New Netgear DNG2200

I have had a Zyxel Prestige 660HW-T1 router for a few years, and I have just bouught a Netgear DNG2200.

The router stats on the old Zyxel were

Noise margin 20 db up 6 db down
Attenuation 31 db up 53 db down
speed 448 3904

with the new DNG2200 the stats are

ADSL Link Downstream Upstream
Connection Speed 3168 kbps 448 kbps
Line Attenuation 61.0 db 29.0 db
Noise Margin 12.3 db 19.0 db

It looks like the noise margin on the DNG2200 is double that of the noise margin on the zyxel and hence why the connection speed is less.

I have tried rebooting, unplugging, the routers a couple of times but I still get the same stats.

Why should the new DNG2200 connect with a different noise margin than the Zyxel.

I was hoping that the new router would have given me a better connection speed than the zyxel.

Stephen

The noise margin is not double. How do you work that out?

The downstream noise margin for the Zyxel is 6, but for the Netgear its 12.

It's a logarithmic scale. More like 5 times.

If you want to be pedantic about it, since its logarithmic, assuming base 10, 12 is 1 million times bigger than 6, or if you are working in natural logs its 400 odd time bigger. But since most people look at the absolute figures quoted by the router, then I say it is double.

If you drop the zyxel back in now its still showing the 6dB? Have a dgn2200 and not seeing anything resembling this behaviour.

Suggestion is ISP has raised your target noise margin to 12dB now.

I have tried swapping the modem cable a couple of times between the routers and it still gives me 6db on the Zyxel and 12 on the DNG2200.

I'm trying to remember from the back of my mind if I might have tweaked the noise margin target on the Zyxel sometime in the past, but I don't think so.

What I can remember is that when I got the Zyxel it was to replace a Belkin router, and the noise margin dropped immediately when I started using the Zyxel because I had been having problems with the Belkin that had caused the noise margin target to be raised at the exchange ( my connection speed went from about 2.5meg to 3.5meg the instant I started using the zyxel). It did look at the time like the Zyxel was targeting a lower noise margin than that set at the exchange, becasue the Belkin had been about 12 and the Zyxel was 6. Would that be possible?

Is it possible to tweak the noise margin target on the DNG2200, or can I check somewhere what my target noise margin is and can I get that changed with my ISP (ADSL24)

In reply to a post by BatBoy:

It's a logarithmic scale. More like 5 times.

That would be kind of true (ish) if you were expressing noise margin as the signal strength difference between signal and noise (as a measurement of power).
Even then, and using that logic, it would be four times greater (approximately double the signal/power for every 3dB).
But expressed as a dB figure (which it is), it's just double.

Well that would be the case if you were hearing the noise, it would sound twice as loud - that's the idea behind a logarithmic scale. Unfortunately in this case, it's affecting a device which is hearing five times the noise. Unlucky

I'd stick with the Zyxel...

In reply to a post by soapysoutar:

I'm trying to remember from the back of my mind if I might have tweaked the noise margin target on the Zyxel sometime in the past, but I don't think so.

Didn't think you could do any tweaking on Zyxels unless you know some cli stuff that I don't I've used a couple of them in the past and found them to be very good, despite the fact the majority use an AR7.

Is it possible to tweak the noise margin target on the DNG2200, or can I check somewhere what my target noise margin is and can I get that changed with my ISP (ADSL24)

If you can track down a DGteam firmware for your DNG you can tweak the SNRM yourself. Luckily you (we) are with ADSL24, and they should be able to help.

DrT

On second thoughts I know I did not tweak anything on the Zyxel. I can look at settings on the zyxel via telnet, but I can't change anything.

I have just got a DGN2000 and am wondering if anyone has any experience of using the DGTeam firmware on it. I have read some instances of it locking up the router but am not sure if they are referring to the 2000 or 834.

I have found a link to the 1018 version here

http://jake-tm.co.uk/?page_id=259

I am able to tweak the SNR target by telnet'ing in and using the command

adslctl configure --snr 1

When I reconnect the line, this takes my noise margin reported on the DNG2200 down to about 6, and the sync speed up to about 4300 (from 3300).

However, that tweak is not permanent, and if I reboot the router the the twaek is lost and I reconnect with a noise margin of 12 and a sync speed of 3300ish. So I don't see this tweak as a solution.

Can I make the tweak permanent without finding some other un-official firmware? Or are there other routers that are good for long lines and easiy allow the noise margin target to be tweaked, that I should look at instead.

Edited by deleted (Sat 13-Nov-10 18:03:01)

The tweak should stick with the DGTeam firmware. That's all I used on my 834GT and bought it specifically so I could use DGTeam wares.

DrT

The easiest way of finding your connection time noise margin is to read the stats immediately after a connection is established, unless your router records it in the log. I'm fairly sure your Netgear will not.

From your earlier posts it looks possible you have done multiple reconnections in a short period. Assuming you are on a BT Wholesale based service this could have triggered the DLM into raising it from 6dB to 12dB.

Tweaking to --snr 1 is extreme. Starting from a 6dB setting (if done in the daytime) you would almost certainly (on BT Wholesale) cause a rise as above. Certainly a permanently lowered IP Profile.

If you are now starting from 9dB or 12dB then you may get away with it, but your speeds are already severely compromised.

If you find your sync-time margin is 9dB and apply --snr 50 I think you will find you get around 6dB result. If you are starting from 12dB you will get about 9.5dB.

Just to clarify, the netgear currently reads a noise margin of 12 when I reconnect with it. If I then use the Zyxel I get 6 when I reconnect. Switching back to the netgear then gives 12, and going back to the Zyxel gives 6. I get different connection noise margins depending upon which router I use.

On the netgear, using the --SNR 1 command took my connection noise margin down to about 6.2 or so (starting from 12)

I have not tweaked the zyxel, so why should it connect at a different noise margin than the netgear?

It cheeses me off a bit because I would rather use the DGN2200, but as the Zyxel connects at a higher speed, I think I will need to keep using that and just return the Netgear

I have raised a ticket with my ISP to see if they can get the target SNR reset to 6, and they have came back with the following comment from their supplier;

"They also advise that the target SNR for the line and can vary upon different equipment (which might explain the disparity on stats between the 2 modems)."

Is this the case? or is this lust a fob off from my ISP's supplier?

The noise margin your router reports is an average of the margin on each tone/frequency available. The attenuation reported is also an average of the attenuation of each tone/frequency.

They are correct to say that different routers (actually the modem component) can calculate the averages differently. For example I changed from a Linksys router to a SpeedTouch one and my reported attenuation fell by 3dB.

However they are incorrect to say the target (sync-time) margin varies depending on the equipment. That is a figure set in the database referenced by the DSLAM at sync time.

Have a read of my Noise margin page to see how the two relate.

Re-reading your opening post, your logic is correct in a way, but there is something odd happening which right now I can't put my finger on. If during the sync negotiation the router reports different values for attenuation and SNR (note not SNRM = margin) I can understand the sync speed being different. But not the reported margin straight after the connection is made.

Also note the expected speed difference per 3dB on my web page.

One of those routers is mis-reporting to you and probably to the DSLAM. I suspect the Zyxel.

Why did you get the Netgear? Have you been getting even more instability with the Zyxel than with the Belkin. Have you still got the Belkin and if so what does that report? I think it will show 12dB if you have it.

Incidentally your quote from ADSL24's supplier looks to have a bit missing.

Whilst typing all the above, and looking back through the thread, I begin to think you have a long-standing line problem of some sort that you are trying to solve. But looking in the wrong place for the solution.

For info, the full reply from my ISP was

"Our suppliers have tested the line & advise they see no faults on the service being flagged up, but note that the line is a long line and no fault conditions can be seen on the connection that would cause a low sync rate
They also advise that the target SNR for the line and can vary upon different equipment (which might explain the disparity on stats between the 2 modems). "

I got the Netgear router because I have been starting to have wireless connection issues with the Zyxel, and the ADSL sync speed had drifted down a bit over the last couple of months. I realise that the connection speed would probably not be a router issue, but I thought I would change anyway. And also, I had only ever been able to get the Zyxel to work using WEP and I wanted a router that used better encryption.

With regards to the old Belkin router, I changed from that because it had started to have problems with it. The Belkin used to connect with a noise margin of 6db, but I had a connection issue one day and I did a lot of router resets (I did not know at the time that would affect the target SNR) and my noise margin increased to 9, then 12 (and the sync speed dropped accordingly). So, I changed to the Zyxel and without doing anything the SNR dropped back to 6 (and the sync speed went up). I have not tried the Belkin again, but I agree with what you suspect that it will connect at 12 as well.


I have read many times, that the target SNR is meant to decrease automatically if the line is stable. But I have also read that that is not always the case. So, if I keep the Netgear connected at a noise margin of 12, will the exchange eventally drop my target down. My connection has been stable over the last couple of years, but as I've always connected at 6db, has the exchange been getting fooled into thinking it doesn't need to drop the target?

They also advise that the target SNR for the line and can vary upon different equipment (which might explain the disparity on stats between the 2 modems

Well the missing bit seems to be "...for the line can and does vary ...". But even when corrected, the statement is simply wrong. Barring faulty equipment such as we suspect the Zyxel may be.

...as I've always connected at 6db, has the exchange been getting fooled into thinking it doesn't need to drop the target?

This seems to me quite likely, though with a roundabout sort of logic.

I would be fairly sure the Zyxel is reporting 6dB when the target figure is actually 12dB. Quite how you get the decent sync I'm not sure, but that will be down to how the software at each end negotiates.

It is relatively simple these days for ISPs to get the noise margin reset, but if you do keep the Netgear connected then you should find out in 10-14 days whether the DSLAM is going to lower the target automatically. A longer wait, but less hassle if it works.

Just to update this, tonight I checked the stats on the DNG2200 and the noise margin was down to 9db and the sync speed up to 3774. A reboot of the router gave the same stats.

The router been connected for over 12 days without a resync, so it looks like the noise margin has been lowered at the exchange. I just hope that in another 2 weeks or so the target will drop back down to 6db.

A day-time connection would give a bit more speed, but don't do several!

Today it looks like my target noise margin has been reset back down to 6db, as my router re-connected this morning with the following stats

Connection speed 4416 kbps
Line attenuation 60db
Noise margin 6.3db

This change is just over 14 days since the last step change of the noise margin, so it does look like the automatic system for dropping target noise margin does work.

Nice to know it is still working.

Your Zyxel certainly seems to have been fooling the system.

4416kbps on 60dB attenuation is really good! It was worth waiting to see what the system did I reckon.

One of the reasons I have been reluctant to replace my 2Wire BT 2700HGV (see other thread) is its ability to run at low SNRs.

It is currently showing
Down Up
Current Rate: 7392 kbs 448 kbs
Current Noise Margin: 4.0 dB 17.0 dB
Current Attenuation: 30.6 dB 16.0 dB

The BT1800 offers the same characteristics.

> Tweaking to --snr 1 is extreme. Starting from a 6dB setting (if done in the daytime)
> you would almost certainly (on BT Wholesale) cause a rise as above. Certainly a
> permanently lowered IP Profile.

... interested to see you say about permanent profile lowering. I've just started tweaking SNR because frequent power cuts make the line seem unstable to the DLM. My 6dB connect rate is 2.2 meg G.DMT due to long line length. When the SNR margin goes to 15dB I'm often down to 700kbps and the wait for it to increase seems like an eternity. Just got a DG834GT, loaded DGTeam connected @ 15dB then changed the percentage to 5% to give a daytime 6dB margin (night is about 4dB). This gave me a sync connect of 1.8 meg.

I'm a bit concerned, can you elaborate on what you mean?

I'm with BT for phone and broadband. Attempting to get a fixed SNRM out of India is like pulling teeth. Whenever you mention slow internet they suggest change wifi channel to better it