It's a long time since I used DGTeam, but it is on one of the pages.

However, you now have something else to consider. You have confirmed earlier instability, and that the noise margin was raised to stabilise the line. Raising the sync-time (connection time) margin reduces the sync speed and improves stability. Lowering the sync-time margin increases speed but makes the line less stable.

That's what it is for.

So if you lower it you are undoing what ADSL24 did to help you.

Not good .

If the instability was with the other router, that gave the dodgy stats, I suggest the first thing is to you see if you stay connected for considerably longer now than you did with that one.

If you do, then it would point to the router being faulty in other ways, and it is safe to try a bit of tweaking as earlier. If that works and also stays stable then ask ADSL24 to reduce it at their end.

If and when you do try a tweak, you can try any number between the 10 I suggested, and 90, and see what is stable. They are percentages but it's a logarithmic scale, so discussion gets complex. Basically 50% of the ISP end setting will reduce it by 3dB and 200% of the ISP setting will add 3dB.

In reply to a post by RobertoS:

If and when you do try a tweak, you can try any number between the 10 I suggested, and 90, and see what is stable. They are percentages but it's a logarithmic scale, so discussion gets complex. Basically 50% of the ISP end setting will reduce it by 3dB and 200% of the ISP setting will add 3dB.

If the sync is capped by rate, rather than noise margin, how do such tweaks work? Can they even overcome the cap?

I'm not saying this is necessarily the case here (although I think it might be). I've just never tried a tweak in such circumstances, and always wondered about this...

If sync is rate caped then no fiddling with target noise margin in the modem will change things

In reply to a post by RobertoS:

They are percentages but it's a logarithmic scale, so discussion gets complex. Basically 50% of the ISP end setting will reduce it by 3dB and 200% of the ISP setting will add 3dB.

It's not too tricky to do a direct calculation ; )

So, if you need precision and want to change the target noise margin by x dB, then just punch the following into Google, and this is the % you need to enter: 10^(x/10+2)

Eg you want to reduce the target noise margin from 14dB to 6dB. This is a change of -8dB. 10^(-8/10+2) = 15.8%

In reply to a post by MrSaffron:

If sync is rate caped then no fiddling with target noise margin in the modem will change things

Thanks. Makes sense, lol.

In reply to a post by vimto_girl:

In reply to a post by RobertoS:

They are percentages but it's a logarithmic scale, so discussion gets complex. Basically 50% of the ISP end setting will reduce it by 3dB and 200% of the ISP setting will add 3dB.

It's not too tricky to do a direct calculation ; )

So, if you need precision and want to change the target noise margin by x dB, then just punch the following into Google, and this is the % you need to enter: 10^(x/10+2)

Eg you want to reduce the target noise margin from 14dB to 6dB. This is a change of -8dB. 10^(-8/10+2) = 15.8%

That might be a little difficult for over 90% of the posters here I'm afraid. Many won't even know what a logarithmic scale is, or have a clue what your notation means.

A simpler method which suffices for the case in point, is to assume the target noise margin is 15dB, as it is usually set in 3dB steps. Then a simple iteration gives 6dB at 12.5%, based on what I posted earlier of halving or doubling equating to a 3dB change. As the tweaker in question works in 10% steps ....

Oh, by the way, punch the following "what" into google?

In reply to a post by RobertoS:

Oh, by the way, punch the following "what" into google?

She means the arithmetic expression after the colon in said sentence. Google will evaluate arithmetic expressions presented to it rather than doing a search. There is no reason it should present any difficulty to any poster here; they are quite capable of "punching" it in blindly w/out knowing anything about logarithmic scales or the notation employed. All they have to do is decide the x dB increase they wish to achieve.

BTW: Her formula is the inverse of one I showed you earlier in PMs.

Edited by XRaySpeX (Fri 13-Jan-12 00:40:59)

I mis-read the relevant paragraph in her post.