Master socket splitter?

Haven't really got a major issue but I've looked (skimmed) through the forums and can't really find an answer to this.
(Dunno which pieces of information are useful so I'll just post what I can think of)

Was on Tiscali ADSL (which was rubbish, obviously) after they bought Pipex and now on O2 ADSL2+ (much better). Samknows says I can get up to 6Mbps download on 2+ or 5Mbps on ADSL, though obviously these are just estimates. With Tiscali I was lucky to get more than 2 and O2 has been steady at around 3 (using speedtest.net). Upload has never got above 1 for either of them using speedtest.net, though a little better with O2. With both Tiscali and O2 I have (had) quite a lot of dropouts, but nothing too bad.

Anyway, bought and fitted a filter plate to split the line for ADSL and voice at the master socket and isolate the bell wire, and this has made download on the modem stats page about 50% faster and substantially increased the SNR for download. What I am wondering is why it has lowered the SNR for upload.

Here's before and after, though I didn't repeat and average--just two snapshots:
Before After
Bandwidth (Up/Down) [kbps/kbps]: 1,178 / 3,099 1,200 / 4,802
Output Power (Up/Down) [dBm]: 12.0 / 15.5 12.0 / 19.0
Line Attenuation (Up/Down) [dB]: 25.0 / 45.5 25.0 / 45.5
SN Margin (Up/Down) [dB]: 7.0 / 6.0 6.0 / 9.5

I didn't do a great job inserting the wires into the new faceplate since I was using the lid of a felt tip pen, but I wouldn't have thought that would be a problem since that is just the voice side of things anyway, and there is a cordless phone base station by the modem, but that hasn't moved at all. No fluorescent lights on for either test, no large metal objects, same weather etc.

Anyway, as I said, it's not massively important as I am quite chuffed with the improvement in download, but I am just a little confused, and would be grateful if you had any ideas.

Thanks,

Martin.

Edit: I had an epiphany and realized that the maximum upload speed possible is 1.3Mbps, but that still doesn't explain why the SNR is lower and also begs the question of why I am getting so close to the maximum on upload and comparatively worse on download.

Edited by martyp (Sat 26-Jun-10 15:50:44)

The SNR margin varies according to the prevailing line conditions and goes up and down during the day, generally lower at night (more radio interference).

Upload uses lower frequencies than the download and is less susceptible to radio interference. I think most people get close to the maximum upload.

http://en.wikipedia.org/wiki/Asymmetric_Digital_Subs...

Thanks, I guess that explains the speed, but it doesn't explain why the SNR has gone down. I took both readings at about the same time (around lunchtime) and since changing it hasn't been past 6.5 at all, and the 6.5 was only on one instance, (otherwise on 6.0, though obviously I haven't refreshed the page every second...). Is the 1dB drop statistically insignificant (it is 1/7 of the ratio, after all..)? Is it that when I got the 7.0 it was possibly only 6.75 and now I am getting 6 is it actually just under 6.25 (making the drop only 1/14 of the ratio)?

Anyway, since it clearly isn't causing any great problem (unless dropouts get worse, rather than better) I guess it doesn't really matter.

Martin.

FYI - It looks like your Target SNR has increased to 9db from your figs!
Does the d/load SNR show around 9db as soon as you restart the router (during daylight hours)?

My attenuation is 51db and I sync around 53xx on a 6db Target . So you should expect better than that.
Did you perhaps restart/connect the router a lot while you were making your changes?

If it has increased then it should come back down if you maintain a stable connection for a few weeks (or so!).

That upload SNR of 6.0 is nothing to worry about, that is the target SNR set by the exchange equipment. The 7.0 you had before is odd, probably just a fluctuation in noise levels.

The download SNR of 9.5 suggests that for some reason you have been put on a higher target SNR, usually done to make an unstable line more stable. Hopefully with time this will be set back to 6 and you will get a little more download speed.

Sounds like a fluke to me. Try posting another 'after' measurement please.

In reply to a post by Backpack:

That upload SNR of 6.0 is nothing to worry about, that is the target SNR set by the exchange equipment. The 7.0 you had before is odd, probably just a fluctuation in noise levels.

That's interesting, I hadn't heard of a Target SNR for 'upload' before only 'download'.
There's plenty of documentation about d/load Target SNR, is this a BTw thing or non-BTw ?

OK, well it's dropped out twice (that I have noticed) since I changed it, which is pretty typical for the timeframe. First time, nothing happened but this time the download has jumped up again, I was untangling cables at the socket so I might have accidentally made it lose connection. Anyway:

Bandwidth (Up/Down) [kbps/kbps]: 1,196 / 5,771
Output Power (Up/Down) [dBm]: 12.0 / 19.0
Line Attenuation (Up/Down) [dB]: 25.0 / 45.0
SN Margin (Up/Down) [dB]: 6.5 / 9.5

Think download was possibly even a little higher than that straight after it came back on, which is making the Samknows prediction look pretty damn accurate.

And badger, I just turned it off, changed it and turned it straight back on (once). Also, it's a BT line but O2 has kit in the exchange.

Martin

Update (01h00):

SNR down is now at 8.0 but everything else is the same.

Edited by martyp (Mon 28-Jun-10 01:00:28)

There does seem to be a target for upstream too. It is probably never discussed as everyone is always obsessed with downstream. Any change in upstream speed though does not cause a resync.

I have seen my upstream speed increase until the margin reaches 6dB +/- 1 and then stay there. If for some reason the margin drops, then the upstream rate will drop too - just enough to get the margin back to 6dB.