Another O2 3dB profile question thread

Man of Honour
Joined
5 Apr 2009
Posts
26,454
Code:
Uptime:						4 days, 16:48:31
DSL Type:					G.992.5 annex A

Bandwidth (Up/Down) [kbps/kbps]:		1,235 / 3,944
Data Transferred (Sent/Received) [kB/kB]:	0.00 / 0.00
Output Power (Up/Down) [dBm]:			12.0 / 17.5
Line Attenuation (Up/Down) [dB]:		24.5 / 46.0
SN Margin (Up/Down) [dB]:			6.5 / 7.5

Vendor ID (Local/Remote):			TMMB / BDCM
Loss of Framing (Local/Remote):			0 / 0
Loss of Signal (Local/Remote):			0 / 0
Loss of Power (Local/Remote):			0 / 0
Loss of Link (Remote):				0

Error Seconds (Local/Remote):			7 / 0
FEC Errors (Up/Down):				1,315 / 911,100
CRC Errors (Up/Down):				1,315 / 4,167
HEC Errors (Up/Down):				86,177 / 2,243

Significant benefit from trying 3dB or will it be pointless?

I'm a bit concerned about the number of errors at the bottom though i've no idea what significance any of them carry :p
 
Last edited:
I would say no, my friend has very similar attenuation to you (47.5db) and is on the standard 6.5db noise margin. He syncs up in the morning at around 7000 (6Mbps throughput) but by are 11pm in the evening he is on 0.5db and he eventually drops out and re-syncs around 5500. The onll way to keep his line stable was to boot up in the morning, wait for it to drop to 3db, then re-boot it again. That way it would just stay stable through the night.
 
You'll always get a benefit from reducing the target SNR margin, it's whether you'd be left with a connection that was usable or not. Either way, you only have to ask them to change it back.
 
I agree, you should definately request the 3 dB profile and give it a try.

In my opinion all the router stats are there for information but what matters is the perceived quality of the connection; is it stable enough for the user and does it perform to their satisfaction.

My understanding of the errors are that FEC errors are corrected and "often over reported by the O2 box" and thus nothing to worry about. CRC and HEC errors are ultimately uncorrected and, while I could easily be mistaken, provide redundancy checking for the ATM cell payload (48 bytes) and ATM cell header (5 bytes) respectively. My guess is therefore that one CRC error is equivalent to one corrupt 48 byte ATM cell. So, 4000-ish downstream CRC errors in four-and-a-half days is a miniscule fraction of bad data for an active user.

Back to the 'quality' of your connection, with respect to gaming I think we would watch out for packet loss and any effect this had - 'smoothness' of display is a characteristic I'd associate with that. For web use check that the thoughput is as expected; look out for trouble with streaming YouTube/iPlayer (bearing in mind these can be bad based on the time of day) and whether you can 'max out' the connection with a reliable high-bandwidth server such as one serving Linux distros.

I have a long line with 58-59 dB attenuation and in the first few months of this year have seen 3400 kbps @ 9 dB, 3900 kbps @ 6 dB and around 4600 kbps @ 3 dB with a Bebox. I moved to an ex-Sky DG834GT and gained a few hundred kbps on the 3 dB profile and now that the season moves on I have found it increase a little bit more, especially if it syncs in the morning. I'm not advocating restarting the router every day, mind.

@Ryan111: Wouldn't it be wise for your friend to try and determine just why the noise margin drops so significantly each evening. There are a number of threads around with troubleshooting advice and mentions of diagnostic tools, in addition to getting the ISP involved.
 
@Ryan111: Wouldn't it be wise for your friend to try and determine just why the noise margin drops so significantly each evening. There are a number of threads around with troubleshooting advice and mentions of diagnostic tools, in addition to getting the ISP involved.

How would he go about this? I am going to try and get him to run router stats for 24hours and see the exact change in SNR

He is currently using the default thompson O2 router, would he benfit from a new one? A DG834GT maybe?
 
Going to ask this question again, His line will sync at 7400 in the morning, but then by the evening the SNR drops soo low it eventually cuts out. Any reason for this?
 
It's normal for the SNR margin to fluctuate. If you connect when it's at its highest it's not unreasonable that you'll lose sync when it gets to the lowest point.
 
Ryan111, sorry for the delay in replying.

Using Routerstats (the Lite version will do) is one way of monitoring the modem's SNR margin over time. What you can look for is whether there is a recurring point during the evening at which sync is lost. You might be able to differentiate, for example, whether there is an event which causes an instantaneous drop in SNRM or whether it just gradually decreases until the modem can't hang on.

Another software which people play with is the DSL Modem Tool (DMT). Version 7.35 is compatible with many Thomson devices. I can't vouch for its utility in actually tracking down a connection problem although it does provide a measure of any frequency ranges which are not "being heard" by the modem. Apparently, a couple of years ago you used to hear of high level BT 'SFI' engineers who would help in locating the source of radiofrequency interference. By some accounts this is sadly very rare these days; the SFI service now provided seems dumbed down and genericised. My personal feeling is that when real line problems are present you're more likely to get it resolved if you are a clued-up user.

Asking why the SNRM drops low and causes LOS is quite a broad question. Kitz has already covered it so I'll direct you there:

There are some notes there, derived from their forum, on the fact that some modem/routers hold sync with lower noise margin than others. I really don't know how common this is but I will say that coming from the standard Bebox (Thomson 585v7) I have managed to get more speed (lower target SNRM, aka 'profile') and keep it satisfactorily stable on my high-attenuation line.

Heh, I've just realised if I had started with the Kitz links I could have saved myself a lot of typing. I hope this helps :)
 
Last edited:
^^ Very helpfull.

He is running routerstats now, he is currently connected at his lowest sync so wont disconnect throughout the day/night. Will post the results back in 24 hours.
 
If he isn't already, the modem should be connected at the master socket with any extensions on the other side of a filter.
 
He is connected at the master socket with two hardwired extentions coming from it, both with the ring wire removed. All sockets that have a phone connected have a filter. Would it be worth swapping filters?
 
In which case try the test socket, i.e. so the extensions are removed from the equation.
Changing filter isn't going to hurt but they probably won't change anything.
 
Back
Top Bottom