SNR keep dropping, streaming buffers, line out of sync

[elriconess]

Good evening

What may the problem be if the line does the above mentioned? It;s after the rain and I've asked Telkom to come look at it a few weeks ago and they couldn't find a fault on the line.

Stats below:
Operational Mode G.Dmt.BisPlus
Annex Type AnnexA
Upstream 511800
Downstream 4095900
SNR Margin(Upstream) 18.2 dB
SNR Margin(Downstream) 20.1 dB
Line Attenuation(Upstream) 2.0 dB
Line Attenuation(Downstream) 9.5 dB

Operational Mode G.Dmt.BisPlus
Annex Type AnnexA
Upstream 511800
Downstream 4095900
SNR Margin(Upstream) 30.5 dB
SNR Margin(Downstream) 32.30 dB
Line Attenuation(Upstream) 2.0 dB
Line Attenuation(Downstream) 9.5 dB

It changes continuously

 

[Pada]

At least your attenuation values remained the same, meaning the physical connection's state remained the same (like no loose connections or corroded terminals, etc).
Seeing that the SNR margins changed quite drastically, it would mean that some kind of other interference is present: lights, other electrical appliances, etc.

If you can see a pattern in the SNR margins - for example that it is low between say 7pm and 7am, then it is highly likely due to street lights. There are tonnes of other possibilities (eg. rain) though.

Seeing that you're so close to the DSLAM/exchange, your stats are excellent for a 4Mbps connection! Do you suffer from disconnects or have you just noticed a fluctuation in the SNR margins?
The reason I ask is because SNR margins above 12dB should provide you with a stable enough connection to not even have to worry about the SNR margins fluctuating The stats that you posted indicated no change in the sync rate or any sign of an "unstable" connection.

Lastly, you could perhaps force your modem on G.DMT mode, instead of ADSL2+. G.DMT mode would disconnect completely when it resyncs at a slower/faster speed, where as ADSL2+ should be seamless if I'm not mistaken?!

 

[elriconess]

I do not get disconnections. I forced it to G.DMT mode and I get these CRC errors they are in thousands more when it starts raining here like now. Goes up to 40000+:

Operational Mode G.Dmt
Annex Type AnnexA
Upstream 512000
Downstream 4096000
SNR Margin(Upstream) 29 dB
SNR Margin(Downstream) 28.0 dB
Line Attenuation(Upstream) 3.0 dB
Line Attenuation(Downstream) 4.5 dB
CRC Errors(Upstream) 1031
CRC Errors(Downstream) 0

 

[elriconess]

It's happening again! Zeus has come again! It's raining with thunder like it's no one's business and I have disconnected this time :|

 

[davidlight]

Could be water in the cable joints.

I'm living in a complex (no, the security type, not the mental type) and had continuous sync and streaming issues during the wet season. Telkom techie after another until eventually a month ago they sent one that took a closer look at the infrastructure from the MSAN to my house. Turns out, inside a manhole in our complex, there is a 'capsule' where the cables are spliced to the various houses. This entire manhole had filled with water (presumably rainwater and seepage) and eventually so did this watertight capsule (which over the years lost its watertightness). We emptied the manhole, left the cables out to dry and nearly instantly my line stats improved dramatically! After the heavy rain this weekend, same story, I opened the manhole, emptied it and by last night my Apple TV streaming had returned to normal (no delays on 4Mb).

Could be your problem too?

 

[ponder]

I do not get disconnections. I forced it to G.DMT mode and I get these CRC errors they are in thousands more when it starts raining here like now. Goes up to 40000+:

Somewhere along your line you have exposed copper if it changes during rain. Sounds like a physical cable fault.

 

[Superspeed]

Log a fault with Telkom. This is a physical issue for sure.

 

[elriconess]

Today it's slow speeds and high latency. What the hell is going on in this area.


Tracing route to www.l.google.com [74.125.233.83]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms Billion.400G [10.0.0.2]
2 11 ms 10 ms 11 ms 41.177.38.1
3 521 ms 513 ms * ct3-ipc.dsl.rsaweb.co.za [196.6.121.50]
4 524 ms 522 ms 519 ms ct2-ct3.dsl.rsaweb.co.za [196.6.121.114]
5 514 ms 523 ms 520 ms ct6-vlan196.ctcore.net [196.41.96.236]
6 660 ms 664 ms 662 ms th006-lon-po-wacs.borwood.net [217.78.80.241]
7 658 ms 680 ms 660 ms th102-lon-vl-303.borwood.net [217.78.88.37]
8 664 ms 660 ms 661 ms gi-0-3-0-1-63.r02.londen03.uk.bb.gin.ntt.net [21
3.130.48.1]
9 651 ms * 659 ms xe-0.level3.londen03.uk.bb.gin.ntt.net [129.250.
8.138]
10 666 ms 659 ms 657 ms ae-52-52.csw2.London1.Level3.net [4.69.139.120]

11 673 ms 652 ms 668 ms ae-228-3604.edge3.London1.Level3.net [4.69.166.1
58]
12 666 ms 658 ms 673 ms 212.113.15.186
13 659 ms 657 ms 667 ms 209.85.255.78
14 676 ms 692 ms * 209.85.253.94
15 667 ms 666 ms 670 ms 72.14.232.210
16 678 ms 681 ms 677 ms 216.239.43.43
17 848 ms 836 ms * 72.14.232.103
18 831 ms 841 ms 850 ms 64.233.174.57
19 843 ms 922 ms 834 ms jnb01s02-in-f19.1e100.net [74.125.233.83]

 

[Pada]

I'd suggest that you try a different ISP account, because the second hop's latency wasn't high.

 

[elriconess]

I'd suggest that you try a different ISP account, because the second hop's latency wasn't high.

I have tried different ISPs with the same results. Can it be that the connection from the UMC to the Exchange seem fine but from the Exchange to the Telkom network maybe broken?

 

[davidlight]

I had the same issue, no matter how many accounts I tried, on some pings it's perfect, and then seconds later dismal. From my experience it's a cable fault on the last mile. Had exactly the same issue you are describing here at the office just recently. Eventually it was found that the cable from our box in the street (industrial area) to the exchange in Meyerspark some 3km away had an ongoing fault which they are now finally attending to. How do you attach a photo to these posts?