxx.xxx
12/29/2010 11:17:18 	ATM1 start PPP
12/29/2010 11:17:18 	ADSL Media Up !

Your line profile looks like it's dropped to 750 following a fair few disconnections - definitely more than once per day, in December. Unfortunately the lower profile has stopped the disconnections but we can still definitely raise this as a fault for you and get it investigated.

First though, could you test the connection with the router in the master socket or even better the test socket (if there is one behind the master socket front plate) and check if there's any difference in the sync speed shown in the router before and after?

This will give us an idea if the drops were to do with internal wiring - I don't think you'll get any disconnections on the lower profile so we can't monitor for those in the master unfortunately.

If you update the ticket (or post here) with that information I'll get the line checked and get things pushed on for you.

Thanks for your reply.

The connection is in the master socket and there are no extensions wired in - cordless phone base station & router are both plugged into microfilter. I'll move things into the test socket (& swap the cordless for an old corded phone for a few hours) as soon as I've posted this & see if that makes any difference.

When you say more than one disconnection per day is that in a constant stream as my experience suggested that it was less frequent but there would be several reconnection attempts that only held for a second or two all one after another until it finally stuck. So, although there might be say 6 reconnections in a given day they would all be in the space of a couple of minutes with the rest of the day OK.

Can you tell me if there are less errors since 28/29 December as this was when I swapped routers (And lost any access to figures - whether accurate or not as the old router reported very few errors between disconnections via telnet). According to the figures reprted by the routers, the new(er) one is only connecting at an 8-10 SNR margin downstream but with the power backed right down. The old one was reporting a much higher downstream power & 30+ SNR. But they both report vastly different attenuations (41 v 51) & the old router sad it was on fast path where the new one says interleaved. Could they both be right (different chipsets reacting differently?)? Or has one got its reporting knickers in a twist? As I said previously I am not aware of any disconnections (apart from a few in the first few hours caused by settings changes requiring a router restart) since putting the new router online.

I'll copy & paste the info above along with answers to the questions in the recent reply to my ticket & continue things there.