ADSL2+ stats

I have recently been moved to an ADSL2+ connection. Before the move I was getting rock solid 8128kb/s downstream and 448kb/s upstream. Here are the current stats from my DG834G router. It's stayed fairly consistent for a week or more.

ADSL Link Downstream Upstream
Connection Speed 13029 kbps 1072 kbps
Line Attenuation 16 db 4 db
Noise Margin 6 db 7 db

This is a worthwhile speed boost, especially to the upstream speed. Do these numbers look sensible? Can I expect any more speed?

I'm sorry to say in my opinion those are terrible speeds.

With that attenuation I would expect over 18Mbps downstream and whatever upstream the WBMC connection allows - I thought it was 1.3-1.4Mbps.

Have you optimised your home wiring? Have a look at the Ring/Bell wire and Miscellaneous nasties links on my Troubleshooting page.

Agreed, very poor with that attenuation. Some noise going on somewhere.

I don't think there's much scope for improving my setup.

There is an ADSL splitter faceplate at the master socket (scrounged from the BT man who took out my Home Highway some years ago). Just a short lead from there to the DG834G. With this setup I presume the wiring on the phone side will not affect speeds.

The DG834G is old but flashed with the latest firmware.

Try a normal filter plugged into the test socket on the wall at the back in this pic.

The faceplate you say you have will not be filtering the ring/bell wire. Only the latest Openreach ones do that - or third-party equivalents.

Did you explicitly request the maximum upstream as normally we'd cap it at 448kbps?

In reply to a post by RobertoS:

Try a normal filter plugged into the test socket on the wall at the back in this pic.

The faceplate you say you have will not be filtering the ring/bell wire. Only the latest Openreach ones do that - or third-party equivalents.

I don't understand - an ADSL faceplate has a separate ADSL output so what does it matter whether the bell wire is filtered on the phone output?

Yes - it does matter.

The ring wire acts as a radio antenna/aerial and feeds noise at ADSLx frequencies back onto the line unless it is filtered or disconnected. Thus screwing up the broadband.

In reply to a post by RobertoS:

Yes - it does matter.

The ring wire acts as a radio antenna/aerial and feeds noise at ADSLx frequencies back onto the line unless it is filtered or disconnected. Thus screwing up the broadband.

OK thanks, I'd better check mine again because I can't remember whether I disconnected the ring wire. It's one of these:

http://www.clarity.it/xcart/product.php?productid=16...

In reply to a post by kasg:

I hadn't spotted on your first post that you are not the OP to whom my suggestion was addressed.

In your case, with the Clarity faceplate, I don't know whether a ring wire filter is included in the circuitry. I know it is with the ADSL Nation version and the latest Openreach NTE5A. The OP has a standard ADSL filtered faceplate which definitely did not filter the ring wire.

If you are in doubt do as I suggested to the OP. Re-sync and take stats, (if the connection has been up for a while the comparison becomes invalid), then connect with a normal filter to the test socket and take the stats again.

Just connected direct to master socket. Modem and microfitler only, no phones or anything else. Here are the stats. Effectively no change.

ADSL Link Downstream Upstream
Connection Speed 12581 kbps 1067 kbps
Line Attenuation 16 db 4 db
Noise Margin 6 db 7 db

Upstream speed is what I was given, didn't specifically ask for anything special, and I'm very happy with it.Makes more practical improvement to me than a very high downstream speed.

EDIT:
And again after returning to my normal setup with all phones connected

ADSL Link Downstream Upstream
Connection Speed 12773 kbps 1095 kbps
Line Attenuation 16 db 4 db
Noise Margin 6 db 7 db

Edited by ppppenguin99 (Wed 20-Jan-10 07:52:25)

You do mean the test socket inside the master socket I hope? The external socket on the master often doesn't show any difference.

If you do mean the test socket then the low speed is baffling.

Yes, I do mean the test socket inside the master socket.

I'm not worried about the speed I'm getting, just idly wondering if ti could be better.

It's definitely very low for the attenuation. For example look at this graph and put your stats into this estimator.

Which version of 834G is it? There is a possibility it doesn't get on well with the BT line cards on ADSL2+.

Says DG834G v2 on the case.

I flashed it with V2.10.09 firmware before the change to ADSL2+.

So that's an AR7 chipset. I think that might be the problem. But getting out of my depth now.

It might be an idea if no-one "chips" in here to start a thread in a forum such as BT Wholesale Implementation or DLS Hardware asking about how it gets on with the BT MSANs. Certainly some "ADSL2+ capable" routers are less than good with them.

In reply to a post by RobertoS:

It's definitely very low for the attenuation. For example look at this graph and put your stats into this estimator.

It's a bit disappointing that, with a 54dB attenuation, I'm not going to get any benefit from ADLS2+

In reply to a post by kasg:

In reply to a post by RobertoS:

It's definitely very low for the attenuation. For example look at this graph and put your stats into this estimator.

It's a bit disappointing that, with a 54dB attenuation, I'm not going to get any benefit from ADLS2+

Have you checked the stats using the test socket as I suggested.

(It would help if you clicked the Reply button on posts where I have replied to you, not just my latest post in the forum - most of which are addressed to the OP. As I generally use Threaded mode it is very confusing and I have to search back to find out what you are talking about. Like to find the stats you mention. Even in Flat Mode, if you read the side headers you will see what I mean. There is a link in them to an earlier post, and if it is always just the preceding post it makes that useless).

Remember that ADSL2+ increases your attenuation by 3dB

Latest firmware for V2 is 3.0.38 (yes that is for V2 despite the 3.x numbering). Might be worth a go.

Mind you, I also have a low sync for my attenuation using a DG834G V2:

ADSL Link Downstream Upstream
Connection Speed 15227 kbps 1085 kbps
Line Attenuation 15 db 3.5 db
Noise Margin 8 db 6 db

Telnet-ing into my router tells me that the ADSL2+ hardware at the exchange is using Texas Instruments chipset (V2 uses a TI chipset), shouldn't be any problem with it but... who knows?

Maybe time to try a newer router - that's where I'm heading eventually, just need to try and decide which one...

In reply to a post by RobertoS:

In reply to a post by kasg:

In reply to a post by RobertoS:

It's definitely very low for the attenuation. For example look at this graph and put your stats into this estimator.

It's a bit disappointing that, with a 54dB attenuation, I'm not going to get any benefit from ADLS2+

Have you checked the stats using the test socket as I suggested.

(It would help if you clicked the Reply button on posts where I have replied to you, not just my latest post in the forum - most of which are addressed to the OP. As I generally use Threaded mode it is very confusing and I have to search back to find out what you are talking about. Like to find the stats you mention. Even in Flat Mode, if you read the side headers you will see what I mean. There is a link in them to an earlier post, and if it is always just the preceding post it makes that useless).

I'm sorry if I have caused any confusion - I didn't think I had replied to an inappropriate post at any point. I will retire from this thread at this point - to clarify, I don't actually HAVE a problem that I am trying to solve and the comment about ADSL2+ with a 54dB attenuation was just an observation - I do not have an ADSL2+ connection and there seems no point in my attempting to obtain one.

In reply to a post by kasg:

I'm sorry if I have caused any confusion - I didn't think I had replied to an inappropriate post at any point. I will retire from this thread at this point - to clarify, I don't actually HAVE a problem that I am trying to solve and the comment about ADSL2+ with a 54dB attenuation was just an observation - I do not have an ADSL2+ connection and there seems no point in my attempting to obtain one.

It's OK, but it does make things difficult.

If you have attenuation of 54dB on ADSL(1) then that would rise to about 57dB anyway on ADSL2+ . You may find these two links useful:-

http://www2.farina1.com/ADSL/default.aspx
http://www.avforums.com/forums/attachments/networkin...

Where you could gain from an ADSL2+ service would be if there is one of the good LLU ISPs at your exchange, as higher performance is usually obtained than on BT Wholesale based lines even at similar speeds, because of the different systems used.

I suggest if you want to find out more you start a thread in Which ISP with full line stats and the link to what samknows says about your exchange. See the Useful links page on my website for samknows and help finding your stats.

Thanks for all your help and suggestions.

I think I'll leave well alone for now. The speeds I'm getting are fine for my usage and the last time I flashed the DG834 it didn't go very smoothly.

I'm not using wireless networking at the moment. AFAIK, my modem/router can't do WPA2 so at some point I'll probably get a modern "N" capable modem/router.

Edited by ppppenguin99 (Thu 21-Jan-10 08:45:28)

I knew it was too good to last:-

ADSL Link Downstream Upstream
Connection Speed 9122 kbps 1095 kbps
Line Attenuation 22 db 4 db
Noise Margin 6 db 7 db

Attenuation gone up to 22dB? Something very strange there. That could just about account for the speed drop.

In reply to a post by RobertoS:

Attenuation gone up to 22dB? Something very strange there. That could just about account for the speed drop.

Odd isn't it. I reckon about 19000kbps for that attenuation anyway

Yes, we've already covered the fact the speed at 16dB is pathetic.

I think the poster needs to try a different router.

The only other router I have to hand is an old Speedtouch 510. About 2002 vintage but flashed at least reasonably up to date. I don't think it can even do ADSL2+. I think a friend may have a Billion 5200 series modem/router (definitely ADSL2+ capable) spare so I'll try to borrow that.

It's worth a go .

Tried rebooting the modem/router and I've got my old speeds back again. Don't see how rebooting the modem can change the line attenuation.

ADSL Link Downstream Upstream
Connection Speed 12517 kbps 1104 kbps
Line Attenuation 16 db 4 db
Noise Margin 6 db 7 db

I'll try anther modem/router when I get hold of one.

Try that Speedtouch, 510 is ADSL2+ compatible..

The DG534 has been up and down from 9000 to 12000kb/s with attenuation from 16 to 22dB.

Just connected the Billion 5200 and there's an instant improvement.
Downstream Upstream
SNR Margin
6.1 6.3 db
Line Attenuation
17.0 10.3 db
Data Rate
15347 1067 kbps

Thanks for the info on the 510. I'm not 100% sure mine's in good condition. It went intermittent a few months ago so I prodded around inside and changed all the electrolytic caps. This seemed to restore it health but after that it was never quite as solid as the DG534 at 8128kb/s. So I stayed with the DG534 which also has a much nicer HTML UI.

If you re-sync tomorrow well after dawn it will be quite a bit faster.

Why should this happen? I've just re-synced and am getting almost exactly the same speed.

Don't laugh - the sun well above the horizon has a significant effect in reducing background noise. For most installations it varies by 3dB or more.

So if you connect/sync when it is low or or at night you get a lower speed because the noise margin and the background noise both affect the sync speed.

See Noise Margin/SNRM on my website for more detail.

I'm surprised it wasn't better at 9am. Try nearer mid-day, but don't re-sync too often or the BT DLM gets upset. (IIRC and without reading back it is a BT Wholesale connection?)

It does seem a bit unlikely, especially on a line that's entirely underground.. But I suppose anything's possible with cruddy bits of copper with loads of joints down damp holes in the ground. It's still pretty miraculous that ADSL works at all on rotten old POTS lines. It still seems miraculous even though I'm an engineer and understand at least the basics of COFDM, FEC and the other wizardry used to make ADSL happen.

At present, DLM is limiting me to 8000kb/s anyway, presumably because my sync speed has dropped as low as 9000 odd. With a bit of luck my sync will now be stable at 15000kb/s odd so in a few days my DLM profile will get raised.

Dumb me doesn't need to read back on a Plusnet connection - doh!

Profile should rise well within the maximum 5 days with a jump of that magnitude. It would be worth the occasional BT speed test as sometimes that shows a profile rise and the corresponding Plusnet update fails. (Allow several hours extra for that though).

BT speedtest now giving 11343kb/s. Getting better. Perhaps I can expect a little bit more when DLM gets round to it.

Router (Billion 5200) stats seem stable at:
ADSL Firmware Ver FwVer:3.5.10.1_A_TC3085 HwVer:T14.F7_1.0
Modulation ADSL2 PLUS
Annex Mode ANNEX_A

Downstream Upstream
SNR Margin:
6.0 6.0 db
Line Attenuation:
17.0 10.3 db
Data Rate:
14720 1099 kbps

Out of curiosity I tried switching my modem/router to Annex M but it didn't connect at all.