In reply to a post by camieabz:

I'd be looking at the shielding of...well everything really. Shielding and proper grounding / termination of anything electrical or comms related might not deal with the interference, but it might minimise it.

There's always this - https://fishofgold.files.wordpress.com/2014/06/13964...

Unfortunately (from a fix point of view) 95% of the electrical wiring is brand new (including meter tails, consumer unit, all light fixtures, all sockets). It was all wired by my father who's a spark with my supervision to ensure it meets my OCD standards which includes much better spacing than required and decent kit.

In addition all the network cables and coxial cables are brand new and kept away from power ). The NTE5 is brand new and wired directly from the underground cable.

In reply to a post by broadbandjockey:

In reply to a post by Zarjaz:

I’d suggest taking a ‘screenshot’ of all the router stats when it isn’t happening to start with ... to have baseline as such. Then have the suspect switched on at an agreed time and then watch the stats.

Is an overall wider picture of the event not possible by examining the DSLAM logs ?

Though possible, I doubt the poster has access to such info.

I was talking about BT using the logs as an aid to solve the original problem, to assess the likely spread of the affected punters,

In that case, Openreach, have a few systems that can be used to show its evidence in a location, but not to pinpoint it. That requires good old fashioned legwork.

In reply to a post by Zarjaz:

In that case, Openreach, have a few systems that can be used to show its evidence in a location, but not to pinpoint it. That requires good old fashioned legwork.

Reminds me of a problem I had with ADSL when I lived in a different village many years ago.....

I noticed the SNR dropping from 8dB to 1dB at regular intervals for up to an hour at a time, then returning to perfect levels. The ISP were helpful and checked everything out - no fault found. This drop was being handled so well by the OR adaptive kit at the exchange that the connection never dropped, but of course the speed was lower than it could be.

I went into detective mode over the course of a week……

What could be the source of the interference? Sure enough with a SW radio I could clearly hear the interference (coinciding with the drop in SNR), but where was it coming from? Street lights? No. Neighbours? No. My equipment? No. Over two weeks my net widened, with considerable 'leg work'....

Eventually I went for a long walk and stumbled across a sewage processing plant over two miles from my property and that of any other village property. I noticed the plant machinery ran to roughly the same timing. I returned the next day, armed with my laptop, VPN connected to the live SNR output of the modem, the SW radio and a six pack of beer, and made myself comfortable on the grass by the fence outside the site. With only one can of beer (maybe two / three...) down - Bingo - as soon as the machinery started up, the SNR went down to almost zero and the radio nearly blew my ears off with the noise! When it stopped the SNR went up by at least 6dB.

I contacted the sewage processing company, who to my surprise, were very helpful and quickly setup a site visit. I showed the Engineer my data, both historic and live, and we went inside the plant control room.

I will never forget the look of horror on his face when he discovered that the racks had never been Earthed since installation - there was the fault! Ironically, he even said that they wondered why in the office, their ADSL connection to the plant dropped out when the machinery ran.

A couple of weeks later, the problem went away, for not just me, but everyone in the village (about 500 properties), most of whom were probably unaware.

So the moral of this story? To find interference often requires time, a SW radio and as Zarjaz says 'leg work' (and, perhaps, a six pack of beer....:-))

Mendip.

Edited by deleted (Tue 22-Sep-20 15:34:08)

An excellent and amusing tale about the sewage plant, now the problem has been solved. Certainly not amusing while it was present.

What is does still illustrate is that even where electrical and electronic equipment has many regulations and standards to prevent such problems, we can never rule out human error or subsequent undetected faults in such equipment or preventative measures.

How was the person to know that they where doing anything wrong? How do you know that there is no faulty equipment in your house that is not impacting broadband performance? I am sure you would be similarly mortified if an apparently perfectly good appliance was causing your neighbourhood such problems.

Personally I would invest in a SDR (software defined radio) USB dongle, plug it into your laptop/computer and start monitoring the RF environment if you think you have a problem. They are not expensive, prices starting at 25USD. I personally have one of these

https://www.rtl-sdr.com/buy-rtl-sdr-dvb-t-dongles/

In direct sampling mode it does 500kHz to 24MHz so ideal for VDSL, but you will miss some of the bottom bands of ADSL. There are other dongles that will cover that for you.

It is in effect a very cheap spectrum analyser. I note it would have been cheaper for Openreach to have posted one of these out along with a raspberry Pi to a customer in the village on a none return basis and just collected the stats and analysed remotely than paying for engineers for overnight stays and being out at the crack of dawn.

It is in effect a very cheap spectrum analyser. I note it would have been cheaper for Openreach to have posted one of these out along with a raspberry Pi to a customer in the village on a none return basis and just collected the stats and analysed remotely than paying for engineers for overnight stays and being out at the crack of dawn.

..... and these remotely gathered stats would have pinpointed the exact faulty equipment/culprit ? (sigh)

They would have pinpointed that the problem was RF noise, potentially months earlier to then justify sending a full team out to pin point the dodgy appliance.

My guess is not investigating RF noise earlier is down to the cost factor in doing so. When they sent the expensive team out to look for RF noise they had no idea if this was a problem, it was by Openreach's own admission a last throw of the dice.