Help with diagnosing ongoing issues with my ADSL2 connection

Forums › Spark New Zealand (including Skinny and BigPipe) › Help with diagnosing ongoing issues with my ADSL2 connection

1 | 2 | 3

8953 posts

Uber Geek

#1443144 8-Dec-2015 21:03

How inclined are you get your hands dirty with checking the wiring yourself? If you can figure out where the leadin comes in from the street (either aerial or underground), you should come across a master splitter fairly easily if one is fitted. You can then temporarily connect the modem and see what happens.

dclegg

2806 posts

Uber Geek

Trusted

#1443302 9-Dec-2015 08:33

RunningMan: How inclined are you get your hands dirty with checking the wiring yourself? If you can figure out where the leadin comes in from the street (either aerial or underground), you should come across a master splitter fairly easily if one is fitted. You can then temporarily connect the modem and see what happens.

That may be a wee bit outside of my area of expertise. And I'd hate to make the situation worse by completely buggering up the Internet or phone line.

Here is what we have outside, which I've always assumed is where the wiring comes into the house. The screws on this looks like they require a non-standard screwdriver, so I'd probably not be able to get the cover of it without sourcing one.

RunningMan

8953 posts

Uber Geek

#1443319 9-Dec-2015 08:55

That is your external termination point (ETP). If you have a master splitter, then it should be inside that box, with cables running from there to your phone jacks, and a separate feed to your modem outlet. There will also be the leadin from the street coming up the conduit that you can see. Everything should be connected inside with gel crimps.

Jase2985

13457 posts

Uber Geek

ID Verified

Lifetime subscriber

#1443325 9-Dec-2015 09:01

if you get the right sized flat head screwdriver you can get those triangle heads off. ultimatly you would want to connect your modem to the 2 incoming pairs there which will completely eliminate your houses wiring.

couple of scotchlocks, some cat5 cable and an old rg11/rg12 cable with the end cut off and you can be away testing.

dclegg

2806 posts

Uber Geek

Trusted

#1449602 13-Dec-2015 12:09

Well this is getting curiouser and curiouser.

@BarTender kindly reached out to me via PM to assist, even though he had to do it outside of work hours due to other commitments. One of the things he observed was that the Spark logs were showing a request that would tend to suggest we were manually disconnecting (PACKET-TYPE="Stop", TERMINATE-CAUSE="User-Request"). He stated that his is normally caused by a disconnection manually invoked by a modems GUI. But these disconnects happen automatically overnight.

Our woes started around about the time I purchased a new ASUS router, so I was instantly suspicious of this. But I took it out of the equation completely by using my old NetGear modem/router. We had the same pattern of an overnight disconnect, with full sync rate not achieved until around 8AM-10AM.

On Friday night, I disconnected all wired equipment and disabled the wifi on the NetGear. When I woke up on Saturday morning, for the first time in over a week we had Internet syncing at its normal full capacity.

Last night I left wifi enabled, and only disconnected the devices running via my TP/Link Powerline adapters (by unplugging the Ethernet cable running to the Powerline adapter attached to the router). This took my PS4, Apple TV3, and MySky+ with new firmware out of the equation. Oh, and the Powerline adapter itself, of course. Once again, I woke up to full sync speeds.

I don't know if this is all co-incidence or not yet. Today I've reinstated the ASUS/Draytek combination, and will run the same test I ran last night using this equipment. If I wake up to full sync speeds again, I'll run it again with everything connected.

This raises an important question though. Is it even possible for equipment that isn't a modem, or router attached to a modem in bridge mode, to initiate a disconnection request either directly or indirectly?

RunningMan

8953 posts

Uber Geek

#1449607 13-Dec-2015 12:49

If that's the case, you may well have 2 separate issues. There is something causing your widely varying downstream sync rate. A variation from 15Mb down to 300kb continues to suggest a physical wiring issue (somewhere!).

On top of that, you may be having further reliability issues due to something closing the connection.

Rather than try and solve both at the same time, it may pay to resolve one then the other.

dclegg

2806 posts

Uber Geek

Trusted

#1449670 13-Dec-2015 14:27

Yeah, the low sync rate when it does reconnect had me wondering if we have two separate issues. Unless the modem initially requests the disconnect because it detects a sync rate below a certain threshold, in which case both may be related.

Equinox IT

Cloud spending continues to surge globally, but most organisations haven’t made the changes necessary to maximise the value and cost-efficiency benefits of their cloud investments. Download the whitepaper From Overspend to Advantage now.

RunningMan

8953 posts

Uber Geek

#1449678 13-Dec-2015 15:02

I'd look at solving the sync rate issue first. If you note down sync rate / attenuation and SNR figures over the course of a couple of days as they vary, that should be enough for your ISP to start investigating with Chorus. Absolute worst case scenario is the issue is within your premises and you pay the Chorus call out fee - at least the issue is resolved, and you get on with life!

Jase2985

13457 posts

Uber Geek

ID Verified

Lifetime subscriber

#1449723 13-Dec-2015 16:56

get something like DSLstats if it works with your device and get it to record the modem stats and graphs for you.

dclegg

2806 posts

Uber Geek

Trusted

#1450025 14-Dec-2015 10:05

Jase2985: get something like DSLstats if it works with your device and get it to record the modem stats and graphs for you.

I couldn't get DSLStats to connect to my modem, but I found a similar app (RouterStats) that does. Thanks for the tip. I'll keep this logging for a while, so I have some meaningful data to pass on to Spark/Chorus.

We had another outage last night, and this time we lost connectivity completely (around 8:55PM; 35 minutes before I was due to start my testing). That seems to be the pattern if I use our DrayTek Vigor120. If I use our NetGer DGN3500, connectivity returns but with a sync rate of around 300Kbps. For both modems, connectivity returns at a sync rate of around 15Mbps mid-morning the next day.

RunningMan

8953 posts

Uber Geek

#1450036 14-Dec-2015 10:33

Given the huge drop in sync rate at about the same time every day, I'd guess some sort of external noise feeding into the line - Some large electrical appliance nearby starting up?

To narrow it down, finding the pattern of when it occur is the key, as it could be time, temperature, moisture, movement related...

dclegg

2806 posts

Uber Geek

Trusted

#1450056 14-Dec-2015 10:43

RunningMan: Given the huge drop in sync rate at about the same time every day, I'd guess some sort of external noise feeding into the line - Some large electrical appliance nearby starting up?

To narrow it down, finding the pattern of when it occur is the key, as it could be time, temperature, moisture, movement related...

It's not always the same time of day. Initially it was occurring some time in the early hours of the morning, but has been known to happen this side of midnight on a few occasions. Last night was the earliest time it has occurred.

Your theory is not without merit though. I live on a peninsula where they are doing some major roadworks at the base. This has me suspicious that this may be a contributing factor.

I have had Chorus out here once already, but as it turns out it was pretty pointless. By the time he came out, my sync rate was running at its full capacity. He reseated an external connection (which he identified as a fault), reset our port at the Exchange and said we were all good to go. At the time I used this to discount the roadworks as a factor, as that would be affecting thousands of other houses in my area, which Chorus should be aware of. That was quite possibly naive of me.

The one thing that is nagging me is the user initiated disconnect that was observed by @BarTender. But I'll deal with that once I've got a handle on this sync rate issue. It wouldn't surprise me if that is actually related, due to the modem firmware initiating it when hitting a low sync rate threshold.

Jase2985

13457 posts

Uber Geek

ID Verified

Lifetime subscriber

#1450146 14-Dec-2015 14:04

routerstats will be helpful with the logs as you will know exactly when it stops working and it should show any changes in SNR, attenuation or the other stats.

Will be interesting to see what happens

dclegg

2806 posts

Uber Geek

Trusted

#1450483 15-Dec-2015 08:32

So I ran RouterStats overnight, and didn't quite get the results I expected.

tl;dr version: Connection went down completely for 30 minutes. Came back up at vastly reduced sync speed for 1.5 hours. Has been fine since then.

Full version:
While we did have our nightly outage/slowdown, this only lasted just over 2 hours. This is the shortest outage we've had by a significant margin (they're usually around 10 hours).

What the logs showed me was that the modem returns very consistent sync stats when everything is chugging along nicely (16Mbps down/1Mbps up, downstream noise margin 11dB, upstream noise margin 12dB). I get a little movement in the noise margin stats, but the sync speeds are constant.

The DSL connection went down around 9:11PM, with sync stats and noise margin values returning 0. Around 9:37PM, it looks like the router attempted to connect, with sync stats of 4.5Mbps/7kbps. The downstream noise margin was logged as 0. The next few readings had 0 figures for all values (readings are polled at approximately 20 second intervals). There were a couple of readings at 9:41PM which had sync rates of 11Mbps/8Kbps, but with rubbish noise margin values.

Connectivity returned at 9:42PM, with sync rats of 379Kbps/920Kbps, and noise margin rations of 12dB/12.5db. The sync statuses stayed constant to around 9:58PM, when I got a few empty stat values in the file (I'm wondering if the router rebooted; but the admin page isn't allowing me to access the logs).

At 10PM, stability resumed, with sync rates of 379Kbps/888Kbps, and it stayed this way until 11:28PM. I got a couple of log entries with 0 figures (as opposed to the null values reported above), and then the modem started syncing at 16Mbps/1Mbps. The sync speeds have largely remained at this speed since then, but there are a couple of entries in the log with null stats, and about 6 wth really low sync stats, and a downstream noise ratio value of 0 (may be rubbish log entries).

I'll probably reach out to Spark today to see if I can get this more formally investigated. But I'll keep the logging in play so I can continue to build up an idea of what exactly is going on here WRT sync speeds.

dclegg

2806 posts

Uber Geek

Trusted

#1451276 16-Dec-2015 08:20

Spark got in contact with me yesterday to inform they had performed a Port reset from their end. I rebooted my modem as requested, but didn't hold out for a miracle, as Chorus had already taken a similar course of action in the past.

RouterStats showed that overnight our connection went down around 12:50AM, and stayed down until 12:57. Prior to going down, it was synching at 16Mbps/1Mbps. When it came back up, it was syncing at 379Kbps/888Kbps. It is still syncing at this speed.

Observations: Last nights performance degradation happened later, and is going on for longer (matching the pattern I've normally observed). The lack of complete connectivity lasted for a shorter duration (7 minutes vs 30).

I've hit the ball back in Sparks court, and will see what they can do from here.

1 | 2 | 3