Changed phone socket performance problem

I recently took off my old phone socket (ancient old style thing I'm not sure the name of) and put on a more modern NTE5C socket. I didn't play with the master socket, just the one socket that comes into the flat. The usual two wires were in the old socket and I only connected those to the new socket too.

Stats on the router (supplied Fritzbox 3490) seemed to have improved nicely. Attenuation went down to 6db (I think it was in the 20s before hand, I can't remember) and signal to noise ratio has gone from 6db/3db up to 20/17db.

Sync speeds have always been excellent (being so close to the exchange) but I did all this because I was noticing packet loss when gaming, even though any tests in windows seemed clear,I wanted to rule out everything.


The weird thing is, after doing all this, my first hop to Zen has gone from 6ms up to12ms.I know this is a ridiculously low response time to start with but it seems too much of a coincidence this happened at the same time I changed the socket.

The SNR 'spectrum' graph on the router, in the DSL information now has 2 big chunks of 'data' missing from it, where as before it was solid all the way down the spectrum.

As another note, a couple of weeks before I did all this, when I first upgraded to fibre, I called support about my strangely low SNR (6/3db) and the guy said it was fine for it to be so low, as it was profiling itself to be as low as possible to get the best speeds out of the connection, which kind of makes sense, but doesn't explain why I'm back up to 20/17db).

I've given it a few weeks, to see if my connection would settle but I really want to get that ultra low response time I was getting before I did all this. 'Fast track' is opt out by the way. If that's still a thing.

Line stats are below. Does any of this look off? TIA


Max. DSLAM throughput kbit/s 40000 10000
Min. DSLAM throughput kbit/s 128 128
Attainable throughput kbit/s 85682 28622
Current throughput kbit/s 40000 10000
Seamless rate adaptation off off

Latency 1 ms fast
Impulse Noise Protection (INP) 0 0
G.INP off off

Signal-to-noise ratio dB 20 17
Bitswap on on
Line attenuation dB 6 6

Profile 17a
G.Vector off off

Without full before and after data hard to say a lot i.e. what was attainable and current throughput previously.

Changing a socket alone (and not removing anything like star wiring/extensions) and seeing a 14dB drop in attenuation is unusual and may be memory tricks, but as I said lack of data makes it hard to judge.

You are maxing out the 40/10 product and thus have lots of spare noise margin NOW

The DLM system will train down to 3dB on stable lines to squeeze more speed out, but on current stats no need for that, sort of suggesting your sync speed previously was well below the 40000 figure you have now.

There is a high chance my memory of the old attenuation is inaccurate. It doesn't seem right it went down in the first place, just by change the wall socket.

Sync speed was always maxed. Attainable values on the router don't seem to have changed either, but I'm not 100% suree.

I would have thought the DLM would have trained itself by now, if it wanted to. Isn't it 72 hours? Sync isn't really my problem, I just want to know line is stable. Am I correct in thinking SNR will have nothing to do with my ping?

Also, if the DLM is happy with the current sync/stability will it keep the current higher SNR profile or will it still try to achieve the lower profile?

My main question is probably, is there any way any of this has affected my ping, because it just seems too much of a coincidence? Thanks

I'm still paranoid games seem jittery, even if the line shows no packet loss

DLM runs 24/7 always. It never stops training the line. It works with any changes in the line condition at any time to ensure stability.

When you say you did not play with the master socket, what does that mean? It sounds like you removed it entirely, and replaced it - so elaborate on how that is not playing with it, if indeed you did not touch it, you may now have a second master on the line.

Edited by ukhardy07 (Thu 25-Jan-18 13:32:16)

You might just find that its your first hop that has changed, your connection could terminate in their London or Manchester POP's.

For example this time I ended up in the London one...
losubs.subs.bng2.th-lon.zen.net.uk

Some times disconnection / reconnected can change this, but you have to be careful its not detected as line errors

Yeap I think some of the other changes such as routing that others are talking about may be the issue.

The DSL part on the FTTC connection is just a small fraction of the network that can cause issues, so time to explore trace route and ping tools.

Hi everyone, thanks for all your responses, I am sorry for such a rushed response to all of you. I've been trying to get onto the Manchester gateway for a few days now but I'm constantly getting put on London, so I'll keep trying, I just haven't had much chance.

I can't see it making such a large increase in ping difference but as soon I get onto a different gateway I'll let you know.

If anything else comes to mind please feel free to add suggestions as I can't see gateway doing it, but we'll see. Again, sorry for the late and rush response

Hi,

If you PM me your username I'll be happy to take a look for you.

Thanks,

Jon