Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[roseway]

Sorry to be argumentative BK, but it's quite normal for the upstream noise margin to be a lot higher than the downstream, because it's capped at quite a low speed. This isn't indicative of higher than normal noise in the upstream part - if the upstream noise margin were much lower, that would be indicative of higher noise.

The dropping PPPoA session could possibly be the result of temporary bursts of interference causing high error rates, but that's not so very likely. It might be a defect of the router, and it would be worth swapping to a different model to make sure about this. Or it could be caused by the ISP's equipment or the interface between the exchange and the ISP.

[Archie]

@B*Cat & roseway,

I have been in touch with my ISP and they think that by increasing the D/S SNRM from 6 to 12 dB and with interleaving on, it may stabilise the PPPoA session.  What do you guys think?

Having said that, I wish to convey roseway's thinking of ISP's equipement or interface between the exchange and the ISP to cause the PPPoA session to drop; if it is alright with you, roseway.

I suppose, I can temporarily borrow a router and see if that makes any difference.  But the question is which one do I do first.  Any thoughts?

Many regards,
Archie

[roseway]

@Archie:

Increasing the target noise margin will reduce error rates, so it would deal with one of the possible causes of the PPPoA dropouts. But it will also lead to a significant reduction in your connection speed (down from ~3000 to ~2000 kbps perhaps).

I'm quite happy for you to repeat anything I've said here to your ISP.

If you can borrow a different router to try out, that would be my first choice to try. Some routers (e.g. Netgear DGN3500) seem to have a real problem with PPPoA dropouts.

[burakkucat]

Sorry to be argumentative BK

Why, Eric, you are not. I promise not to aim a paw swipe in your direction.

it's quite normal for the upstream noise margin to be a lot higher than the downstream, because it's capped at quite a low speed.

I can only say that I have never observed that. I have seen the U/S SNRM a little higher than that of D/S but never "a lot".

This isn't indicative of higher than normal noise in the upstream part - if the upstream noise margin were much lower, that would be indicative of higher noise.

I agree with you there. Obviously I did not correctly type what I had intended to express, earlier this morning.

The dropping PPPoA session could possibly be the result of temporary bursts of interference causing high error rates, but that's not so very likely. It might be a defect of the router, and it would be worth swapping to a different model to make sure about this. Or it could be caused by the ISP's equipment or the interface between the exchange and the ISP.

I see and now understand.

@Archie -- If you are going to test with an alternative modem/router, I would suggest that you purchase a 2Wire 2700HGV (a.k.a. a BT Business Hub) via eBay. (The rest of the forum regulars issue a collective grin, knowing full-well that the b*cat was bound to type that suggestion.)

[Archie]

@Archie:
If you can borrow a different router to try out, that would be my first choice to try. Some routers (e.g. Netgear DGN3500) seem to have a real problem with PPPoA dropouts.

Currently I have Netgear DG834GT Router and I was thinking of Actiontec ADSL2+ wireless router, the one I had used earlier to test the reliability of PPPoA session.

Having said that, my ISP has already changed the D/S SNRM from 6 to 12db and has enabled the interleaving.  Yes, you are right.  My D/S IP Profile on BT Speed Test has dropped from 2.5Mbps to 2.0Mbps, but achieved the speed of 1.9Mbps, which I can live with. The U/S speed has reduced but very little.

Shall I leave it until my ISP can establish if the stability of PPPoA has been achieved using their solution.

Regards,
Archie.

[Archie]

@Archie -- If you are going to test with an alternative modem/router, I would suggest that you purchase a 2Wire 2700HGV (a.k.a. a BT Business Hub) via eBay. (The rest of the forum regulars issue a collective grin, knowing full-well that the b*cat was bound to type that suggestion.)

Thanks for your suggestion of using 2wire 2700HGV.  Much appreciated.

For reasons given to roseway, above, I think we hold fire for time being until my ISP give up with raised hands to surrender!!!. 

Regards,
Archie.

[roseway]

it's quite normal for the upstream noise margin to be a lot higher than the downstream, because it's capped at quite a low speed.

I can only say that I have never observed that. I have seen the U/S SNRM a little higher than that of D/S but never "a lot".

Archie is (like me) on a BT IPStream connection which has the upstream speed capped at 448 kbps. If it were not capped it would connect at a much higher speed with a noise margin of 6 dB, in the same way as your own connection does. But because it's capped, there's a lot of spare capacity which manifests itself as a large noise margin.

Just for comparison, my own noise margins are currently 5.4 dB down, 23.0 dB up.

[burakkucat]

my ISP has already changed the D/S SNRM from 6 to 12db and has enabled the interleaving.  Yes, you are right.  My D/S IP Profile on BT Speed Test has dropped from 2.5Mbps to 2.0Mbps, but achieved the speed of 1.9Mbps, which I can live with. The U/S speed has reduced but very little.

Shall I leave it until my ISP can establish if the stability of PPPoA has been achieved using their solution.

If you can still make use of the connection in that configuration then, yes, allowing some time for statistics to be gathered will be a good idea.

[burakkucat]

Archie is (like me) on a BT IPStream connection which has the upstream speed capped at 448 kbps. If it were not capped it would connect at a much higher speed with a noise margin of 6 dB, in the same way as your own connection does. But because it's capped, there's a lot of spare capacity which manifests itself as a large noise margin.

Just for comparison, my own noise margins are currently 5.4 dB down, 23.0 dB up.

b*cat, having just learnt something, bows towards Eric and says "Ah so!".

[jeffbb]

Hi
@BC quote : U/S attenuation is ~55% of D/S attenuation. That's reasonable. Tick.

Sorry to be pedantic but the difference is NOT 55% . Remember that db values are logarithmic .Each 3db = a doubling of the value .
so the the difference between 61.5 and 34.1  is actually  ~ 1/500 of the DS attenuation .
a line with say 30db DS and 15 db US would be ~1/32 of the DS attenuation
Regards Jeff

[Archie]

If you can still make use of the connection in that configuration then, yes, allowing some time for statistics to be gathered will be a good idea.

At the moment things are acceptable.  If things take a nosedive, I shall sceam. 

Regards,
Archie.

[Archie]

But because it's capped, there's a lot of spare capacity which manifests itself as a large noise margin.

Just for comparison, my own noise margins are currently 5.4 dB down, 23.0 dB up.

I did not know that spare capacity can manifest as a large noise margin.  You learn every day.  Thanks for this info,roseway.

Regards,
Archie.

[burakkucat]

@BC quote : U/S attenuation is ~55% of D/S attenuation. That's reasonable. Tick.

Sorry to be pedantic but the difference is NOT 55% . Remember that db values are logarithmic .Each 3db = a doubling of the value .
so the the difference between 61.5 and 34.1  is actually  ~ 1/500 of the DS attenuation .
a line with say 30db DS and 15 db US would be ~1/32 of the DS attenuation

b*cat appreciates precision, Jeff.

Perhaps I should say, in similar circumstances, that the unit-less numerical value of the U/S attenuation is ~55% of the unit-less numerical value of the D/S attenuation.