ADSL2+ Speed Drop

Some advice please.

My BB speed (BT line - non LLU) has been steadily dropping since the middle of February. It has gone from over 18 MBS - SNR 3.1 to 11 MBS - SNR 13.8. I do understand that the increased SNR is reason for the speed drop. I have been monitoring my connection with Routerstats to try and see what is going on. There was one day during Feb when my router kept disconnecting and presume this is the reason for the current speed. According to my ISP it is the router which is losing the connection and nothing wrong with the line?? I have subsequently completely reset the router and my connection has remained stable for over 6 days (12 SNR) until yesterday when it disconnected again and yet another drop in the SNR margin to the present 13.8. During this time, the Upstream rate has also been fluctuating from 1265 down to 598 with interleaving being applied and then removed. I have been waiting for the DSLAM?? to kick-in but it does not appear to have done so - except to knock the speed down and not up. Am I being impatient?

Any suggestions/advice would be most appreciated before I go back to my ISP again.

Thanks

Edited by deleted (Wed 13-Mar-13 11:53:07)

The router is most likely reacting to bursts of noise, the provider cannot really tell where these are coming from.

It could be a joint on the line, a solder joint in the router, or a bit of damp in the home causing corroded connections, or as unlinked as your heating boiler now sparking more and the radio interference from this causing problems for the ADSL.

For the DLM to reduce target noise margin you are looking at the system wanting to see you as stable for two weeks.

Before you contact your ISP again can you connect the router to the test socket behind the master NTE5 (if you have a master NTE5?) faceplate (or filtered faceplate?) using a different dangle micro filter (?) and post your full current router stats?

Thanks for the quick response. Unfortunately I do not have a master NTE socket in my property. However, I did replace the filter connecting the router when I reset it and below are my current stats:

Statistics -- xDSL

Mode: ADSL_2plus
Traffic Type: ATM
Status: Up
Link Power State: L0

Downstream Upstream
Line Coding(Trellis): On On
SNR Margin (0.1 dB): 138 88
Attenuation (0.1 dB): 200 104
Output Power (0.1 dBm): 0 128
Attainable Rate (Kbps): 15932 984

Path 0 Path 1
Downstream Upstream Downstream Upstream
Rate (Kbps): 11767 891 0 0

MSGc (# of bytes in overhead channel message): 51 67 0 0
B (# of bytes in Mux Data Frame): 209 27 0 0
M (# of Mux Data Frames in FEC Data Frame): 1 1 0 0
T (Mux Data Frames over sync bytes): 2 1 0 0
R (# of check bytes in FEC Data Frame): 0 0 0 0
S (ratio of FEC over PMD Data Frame length): 0.5696 0.9696 0.0 0.0
L (# of bits in PMD Data Frame): 2949 231 0 0
D (interleaver depth): 1 1 0 0
Delay (msec): 0.14 0.24 0.0 0.0
INP (DMT symbol): 0.0 0.0 0.0 0.0

Super Frames: 6787851 539117 0 0
Super Frame Errors: 0 0 0 0
RS Words: 0 0 0 0
RS Correctable Errors: 0 0 0 0
RS Uncorrectable Errors: 0 0 0 0

HEC Errors: 4294967243 423 0 0
OCD Errors: 52 0 0 0
LCD Errors: 0 0 0 0
Total Cells: 3058772039 1690069 0 0
Data Cells: 5146314 18210 0 0
Bit Errors: 0 0 0 0

Total ES: 689 11
Total SES: 38 0
Total UAS: 159 4294967259

Thanks for yr advice. So really could be anything?? If it is a suspect joint on the line (we're in a rural location here - overhead wires etc) would it not show up as a fault on the line? Central heating boiler - as it's on most of the day wouldn't it keep affecting the router? If it is something to do with the router innards, can try an old Netgear and see what happens but will wait on that in line with the two weeks to see if the DLM does kick in..

It may be the cold weather has allowed some water ingress to a joint outside and resulting in a High Resistance fault.

In theory there are tests for this CIDT is its name, and often is not run and it can detect faults that the traditional tests miss.

For background
http://www.openreach.co.uk/orpg/home/products/llu/ci...

Yes noise affecting the line some where - downstream sync with a 6dB SNRM on a 20dB attenuation should be ~20000Kbps

Your current downstream sync time SNRM may even be 15dB which is, as you suggested, really holding back the sync speed.

As your ISP mentioned, it might possibly be the router causing the problem so can you perhaps try another router?

Well I've taken the plunge and swapped over the router. Back to my old Netgear 834G v5. It is immediately reporting slightly better figures:


System Up Time 00:51:33

Port Status TxPkts RxPkts Collisions Tx B/s Rx B/s Up Time
WAN PPPoA 4409 4229 0 358 886 00:50:55
LAN 100M/Full 9756 10381 0 1422 599 00:51:13
WLAN 11M/54M 9828 7469 0 880 514 00:51:33

ADSL Link Downstream Upstream
Connection Speed 12503 kbps 1279 kbps
Line Attenuation 20.0 db 9.5 db
Noise Margin 12.00 db 5.5 db

I guess time will tell now. Will be interesting to see if it was the other router playing up (TP Link) or problems on the line. Shame really as I am not too keen on the Netgear as I know v5 is not one of the better versions as it does not seem to give as much info as other routers can..

Thanks for the input

That's an amazing upstream sync change.

Give it a day or two with the Netgear and if there is no improvement with the downstream SNRM, ideally down to 6dB, and a sync speed of 20000Kbps then give your ISP a shout. Tell them that you have tried an alternative router, the line is still not performing as it should and insist that you have some tests done on the line.

Good luck

P.S. probably best not to mention that you don't have a NTE5/test socket at this stage unless specifically asked. Also have you tried a quiet line test (dial 17070 with a corded phone.)

Have been patiently waiting for some sort of a change to take place but alas to no avail. I have now had a stable connection with the Netgear for over 8 days now and everything has remained the same.

Have raised the issue with my ISP who have advised that they can see my IP profile has been around 12 for the last few weeks (which I already knew!) and that my throughput speed had been averaging a lot less than 12. They asked that I run the BT speed test then into the further diagnostic mode which I did to enable them to progress the matter further with BT. They also mentioned something called a TAP3 test might be rqd after going into the diagnostic mode but it was not asked for.

Assuming my ISP are trying to sort this for me, what is a reasonable time to wait before asking more questions?

Quiet line test 17070 - Yes done that several times when this all kicked off. Nothing untoward on the line.

Thanks

Noise margin setting drops take 10-14 days stability, then goes down by 3dB (usually with an automatic resync) and starts the next 10-14 days.

Edited by RobertoS (Fri 22-Mar-13 13:33:45)

I have been having similar issues since last Sept/Oct. I ended up getting one disconnect a day (at any time), and each time the SNR went up, IP profile was reduced until I ended up with a profile of 1.5 mbs.

The last few months I have had engineers out, changed routers, swopped wires, re-routed TV aerials cables (clutching at straws) etc.

Finally I found out what was happening from my ISP techy. The DLM was seeing 100's of errors an hour, although at my end on the router didn't see or record as such. Hence, every time I did get a disconnect/resync, the DLM done it stuff and tried to make the line stable (due to the errors it saw) - and each time I just got slower and slower.

Having had enough I ended up getting the connection set to bog standard ADSL, and using tweaks on my router to rx & tx gain + reducing 'bits per bin' to 9, I now have a fairly stable line without any profile being applied if I do get a disconnect.

The thing to note is I now get 100's of CRC and header errors that didn't become apparent until I did get my connection reset to just ADSL (the rx & tx gain tweaks have fixed those - plus reducing bits per bin lowers sync speed which also helps). BT, of course, said there is nothing wrong with my line - but there obviously is.

Nick

Edited by deleted (Fri 22-Mar-13 14:29:59)

But the OP's line has held a connection for 8 days (?) and downstream sync is 12503Kbps, switching to ADSL1 would reduce the speed to 8128Kbps - or less since the downstream SNRM may remain at 12dB during another protracted stabilisation period.

Edited by 4M2 (Fri 22-Mar-13 14:43:34)

Yes, but I am not saying that. What I am saying is IF the DLM is seeing errors (irrelevant if the line doesn't go down and stays connected), then the NEXT disconnect it will apply an SNR increase to make the line stable due to the errors it sees (but you don't!). So you cannot win, no matter how long you stay in sync!

I think the technology is a bit dodgy, to say the least.

Nick

Edited by deleted (Fri 22-Mar-13 14:51:38)

In reply to a post by Lethe:

I think the technology is a bit dodgy, to say the least.

I don't have such problems on partial LLU (SMPF)

It would be nice to be able to log all the people who are getting similar issues that *just* starting happening like this. I have had ADSL for a long time, and my line was always stable, staying in sync for weeks ~ months at a time (with no or little CRC/header errors etc.)... until last Sept/Oct.

Then the OP starts getting similar issues in Feb.

Something got changed somewhere - whether this is to do with BT rollout of fibre or what I dunno, but I feel it has [censored] up a lot of peoples ADSL connections.

Nick

You seem to have missed the point that the line has stabilised since changing the router, with a huge increase in upstream sync taking place immediately.

That implies a fault of some kind on the replaced router.

It follows that there will be no DLM improvement for at least another two days, possibly six. IMO at the moment it is pointless do do anything but wait and see what happens by day 15.
_{Edit - and another typo corrected.}

Edited by RobertoS (Fri 22-Mar-13 18:16:52)

Seems to be particularly the case with Orange (EE or what ever they call themselves) at the moment with upgrades 21CN WBC...

Robert, been there done that etc. After BT enginner called, I sync'ed at 9500 odd - fastest ever. Then it slowly got reduced slower and slower...

I am betting the DLM is seeing loads of errors, so no matter how long you stay in sync, the next disconnect it will try to fix those errors.

Nick

Thanks for the advice. Looks like I'll just have to be patient and wait and see what happens in the next week or so. I suppose looking on the bright side, at least I'm not getting any disconnections at the moment and there are plenty of other people out there who would love to synch at the rate I'm moaning about!!

It does feel like an age the first time you have to wait for it. After that it doesn't seem noticeable .

If it doesn't do anything by the 16th morning, then try a single reconnection yourself. If that doesn't do it, (as just occasionally the auto-resync doesn't happen even though the setting has been lowered), then it isn't going to happen and you need to go back to chivvying your ISP.

If it does auto-resync though, it will probably do it at night. In which case a single reconnection by you well inside daylight hours should improve the new sync.

Interesting to hear other peoples' similar problems and realise you're not alone.

My ISP obviously picked up on something to suggest completely resetting the router which did work for a couple of days and then a further disconnection swiftly followed by an increase in the SNR. Replaced router back to my old Netgear and my connection has remained stable (9 days) with hardly any errors at all.

If it does turn out to be the TP Link 300 mbps wireless N ADSL2+ modem router which is less than 9 months old, I'm not very impressed. It was great while it was working OK but obviously not very reliable. Will have to look around for a replacement when this all settles down. Another Netgear maybe??

If I'm right then a replacement under warranty? Or would you rather get a refund and try something else?

If I'm wrong, as Lethe thinks, then it still looks like a replacement/refund of the TP-Link to me.

I tried several routers - including buying a new one. All end up the same.

There is something going on with the lines IMHO after this roll out of fibre.

And please remember what I said - no matter how long the router stays in sync, if the DLM stills sees errors, then the next disconnect/resync YOU will be lowered.

Nick

Your problem and maggiej's are extremely unlikely to be the same.

Hers is common. Yours almost unique.

The old router seems not to have any issues. Just wait and see what happens please.

Have been patiently monitoring my connection via RouterStats and connection has remained stable . However, after 9 days I notice the PPPoA session dropped for 15 seconds or so with the uptime restarting.. However, as far as I can see from RouterStats there was no disconnection recorded ?? - all graphs logging everything OK, . Was this a 'disconnection' and I'm back to square one??

I ask this as I've heard from my ISP who advise that there is nothing wrong on the line and everything is OK according to the BT speed test. I just want to be pre-armed for any comments they may have to make when I eventually go back to them. I'm assuming that they will say my connection 'disconnected' yesterday. Am I right??

Thanks

Edited by deleted (Thu 28-Mar-13 08:02:10)

That is just the polling interval, so maybe RS didn't 'see' the disconnection.

Say it polls every minute, 1..00,2.00,3.00,4.00 etc. You get at drop at 3.15 and back up at 3.55 - then all RS sees is the measure at 3.00 and 4.00 ~ nothing in between!

Nick

Did either or both the connection speed and noise margin on the graph change at the time of the PPPoA loss? What sampling interval do you have set?

In reply to a post by maggiej:

Have been patiently monitoring my connection via RouterStats and connection has remained stable . However, after 9 days I notice the PPPoA session dropped for 15 seconds or so with the uptime restarting.. However, as far as I can see from RouterStats there was no disconnection recorded ?? - all graphs logging everything OK, . Was this a 'disconnection' and I'm back to square one??

Loss of Internet is not uncommon whilst ADSL remains sync'd to the exchange - generally unnoticed unless one is downloading or uploading without a resume function. A very short drop of the internet (not ADSL) is often not registered by the ISP.

Both the connection speed and SNR remained completely unchanged at the time - in fact they have been constant right the way thru since changing to the Netgear. The sampling interval set on RouterStats is 20 seconds. In fact in checking the Netgear router log, there is no record of any disconnection at the time of the PPPoA loss/restart?? although it is reflected in the Show Statistics/connection status boxes. So I'm none the wiser as to whether it will be classed as a disconnection by my ISP.

Wierd.

I always used to set Routerstats to 5-second sampling. Twenty seconds is useful for seeing trends on a normal line, but poor for spotting what is really going on. A disconnection and reconnection can easily complete in less, so not show anything on the graph.

That's why I asked about the values around the time. As they stayed the same, I agree it is likely there was no disconnection.

All very strange.

Well things have deteriorated today.

Following on from my posting the other day I had a further dropped PPPoA connection for in excess of 3 mins which this time resulted in the router disconnecting completely. I ended up having to reconnect the power and reboot the router. So it looks as though I'm back to square one. An hour later there was another drop in the PPPoA connection (no disconnection this time) I've changed the setting to 6 secs and it still has not picked up anything which would indicate the reason fhe the drop?? so I am very confused as to why this is happening.

Is it possible it could be something at my ISP's end of things? The reason why I'm asking is that when using my other router (which may or may not have been faulty) leading up to this episode, the syslog reported that the primary DNS server was down - switching to secondary - then back to primary server before being able to connect. This happened lots of times with messages from the Exchange reducing the connection speed etc. End result where I find myself now.

Perhaps someone can advise ??

Thanks

DNS would maybe stop internet traffic, but should not cause an ADSL connection to drop.

Do you have the router plugged into a filter/spike/surge extension? If not, try that, and also try to make sure not much else is using the same wall socket.

You could be seeing noise from the mains when something gets switched on (fridge et al).

Nick

Thanks for your advices. I thought I'd mention about the DNS problems just so I could rule them out and not appear too stupid when I go back to my ISP!!

Yes, the router is plugged into a surge protected extension socked (x 6) which in turn is plugged into a double wall socket which has a further extension (x 4) which supplies my other computer, monitor and lamp. Have had this set up for a long time so don't think it can be anything to do with it (OH is an electrician) .

I'm presuming RouterStats would register some kind of noise spike or whatever that could account for the disconnections?? but nothing is being shown.

In reply to a post by maggiej:

Following on from my posting the other day I had a further dropped PPPoA connection for in excess of 3 mins which this time resulted in the router disconnecting completely. I ended up having to reconnect the power and reboot the router.

SNRM would have probably dropped to zero when you lost the ADSL connection.

At disconnections the Routerstats sync graph goes to zero and stays there until the new sync is established. The noise margin graph is similar but may go negative just before the disconnection.

If on 6-second sampling neither is deviating at all when you lose the internet, it's not a line disconnection. To me that would sound like the ISP end, but then that in turn wouldn't cause the re-syncs you do get.

Typical!! for some reason there is no image of the graphs in that particular time - I don't know why?? As such I can only go by the RouterStats Notepad figures at the time of the disconnection:

Fri 29 Mar 2013 12:36:55 80.175.14.216 10.8 5.0 12555 1279 5.0 10.75 ppp0=1day20:02:27
Fri 29 Mar 2013 12:37:15 0.0.0.0 10.8 5.0 12555 1279 5.0 10.75 ppp0=1day20:02:27
Fri 29 Mar 2013 12:37:35 0.0.0.0 10.6 5.5 12555 1279 5.5 10.65 ppp0=1day20:02:27
Fri 29 Mar 2013 12:37:55 0.0.0.0 10.7 5.0 12555 1279 5.0 10.70 ppp0=1day20:02:27
Fri 29 Mar 2013 12:38:15 0.0.0.0 10.7 5.0 12555 1279 5.0 10.70 ppp0=1day20:02:27
Fri 29 Mar 2013 12:38:35 0.0.0.0 10.7 5.0 12555 1279 5.0 10.70 ppp0=1day20:02:27
Fri 29 Mar 2013 12:38:55 0.0.0.0 10.7 5.0 12555 1279 5.0 10.70 ppp0=1day20:02:27
Fri 29 Mar 2013 12:39:15 0.0.0.0 10.7 5.0 12555 1279 5.0 10.70 ppp0=1day20:02:27
Fri 29 Mar 2013 12:52:53 80.175.14.216 12.0 6.5 11952 1271

No drops in the SNR ?? but definitely disconnected at this time (red light - no internet).
The interval between the penultimate reading was while I checked everything out, disconnected power followed by rebooting.

The later PPPoA drop an hour of so later was recorded on the graph and did go to zero for one reading as you have said but I did not lose the internet.

Thanks

In reply to a post by maggiej:

Fri 29 Mar 2013 12:36:55 80.175.14.216 10.8 5.0 12555 1279 5.0 10.75 ppp0=1day20:02:27
Fri 29 Mar 2013 12:37:15 0.0.0.0 10.8 5.0 12555 1279 5.0 10.75 ppp0=1day20:02:27
Fri 29 Mar 2013 12:37:35 0.0.0.0 10.6 5.5 12555 1279 5.5 10.65 ppp0=1day20:02:27

IP address disappeared and PPP session timer has stopped. Dropped by other end perhaps ?

After the reboot sync time downstream SNRM is 12dB and sync speed 11952Kbps - looks like you have lost 600Kbps due to an increase in SNRM. Possibly you had a previous sync time margin of ~9dB and a 12555Kbps sync speed if I'm reading the numbers correctly.

BTW. dropping the DSL connection and PPP link down are two very different things.

At a guess, I would say there are intermittent noise events affecting your line that are causing these issues.

Edited by 4M2 (Fri 29-Mar-13 21:51:17)

Thanks for the input and the clarification on the ' DSL connection and PPP link down' being completely separate. As I am a 'non techie' female I don't really know about these things ! but I do try and believe the only way to learn is to ask others who know far more than I do. At least when I speak with my ISP I will be in a better position to understand what they say and query if necessary.

Will just bide my time now with the aid of RouterStats and try and see if there is anything in my property causing the possible noise issue - if not back to my ISP.

Thanks to everyone else for their responses and help

On a couple of occasions I've lost the internet (PPP link down) for several hours but the router has retained the DSL connection and I have remained sync'd to the exchange. Sometimes rebooting the router can start a new PPP session (usually I would give it 30 minutes) and other times the internet refuses to link for a considerable amount time. Seems to be just one of those things that can mysteriously happen from time to time.

The fact that you don't have access to a master NTE5 box and test socket does tend to make your situation a little tricky and hence it is impossible to rule out problems with the premises wiring. If you had an engineer visit and she/he found a fault exchange side of the test socket then we could fairly safely say that you would not be charged for any diagnosis/repair. However unfortunately this is not the case in your situation and if a problem was found within your premises then you may be liable for any costs.

I would still request that the ISP run tests on your line though and it is just possible that a remote fault may be discovered that way...

In reply to a post by MrSaffron:

For the DLM to reduce target noise margin you are looking at the system wanting to see you as stable for two weeks.

I was wondering what the maximum reduce target noise margin will be ? I think it will be 3.0dB no more than that !

Edited by adslmax (Mon 01-Apr-13 22:57:59)

In reply to a post by 4M2:

On a couple of occasions I've lost the internet (PPP link down) for several hours but the router has retained the DSL connection and I have remained sync'd to the exchange. Sometimes rebooting the router can start a new PPP session (usually I would give it 30 minutes) and other times the internet refuses to link for a considerable amount time. Seems to be just one of those things that can mysteriously happen from time to time.

Well I'm back. Yet another loss of connection today. Checking the RouterStats log it shows that I did not have an IP address - I suppose hence the disconnection. I can see the RX Noise rose from 10.9 to 95 db, TX Noise from 4.5 to 12.37 db and stayed that way for over 30 mins until I went into the Netgear settings and just pressed reconnect What is really strange is that whilst it was disconnected (red light) RouterStats did not show any drop in synch on the graphs during this time!!! with the PPPoA session timer continuing on from when it disconnected???

I don't know if this is relevant or not but at the time it disconnected the Netgear reported the time in its log as Jan 1 2003. I know Negears do lose their time settings from time to time can be rectified by a resynch to the time server. I did this once was up and running and sure enough it showed my fresh PPPoA session . I apologise in advance if this a stupid remark but I'm a firm believer in if you don't understand - ask!

Spoke with my ISP the other day and told them about the dropped PPP sessions but with no disconnection and they could not understand this as they said to lose the PPP would only be as a result in a loss in connection and subsequent noise on the line. Obviously what has happened today is as they have said but does not explain the other occasions. Anyhow, they have said they will reset my line if I wish so I'm just going to wait a bit longer and see if anything else happens.

Edited by deleted (Sun 07-Apr-13 19:44:06)

In reply to a post by maggiej:

they said to lose the PPP would only be as a result in a loss in connection and subsequent noise on the line.

I can't help with the problem, else I would, but that is bilge.

Since when did an effect (disconnection) trigger its cause (noise)?

How come many routers have the facility to stop the PPP session and start a new one without loss of sync?

Thinking a little further, ISPs are well known for not seeing disconnections on their logs as the re-sync can complete before they have detected the loss of connection. In which case the PPP session continues - wrongly.

In Routerstats the expected giveaway is a change in the sync speed.

Another symptom on BT Wholesale links is that if the sync changes with no loss of PPP session the IP Profile will not get reset, whether up or down.

Looking a bit deeper, I see your profile says you are on Newnet. It was when I was with them I first discovered the re-sync not being seen when it had clearly happened. But it does occur with others as well. This was years before they were taken over, but allied to their inability to get a target noise margin reset done with loads of diagnostic evidence for its validity was one of the contributory factors to my leaving rather than upgrading when my allowance became insufficient. (In the end the noise margin was sorted out through MrSaffron's back door and things were fine thereafter).

the time thing is an artifact of losing the internet connection I think - the netgear updates its time from internet ntp servers.

I think it's got to the stage where you really need to check out the condition of your premises wiring if the BB performance is still giving reason for concern. If the wiring is found to be in good order and you are not likely to get a downstream SNRM of 6dB and sync of ~20000Kbps then there is probably an issue with the line from your premises to the exchange which is Openreach�s responsibility to repair.

Why would something change that much time just dropping from an NTP server? I would expect something unreliable to lose/gain a few seconds over many weeks, but not just to give up all hope and reset to some arbitrary date years ago?

Or is this a netgear bug?

Nick

If a router misses a scheduled "SNTP Synchronised again to server" event due to the PPP link being down then it may revert to a default time until SNTP is synchronised again.

BTW. Installing generic firmware can result in a time server having to be specified manually after any router �factory� resets otherwise the log will just show the duration of the current powered-on time.

Edited by 4M2 (Mon 08-Apr-13 14:26:26)

Since when? I never heard so much nonsense.

All the NTP code does is poll a (NTP) time server and use gradual calculations to keep the time correct. If there is no NTP server available, the on board clock will still keep time to whatever accuracy the code/electronics are capable of - why does it just reset to some unknown date/time - that makes no sense at all.

I think perhaps people are too used to running MS stuff.

Nick

Yes the router will keep a record of the duration of time in days, hours, minutes and seconds since the last boot but the actual time of day may revert to the default factory setting if the the PPP link is down due to it's inability to synchronise to a time server and thus is unable to display the actual time of day, i.e. GMT or BST.

After a reboot the router log will show the DSL is up and PPP link is up and then after some time displaying events in only minutes and seconds the router will synchronise to a SNPT server and then display the actual time of day. A loss of the PPP link can result in a similar thing where only minutes and seconds of duration are displayed until the actual time of day is synchronised.

Edited by 4M2 (Mon 08-Apr-13 14:51:26)

Well, I am sorry, but never in all my years of computing have I seen a (adsl)modem/router or such device change date to 'many years ago' because it wasn't on line.

Nick

Since we don't have access to the OP's router then the reason for "...reported the time in its log as Jan 1 2003" is entirely speculative but it does appear that the router reverted to some kind of factory default date/time when the PPP link was down

On routers I see a lot of midnight times with a date of 01/01/longago ie they don't keep time through a reboot and rely on NTP for current time. Whether that is tied up with a reset of PPP I don't know - will have a play.

OK, but forget the reboot - why (supposedly) does a adsl modem/router reset the date/time if it disconnects?

That to me is pants, and I have never seen it happen?

Could the OP be seeing a crash on the router causing it to reboot in an uncertain state?

Nick

completely conceivable that the software is written to reset the clock when it negotiates a PPP session. No idea if this is the case, but possible. A bug perhaps - resets the system clock instead of a timer. Conceptually possible.

Perhaps the thing does indeed crash all or part of the code and that's what drops the PPP session although the modem part continues to be in sync. In the log file posted above the date/time didn't reset with the loss of IP .

Most routers get power cycled or reset when they're misbehaving so the log files with the evidence are lost and the system time with it.

Time Configuration
Time Source: Automatic
Date: 08-04-2013
Time: 17:36:03
Timezone: (UTC)GMT:Dublin,Edinburgh,London
Summer Time: No
Time Since Power-on: 3 days, 8:21:35
Time Server 1: 212.159.6.10

I leave my router like that - it's happier that way, however I do have generic firmware and software variant: AA

Some may also deliberatly keep PPP up to preserve their ip, as some isp's treat ip's like the dialup days setting a new ip on every new ppp session.

I actually think its wrong to use PPP to monitor stability and using IP profiles likewise to monitor speeds. Of course this problem will originate from the fact openreach, BT wholesale and the retail isp are all seperate entities so generally would require sharing of sync information to the isp's.

This all theoretically can be shared as we already know openreach provide sync speed to BT wholesale (so IP profile can be set) and that the dslam somehow shares the sync information with an external system otherwise DLM wouldnt work. I just dont think its in BT's interest to share it, and the isp's probably wouldnt want it either as would confuse tech support staff.

I am quite suprised that no one has so far suggested asking the ISP to manually set the target margins, and disable DLM as ISP's can now and also make sure the IP profile is set up and correct for the sync afterwards do this with BT wholesale adsl2 + products, Saves the all the faffing around waiting for this and that to maybe happen ect ect, and runing a risk of ending up with a call out bill from BT

As for the PPP session timing out, this can be down to the radius servers working incorrectly,

Edited by tommy45 (Tue 09-Apr-13 04:45:33)

ISPs receive sync speed information via delta reports (ADSL) or in the PPP header (FTTC) don't they ?

You seem to be saying ISPs/CPs should not have any information about what speed a customer's line is running at and whether or not it is connected. Including LLU where the DSLAM/MSAN belongs to them.

Uh?

(Edit - re PPP session, as this is established at the CP kit hiding it from them could be difficult, and their not monitoring it would seems to be a very odd approach).

Edited by RobertoS (Tue 09-Apr-13 09:05:05)

You also seem to be saying it should be compulsory for all IP addresses to be static.

Uh?

no, they can do what they want.

But I consider a more modern approach to dynamic ips is where they have a lease eg. as on virgin media and on sky. In that approach they generally dont change for short outages such as a resync event.

Given the ipv4 shortage now, it can never be compulsory.

In reply to a post by tommy45:

I am quite suprised that no one has so far suggested asking the ISP to manually set the target margins, and disable DLM as ISP's can now and also make sure the IP profile is set up and correct for the sync afterwards do this with BT wholesale adsl2 + products, Saves the all the faffing around waiting for this and that to maybe happen ect ect, and runing a risk of ending up with a call out bill from BT

Problem is we don't really know much about the OP's home wiring, she states early on in this thread that she doesn't have a NTE5 and presumably has the micro filter/router connected to a phone socket. So requesting the ISP to manually set the target margins etc. may indeed be one approach but personally I would rather check initially that the feed from an external junction box (?) to the phone socket (?) is in sound condition (e.g. no corrosion, a twisted pair, etc.) and that there are not any other unfiltered extensions or star wiring issues before proceeding any further...

Thanks to everyone for their input/help.

My OH has checked the wiring and this has now been ' sorted' , if you get my drift.
There was an improvement in my connection as you can see from my stats below:


System Up Time 3 days 04:31:01
Port Status TxPkts RxPkts Collisions Tx B/s Rx B/s Up Time
WAN PPPoA 43184 54840 0 83 683 22:17:51
LAN 100M/Full 790023 795198 0 7394 1278 22:17:52
WLAN 11M/54M 339488 117237 0 780 44 3 days 04:31:01

ADSL Link Downstream Upstream
Connection Speed 13976 kbps 1243 kbps
Line Attenuation 20.0 db 10.0 db
Noise Margin 9.00 db 5.5 db

Maybe the previous socket had become faulty ??? Still no where near the sync of 18000 Kbps I had before this all started back in February.

Unfortunately, I had another loss of 'IP address' last night after being connected for over 9 hrs or so - red light on router for approx. 45 mins until just pressed 'reconnect' in the Router status section. No drop in synch, no drop to zero noise margins, noise etc have been recorded by RouterStats- everything remained unchanged. From this I presume I'm right to say that I was still connected to the exchange?? Bearing in mind my ISP comments re the previous losses of the IP address, I'm not too sure if they will take any notice of this latest episode, let alone take any notice of the RouterStats info. Surely two routers cannot both be faulty??

I have got a new Netgear router which I will be connecting at the weekend just in case!! It will be interesting to see if the same losses if IP address continue??

We will see

Edited by deleted (Thu 11-Apr-13 19:46:25)

Is RouterStats showing the downstream noise margin to fluctuate by a considerable amount or is it fairly stable at about 9dB? In other words, for example, after dark does it reduce to around 6dB then return to around 9dB during daylight hours, or fluctuate erratically by 3dB or more at anytime, or remain within 1dB of 9dB throughout a full 24 hour period?

In amswer to your query, I can confirm that the downstream noise margin does not fluctuate wildy. It remains fairly stable around 9 dB (9.1 to 8.8 dB) with no drastic deviations at all. Unfortunately I don't run RouterStats 24 hrs but it has been operational up to 21.00 - 21.30 hrs so has operated in the 'dark' nighttime hrs.

Thanks

I suspect that the 9dB SNRM and improved speed was due to a resync after "My OH has checked the wiring and this has now been ' sorted'" and the SNRM may now be gradually lowering and which also does appear to be very stable

See how you get on with the new Netgear router at the weekend and when that is booted and sync'd to the exchange hopefully you may see the downstream SNRM at 6dB and your sync speed back to it's previous ~18000Kbps level. The loss of PPP link events will hopefully also be considerably reduced.

Good luck

New Netgear router has been up and running since last Sunday. My connection was slightly faster than previous at over 14.6 Mbs. Unfortunately it dropped the connection on Mon (noise upstream) but has been stable ever since. I've decided to sit it out for a couple of weeks and see what happens. Have noted what has been said previously and will wait the 14+ days to see if my SNRM gets lowered. So far it has not, can't help wondering if my connection speed has been 'banded'?? Previously, the DLM automatically kicked in adjusting the SNR/speed up/down as necessary. Will just have to be patient I suppose!! I have told my ISP that I'll be back if nothing changes as I do feel my connection is under-performing.

Listed below are my most recent stats:

System Up Time 190:42:43
Port Status TxPkts RxPkts Collisions Tx B/s Rx B/s Up Time
WAN PPPoA 832668 1186211 0 223 1958 168:17:38
LAN1 100M/Full 2220317 1769152 0 2252 400 190:40:36
LAN2 Link down
LAN3 100M/Full
LAN4 Link down
WLAN 145M 963015 391981 0 1238 64 190:02:35

ADSL Link Downstream Upstream
Link Rate 14383 Kbps 1043 Kbps
Line Attenuation 17.5 dB 10.8 dB
Noise Margin 10.3 dB 9.8 dB

Let's wait and see

In some respects your stats are worst: previous downstream attenuation reported by the router was 20dB, now you are only seeing an increase of ~400Kbps with an attenuation of 17.5dB. The difference in attenuation figures may be due to the way differing routers "see" the loop loss though.

Upstream noise margin has increased so a drop in upstream sync speed is the result although the attenuation is very similar to that previously reported by the other router.

I would say that you still have a problem with line, not serious but sufficient to reduce your sync speeds, particularly the downstream significantly. A 17.5db downstream attenuation should ideally with a 6dB noise margin give you a sync speed of 20000Kbps or more on ADSL2+.

Perhaps the next step might be to have a NTE5 master socket fitted together with the phone wire from a junction box (?) to the master NTE5 being inspected and replaced if necessary. Then you will be able to check your connection speeds etc. from a test socket and if there is no improvement then an OR engineer could check your line from the test socket to the exchange, however if your sync speed stays at around 14 to 15Mbps and the DSL doesn't drop frequently then an ISP may be very reluctant to book an engineer.