Well, connection been up 24 hours now, & not one resync, still getting a rock solid 69 /17.

In reply to a post by Icaras:

A DLM reset wouldn't be applied to all lines on that faulty cable. That's not how it works

If someone gets called out on a fault they can do a reset. But it's just assumed for everyone on that cable that DLM will do its thing, automatically.

It seems to me that if a DLM reset is applied when an engineer makes some changes or repairs on a single line that it would make sense to do it if multiple pairs are repaired/replaced. If not, fair enough, but on ADSL the DLM process seemed very slow to respond and not all lines will be FTTC in a case like this.

Looks like I had G.Inp enable on my line today:

http://www.thinkbroadband.com/ping/share-thumb/5877f...

Router confirms it's G.Inp, latency has doubled to ~20ms. Max attainable rate has jumped from 87Mb to 100Mb, so I'm not sure if it's a good thing or a bad thing..

My line had a minimal errors and a decent SNR - does G.Inp enable by default on all lines, regardless of quality?

It's very odd for latency to increase when G.INP is implemented. In all I've seen it has decreased it.

How often do you disconnect and reconnect normally? It's more likely the implementation was the full depth red at about 2am than when the latency jumped.

Is there anything else in the Sky router stats that might give us a clue what is happening?

Hardly ever disconnect/reconnect - aside from the red at 2am and 10am, the other red spikes are when I commit settings on my Billion 8800 as it mustn't reply to pings for a couple of seconds.

I do monitor downstream SNR via SNMP which you can see below:

http://killsta.co.uk/graph1.jpg

Which does tie in with the G.Inp and the TB graph I posted previously. Crosstalk I wonder? Still doesn't explain the random resync at 10am today though!

Edited by deleted (Tue 23-Jun-15 22:36:01)

It certainly doesn't look like G.INP... though it could be a G.INP on Downstream and then Interleaving has been applied on upstream...

specially as OR are now removing G.INP from existing G.INP enabled (both ways) lines and leaving them on with G.INP on downstream and Fast Path (or what ever is required) on Upstream

.

Edited by deleted (Tue 23-Jun-15 22:55:15)

And here be the relevant line stats:

http://www.killsta.co.uk/stats.jpg

Prior to yesterday, max attainable rate used to be 92Mb ish, with an SNR of 9.3 ish with no interleaving or G.Inp and was perfectly stable with minimal errors and no instability whatsoever. So it looks like G.Inp enabled, max attainable has risen, SNR has decreased and ping has increased. Very odd!

Edited by deleted (Tue 23-Jun-15 22:33:29)

Errmmmm.

That shows significant downstream interleaving - the normal initial depth oh FTTC is 500-800. (The D line).

Interleaving has also introduced its normal first step, Delay = 8ms.

Where does it tell you G.INP is on? The last line, INP (DMP) 3 is totally different.

Edit: See this page, where Impulse Noise Protection appears near the bottom. The stuff above and below it explains interleaving, but ignore the depths/delays etc. figures as they relate to BT Wholesale ADSL2+. Not necessarily FTTC.

Edited by RobertoS (Tue 23-Jun-15 22:54:09)

Yes that page doesn't say G.INP is on at all!

or if it is... your modem/router isn't compatible with it... or needs a firmware update at the very least!