The times that the blocks of CRC errors come on (just after 11, and just after 1) look strange, don't they - and the SNRM graph is not only good, but at its best at those times!

I note a funny drop in the upstream margin, alongside an increase in FEC at the same time.

Yes, well. When it hit 1500+ errors/min I decided to prove that it's either internal or external... hence the lack of data coming out of here. One of the problems is that I have three APC UPS plus about 11 switch mode PSUs in here so establishing a problem with one is a nightmare. Anyway I got down to running just a laptop on battery and the modem running on several different bricks - don't need the router on as can connect the modem via LAN2 directly to the laptop. No difference in CRC levels.

It's all back up currently but running with around a 1000 errors/min. Very, very odd that the SNR is not affected. On ADSL2 this sort of interference would have dropped the conn and come back with about a 12dB margin from 3dB via the DLM. I had 6 months of this year before last but the people who had the faulty monitor PSU are away. Had everything else in the house off as well with no effect.

It almost seems as though these errors are not real! The modem has been powered down and up again four times in the past two hours with no apparent effect. I expect we'll find out tomorow how real it all is...

The Electricity board are working just up from my cab installing mains into a new house and are digging a lot of road up. Wonder whether there's a generator or somesuch or a faulty modem or the firmware is buggy....

Added in Bitswaps/Min graph.

There is a current QLN graph here.

Edit @ 20:00
Actual speeds are around 55/12 which is pretty well much as it has been, Why are there so few FEC errors?

Edited by tbailey2 (Wed 26-Feb-14 20:02:17)

Well the DLM may have had an intervention at 05:40 this morning when there was a sync loss - but seems to have done very little if in fact it was that! The downstream SNR has dropped vy slightly and the upstream has risen a bit. Attainable and Max are now very close to each other in both directions and actual speeds are pretty well unchanged.

However, I may have found the cause of this interference but I am not going to count chickens etc. At 07:53 I made a change and it's looking better at the moment....

Can confirm there is no DLM intervention at the moment. But the noise margin has dropped a lot in the last hour.

Long time local lurker in this thread. Big thanks to all contributors here. It's been great for keeping up with developments, and i've learned a few things along the way.

I am in Albourne and have had to tolerate 7mb for the last few years, i had fibre installed today via Zen and am pleased to say I'm achieving consistently high speed, at around 72/15. It makes a world of difference

In reply to a post by WWWombat:

Can confirm there is no DLM intervention at the moment. But the noise margin has dropped a lot in the last hour.

I'm assuming the DLM will add in some interleaving as a first attempt at killing some of these CRCs? The ones there currently are quite bursty and are external to the house. The earlier sustained high levels yesterday and today I found were due to the Devolo Homeplugs I have taking the LAN down to the downstairs TV. Just pulling the LAN cable out at the TV end stopped it. It was being reradiated around the room I'm in from a 30 feet or so extension mains cable that runs round it. Origin is the TV itself.

Hi and nice to hear from you. It's good to know someone has benefited from the chit-chat here over the past year or so and we got there in the end. Trust you'll make full use of your extra speed - 7Mbps isn't that bad, I have a friiend who is lucky to get 1 on a good day... After all, remember dial-up?

In reply to a post by WWWombat:

Can confirm there is no DLM intervention at the moment. But the noise margin has dropped a lot in the last hour.

Having looked again at the pattern for the whole evening/morning I'll add that the pattern of the change in noise margin reflects behaviour I used to see on my Mum's ADSL2+ connection in Hull. It would start around 4-5pm, and continue until 9pm.

For that, we could only conclude it was some neighbour(s) turning modems on after school or work. I'm not sure that is quite so likely for FTTC.

Another possibility could be central heating running - especially when we see the burst in the morning too.

In reply to a post by tbailey2:

I'm assuming the DLM will add in some interleaving as a first attempt at killing some of these CRCs?

Yes. Strictly, it turns on FEC (error correction) to actually get rid of the CRC errors. It turns on interleaving to make FEC more effective at dealing with certain (typical) sizes of noise bursts.

Here's what my line was like without FEC+Interleaving: TBB-BQM Image, on day of activation

And 2 days later, when DLM intervened: TBB-BQM after 48 hours

Origin is the TV itself.

Plasma?

Thanks. The TBB graphs are pretty dramatic changes.

BTW I've been connected a week now almost to the minute so the DLM can't be too unhappy...

I'm reasonably sure I know what it could be - a monitor switched mode PSU a few doors away that went bananas year before last and was 'repaired'. That used to instantly drop the ADSL line due to about a 6db interference level. I reduced the effect a lot after putting on a load of type 43 ferrite cores on the phone lead up the wall outside and in here on the cabling. Took them off after repair. Now have put them all back on about 10:30 this morning but as you say may not have much effect with FTTC. There is no one in there at the moment.

Part of that problem is that we are on overhead three phase power cabling and it rerediates rubbish.

No not plasma. Its a Sony Bravia LED HD. If you look at the CRC graph I plugged it all in again at 07:50 or so just to make sure it was the TV and you can see the effect! It peaks and clips at 1000 (actual is probably 2500 per min) for a few minutes...

Edited by tbailey2 (Fri 28-Feb-14 17:48:11)

It appears the RFI is not sourced where I thought it was as no one arrived home until about 8 minutes after it started at exactly the same time as yesterday so does sound like it might be a timed piece of equipment. See if it ceases at the same time....

23:00
Which it has...

Edited by tbailey2 (Fri 28-Feb-14 23:11:25)