Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[tristan]

Hello all,

I'm appealing for technical help interpreting what might be going on with my FTTC connection (a far better line that the last time I posted a few years ago!). Installed in Jan, I was getting solid 74 down, 20 up until a month ago, with uptimes >40days. I was monitoring this using an unlocked H612.

Then something happened.

I noticed the connection was dropping / re-syncing (not sure of the correct terminology) every few hours / days in an irregular pattern. During a 'dropping' phase, picking up the phone, I can hear pops and crackles along with static.

An engineer has visited during a 'good' period when I am getting my normal speed. We have ~2 inches of internal wiring (through a window frame), no extensions. No faults detected. He replaced the H612 & power supply just in case. 15 mins after he left the line dropped...

A 2nd visit is scheduled for Monday 13th, so I'd be keen to canvas opinion as to what I can ask him to check.

Attached are 3 stats images:

- 'recent', when I get the speed I have come to expect (1526 on 30/6)
- 'bad', the stats following a drop and what it looks like until it resyncs itself (1433 on 30/6)
- 'normal', stats from April before all this was happening.

My stats are being uploaded to mydslwebstats - username "tristanc".

I experienced a drop yesterday, then started using the dslstats app for RaspberryPi to log all this. So not much data there... It re-synced this morning at 4am-ish but I haven't recovered back to the usual speed, although the attainable is where it usually is.

All help greatly appreciated!

Tristan

[tristan]

Just prior to this post timestamp the connection dropped.

The DSLstats app is wonderful! Watching in real time the issue. The SNR Margin up decreases drastically just before things stopped. Not sure if this is the driver?

[burakkucat]

I noticed the connection was dropping / re-syncing (not sure of the correct terminology) every few hours / days in an irregular pattern. During a 'dropping' phase, picking up the phone, I can hear pops and crackles along with static.

You have recounted symptoms that are indicative of a HR fault.

Have you tried performing a quiet line test, both with the modem/router & telephone connected to the line and then just with a wired telephone connected?

[tristan]

Thanks for the reply.

Yes, I dial 17070 option 2 to hear the crackle / static in all its glory. When the problem first occurred I tried what you suggested whilst on the phone (mobile) to BT; removing the DECT handset, remove H612, use plain old wired phone etc. The crackle was still evident, although their remote checks indicated a perfect line...

I'm hesitant of trying this currently in case it interferes with anything OR / BT are monitoring. But guess I can get it reset on Monday anyhow.

However, for the last hour or so the crackle has gone and the SNMR plot seems to have stabilised somewhat, albeit at lower attainable / synced speeds.

Just googling HR faults, I see this is the 'go out and test each junction' approach?

I'll see if the OR guy will check things at the small junction box, about 4m of cable from where the line comes in to the house. From there the cable goes underground, and the nearest small green box is about 20m from there next to a BT manhole. The cabinet is about 100m away from there along a straight road.

Tristan

[NewtronStar]

Hope you get a morning visit and from what I saw in your WDWS your SNRM dB up/down went wild from 10am to 2pm which suggest the HR fault is intermittent.

[tristan]

Yes - I've a morning slot!

I've just researched some ex-BT engineers in the Oxfordshire area who might, at short notice, be able to race round to help with any diagnosis if the line behaves itself again during the OR visit...

Is there any way of telling what is the driver of the change in SNMR? The reasons I've been given from the Twitter help people (very helpful actually) for my issues include '"it's the line settling itself". If that were true, I'd expect some sort of pattern or predictability to what I see. The observations I've made and communicated to them include that this issue seldom happens at night.

[burakkucat]

Is there any way of telling what is the driver of the change in SNMR?

The wild, erratic, swing of the SNRM over a very short period of time is a function of the changing resistance of the joint.

Everything you have recounted are the symptoms of a HR or semi-conducting joint. Such a joint can be found by --

• Visual inspection
• RFL
• TDR

It really depends upon the experience of the attending Openreach technician as to how s/he will go about the task. (I would look at a TDR trace whilst calling the line from a mobile phone. If there is one particular deflection in the trace that changes its appearance with the application of the ringing voltage, then that would be a likely candidate for visual inspection.)

[tristan]

Thanks - I'll suggest this to the OR tech if he does the same routine as last time.

Interesting that today the line re-synced to 'normal' at ~7am, stabilised, then the SNMR decreased at around the same time as it did yesterday (1030ish). We'll see if things improve at ~1530. A link to temperature perhaps?

[Weaver]

Does weather - water or temperature change - have a noticeable effect? Because I feel that’s something we might well expect.

[Some ISPs are better than others (ie more aggressive with BT) if the customer is having no luck with Openreach. Feel free to PM me in that case if getting nowhere.]

[burakkucat]

Does weather - water or temperature change - have a noticeable effect?

Water, certainly, if it gets to where it shouldn't!

Temperature can have a pronounced effect if there is a wiring defect that is open to the direct glare of the summer sun. Bald_Eagle1's saga, documented in a thread at this forum (which runs into a high double-figure page count), was finally resolved when a broken wire was discovered. Before that discovery, clear evidence had been noted that cold temperatures were good whilst hot temperatures were bad. (The two ends of the broken wire were held in close proximity by the unbroken PVC insulation. In cold temperatures the PVC insulation was stiffer than in hot temperatures.)

[tristan]

I hadn't noticed temperature impacting things before - but then again, I only started being annoyed at the dropped connections when I started working from home and it occurred during the day! If I can work out how to extract the raw data from MDWS (is this possible?) I'll overlay temperature and see if there's a correlation.

<aside>When I had a FTTC issue at my mother's house (extremely rural) and sought help I thought I had found a link to temperature / weather: http://forum.kitz.co.uk/index.php/topic,11424.msg222938.html#msg222938

Indeed, the many lengths of cable they replaced in patches had damage from farm machinery exposing the cores to the elements. Unfortunately, this only decreased the line quality; guessing due to the increased number of junctions along the path.

We never resolved this issue and I've set up an alternative connection using a 4G dongle, a BananaPi running OpenWrt and a VPN running on a VPS to get a static IP routed to the BananaPi. Phew. But that means we get ~20 down, 12 up over the VPN, and 40+ down 20 up without. Latency is ~80ms. Fine for her web browsing and video conferencing / facetime.</aside>

Given I have had a 1st OR visit (internal wiring observed & modem replaced), would the 2nd visit be a lift & shift on to a different pair / use the 2nd pair on the line in to the house if there is one? This might resolve the problem at a stroke?

OR had apparently completed some work that should have improved things - I'll be interested to know what this work was. I will ask the guy on Monday to find out.

I've just restarted the H612 and have got my attainable back up to 78 down, 31.5 up. The sync rate hasn't changed (60 down, 19 up). SNMR is back to 11 down, 14 up.

[burakkucat]

. . . If I can work out how to extract the raw data from MDWS (is this possible?) I'll overlay temperature and see if there's a correlation.

As you are uploading the raw data to MDWS, you should already have it stored locally!    So why bother trying to download it? 

As to what the Openreach technician actually does on the day is really up to her/him. (As an aside, in response to something you typed earlier, it would be best not to directly suggest to her/him what to do. Perhaps, in the course of a conversation, just ask what technique s/he would use to locate a possible HR joint. Make it sound as if you are interested and would like to learn from an expert.    )

[tristan]

As you are uploading the raw data to MDWS, you should already have it stored locally!    So why bother trying to download it? 

Ah - my little RaspberryPi struggles to run the app. I'm not sure how he's coded the sampling, but it seems a bit inefficient as the processor is pegged at 100% during it (the 'old' script + one's I developed myself run at <<100% and are fast). I also think there's a memory leak - after 18 hours or so the Pi crashes and needs to rebooted. I don't think any of the data up to that point has been saved... Might have to switch to a VM running on my Mac. Bit of an overkill!

As to what the Openreach technician actually does on the day is really up to her/him. (As an aside, in response to something you typed earlier, it would be best not to directly suggest to her/him what to do. Perhaps, in the course of a conversation, just ask what technique s/he would use to locate a possible HR joint. Make it sound as if you are interested and would like to learn from an expert.    )

Agreed - I'll be nice this time!

It's when they come back telling me the line's perfect and that the max speed I'll get is x, when x << y, where y is the speed I got for months up to a few weeks ago. Anyhow, they must be able to see all the re-connections (although I think the BT monitoring doesn't sample frequently enough) so there can be no doubt there is a fault...

[tristan]

Oh well....

OR turned up and completed the following:

1) replaced the master socket
2) replaced the H612 I was given last week with an ECI
3) reset the line

His diagnosis is that it is a REIN issue.

Alas, according to speedtest.btwholesale.com I'm left with max achievable 63 down and 20 up. And the actuals are 57 and 12. My pings (to BBC) have gone from 14ish to 20.

So I've no easy way of getting my line stats and my connection is worse than it was... At least I know that I'm 300m from the cabinet according to his box.

Any advice?

[GigabitEthernet]

Has he built a REIN case?