Two lines, but different broadband speeds (was Jos390)

Hi,

I'm a new member here, so I hope I'm asking in the right forum.

We run a business from home and recently took a second line so that we can provide internet to our customers. The new line was taken from the first from where it joins the house, so both lines share the same outer cable from the cabinet (but presumably have different lines?). The old line was installed 10 years ago. Both lines are with TalkTalk Business and often have similar IP addresses (both start with 92.15.43. today, for instance).

However, the new line is usually about 1Mb faster than the old and the speed is far more consistent. The new line runs consistently at about 3Mbps whilst the old line varies between 0.5-2.9Mbps. The old line has always had periods of slower speeds followed by periods of good speeds, so sometimes we can stream TV and sometimes not. It has been too slow for streaming for about 2 weeks so far this time.

Things I have tried: I have plugged the router into the BT test socket and have tried connecting to the router with a cable instead of WiFi but the speed is always about the same. I have tried 3 different routers.

Can the wiring of the BT socket or the cable from the cabinet make this type of difference? Or is the difference caused by some setting in the exchange or elsewhere? Can equipment near the BT socket cause problems?

Any help with this much appreciated - teenagers are revolting!

Edited by deleted (Tue 18-Nov-14 16:51:26)

Once the cable gets to the pole, the lines could then be in separate cables to the cabinet and then from the cabinet again, separate cables are possible and even different routes. Even if the lines are in the same cable one could be close to another noisy line ... so yes, different speeds are quite possible.

However, as one is stable and the other varying does suggest either an intermittent fault or an intermittent source of noise. And yes, local equipment can cause the problem.


First stage is to get line stats - Attenuation and SNR (margin) for both lines at times when they are both running well and also when one has a problem. That will be a start.

Have two lines and one was nice and stable and one would be less stable in the winter months, different drop wires to the property but go to the same pole.

When things were good speeds with ADSL2+ were similar.

Since getting FTTC there is very little between the two lines in terms of performance.

So original posters situation is far from unique, but as suggested line stats for both lines and then doing the usual ring wire, better quality RJ11 modem lead and avoiding putting modem in an electrically noise area are all good things to do.

Hi, thanks for your reply. I have copied the current connection details below. If/when the old line improves again I'll check the details and post the results.

NEW LINE

DSL Type: ITU-T G.992.3
Bandwidth (Up/Down) [kbps/kbps]: 945 / 3.581
Data Transferred (Sent/Received) [GB/GB]: 1,07 / 1,23
Output Power (Up/Down) [dBm]: 12,6 / 0,0
Line Attenuation (Up/Down) [dB]: 31,1 / 53,0
SN Margin (Up/Down) [dB]: 6,7 / 4,7
System Vendor ID (Local/Remote): TMMB / ----
Chipset Vendor ID (Local/Remote): BDCM / IFTN
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): -
Error Seconds (Local/Remote): 31.072 / 0
FEC Errors (Up/Down): 171.482 / 1.463.354
CRC Errors (Up/Down): 9.358 / 101.367
HEC Errors (Up/Down): 8.240 / 828.733


OLD LINE

DSL Type: ITU-T G.992.1
Bandwidth (Up/Down) [kbps/kbps]: 800 / 2.400
Data Transferred (Sent/Received) [MB/GB]: 68,71 / 1,99
Output Power (Up/Down) [dBm]: 12,6 / 17,8
Line Attenuation (Up/Down) [dB]: 31,5 / 56,5
SN Margin (Up/Down) [dB]: 10,0 / 8,6
System Vendor ID (Local/Remote): TMMB / ----
Chipset Vendor ID (Local/Remote): BDCM / IFTN
Loss of Framing (Local/Remote): 0 / -
Loss of Signal (Local/Remote): 0 / -
Loss of Power (Local/Remote): 0 / -
Loss of Link (Remote): -
Error Seconds (Local/Remote): 39 / -
FEC Errors (Up/Down): 277 / 32.198
CRC Errors (Up/Down): 183 / 346
HEC Errors (Up/Down): -83 / 346 / 3.443

Best wishes,
Jos

Edited by deleted (Tue 18-Nov-14 17:01:04)

Some of the speed difference is likely to be from the fact the old line is running at an 8.6 SN margin and the new is at 4.7.

The 4dB difference between the 2 could be making a reasonable difference on connection speeds. What you may find though is that the new line may have more errors and be a little more unstable and may well slow down over time in order to improve the stability.

Looking at the errors (bearing in mind we don't know how long these have been up) the new line does seem to have more errors than the old - do you have the uptime for the 2 connections so we know if we are comparing like for like?

In reply to a post by ian72:

Looking at the errors (bearing in mind we don't know how long these have been up) the new line does seem to have more errors than the old - do you have the uptime for the 2 connections so we know if we are comparing like for like?

The new line was installed a year ago and the current connection has been up for 15 days. I have restarted the old line's modem a couple of times today (sometimes it re-connects at a higher speed) so it has only been up for a couple of hours.

Best wishes,
Jos

It could just be the DLM system is treating the 2 lines differently. The SNR on the new line is definitely lower than would normally expect. You may find if you rebooted the router for it that it would reduce the speed as the SNR target is probably 6dB.

Rebooting both at the same time and getting stats could show the real differences between the 2 and would give us a level playing field to compare error rates.

In reply to a post by MrSaffron:

So original posters situation is far from unique, but as suggested line stats for both lines and then doing the usual ring wire, better quality RJ11 modem lead and avoiding putting modem in an electrically noise area are all good things to do.

Thanks for your reply. I've just switched the RJ11 modem leads and the old line is still slower than the new. But if you think it might speed up both connections I'll certainly try a different RJ11 lead - how do I tell a good from a bad lead if I look for one?

I disconnected the ring wire about a year ago.

The modem is located on my desk next to a laptop, a laser printer, a desktop system and a 4 socket extension for the laptop printer and router. Is this the sort of equipment that will cause electrical noise problems? For comparison, the new line's modem is next to a wireless access point, a couple of switches, a network drive and 2 x 4 socket extensions (for all these items plus a couple of wireless access point located on the roof). The router/modems are the same model, Technicolor TG 852n. We had the same problems with the old line with an old Netgear router.

Best wishes,
Jos

In reply to a post by ian72:

Rebooting both at the same time and getting stats could show the real differences between the 2 and would give us a level playing field to compare error rates.

Do a couple of minute matter? These are the stats after switching the RJ11 leads:


NEW LINE

Uptime: 0 days, 0:25:25
DSL Type: ITU-T G.992.3
Bandwidth (Up/Down) [kbps/kbps]: 904 / 3.339
Data Transferred (Sent/Received) [GB/GB]: 1,08 / 1,24
Output Power (Up/Down) [dBm]: 12,7 / 0,0
Line Attenuation (Up/Down) [dB]: 30,8 / 53,5
SN Margin (Up/Down) [dB]: 6,0 / 5,9
System Vendor ID (Local/Remote): TMMB / ----
Chipset Vendor ID (Local/Remote): BDCM / IFTN
Loss of Framing (Local/Remote): 9 / 0
Loss of Signal (Local/Remote): 1 / 0
Loss of Power (Local/Remote): 0 / 0
Loss of Link (Remote): -
Error Seconds (Local/Remote): 31.099 / 0
FEC Errors (Up/Down): 61 / 230
CRC Errors (Up/Down): 6 / 16
HEC Errors (Up/Down): 3 / 47

OLD LINE

Uptime: 0 days, 0:27:33
DSL Type: ITU-T G.992.1
Bandwidth (Up/Down) [kbps/kbps]: 832 / 2.304
Data Transferred (Sent/Received) [MB/GB]: 87,13 / 1,99
Output Power (Up/Down) [dBm]: 12,6 / 17,6
Line Attenuation (Up/Down) [dB]: 31,5 / 56,5
SN Margin (Up/Down) [dB]: 10,0 / 8,9
System Vendor ID (Local/Remote): TMMB / ----
Chipset Vendor ID (Local/Remote): BDCM / IFTN
Loss of Framing (Local/Remote): 9 / -
Loss of Signal (Local/Remote): 1 / -
Loss of Power (Local/Remote): 0 / -
Loss of Link (Remote): -
Error Seconds (Local/Remote): 50 / -
FEC Errors (Up/Down): 87 / 3.872
CRC Errors (Up/Down): 22 / 0
HEC Errors (Up/Down): -2 / 0 / 2

Thanks,
Jos

Interestingly the new line is now on ADSL2 (G.992.3). Both were previously on ADSL (G.992.1).

There is a lot of potential for noise from all of that. How far from te master sockets are the two modems?

I always recommend that the modem is located as close as possible to where the line enters the house to minimise any problems from noise or wiring issues. I know you have a combined modem/router but that can be overcome.


A quick look at your stats does not show anything significant - the downstream SNR margins are different but that may be because the DLM has intervened.

Edited by MHC (Tue 18-Nov-14 17:00:33)

In reply to a post by RobertoS:

Interestingly the new line is now on ADSL2 (G.992.3). Both were previously on ADSL (G.992.1).

Sorry, my copy and pasting error, I've corrected it now!!!

In reply to a post by MHC:

I always recommend that the modem is located as close as possible to where the line enters the house to minimise any problems from noise or wiring issues. I know you have a combined modem/router but that can be overcome.

This is potentially an issue, though I have plugged the router/modem directly into the BT test socket and it connected at the same speed. But maybe I would need to leave it for a few days to allow the DLM to adjust? Would any useful information be gained by plugging it into the test socket and checking the stats?

The old line's BT box is in my daughter's bedroom and our small hard-wired system does not include that room (because it used to be a workshop/store room) so we may have problems if we need to move the router/modem. And it also doesn't explain why it's OK and usable for long stretches (though it's always a little slower than the new line).

Thanks,
Jos

In reply to a post by RobertoS:

Interestingly the new line is now on ADSL2 (G.992.3). Both were previously on ADSL (G.992.1).

Given that the new line is on ADSL2 (G.992.3) and the old line is on ADSL (G.992.1), could this be the cause of the speed difference?

Thanks,
Jos

In terms of leads

This lead on Amazon would fit the bill, the twisted pair is the key bit. No need for fancy gold or oxygen free stuff.

In reply to a post by Jos390:

Given that the new line is on ADSL2 (G.992.3) and the old line is on ADSL (G.992.1), could this be the cause of the speed difference?

That and the old line is 3dB longer in attenuation.

They are both underperforming by about 1 Meg.

Is it possible that old line is on a legacy 'Up to 8 Meg' product?

In reply to a post by XRaySpeX:

Is it possible that old line is on a legacy 'Up to 8 Meg' product?

Going by the upstreams it shouldn't be as the upstream should be fixed to 448 on legacy?

Another thought. Are they exactly the same make and model of router? If the new line has a different router then that may affect the performance of the connection.

I still think the fact the old line seems to have a 3dB higher target SNR will be affecting the speed by a reasonable chunk.

Might also be worth posting another set of stats now that they have been on for a while just to see that the error numbers are not wildly different for the 2 lines.

In reply to a post by XRaySpeX:

That and the old line is 3dB longer in attenuation.

I would say the lines are not much different in length - look at the UPstream attenuations 31.1 to 31.5 This is a better comparison of line length - assuming same wire gauge and type. The loss of tones due to distance and attenuation only affects downstream. The difference in downstream could well be due to measurement techniques used and across the tones in use or available.

In reply to a post by ian72:

Going by the upstreams it shouldn't be as the upstream should be fixed to 448 on legacy?

You are thinking BT 20CN ADSL Max. This is LLU. LLU ops used to offer 8 Meg capped products with the upstream uncapped. AOL LLU did so until recently and years ago I had such from Orange LLU. Tho' admittedly they were usually run on ADSL2+. It was just a thought to explain the diff ADSL Modes.

Could be - thought that the increased upstream was only with ADSL2+ style services rather than the ADSL capped services.

ADSL Max has an 832kbps upstream option, normally only available on "business" products.

In reply to a post by MrSaffron:

In terms of leads

This lead on Amazon would fit the bill, the twisted pair is the key bit. No need for fancy gold or oxygen free stuff.

Thanks - have ordered one to try.

Jos

In reply to a post by XRaySpeX:

Is it possible that old line is on a legacy 'Up to 8 Meg' product?

Yes. I rang TalkTalk yesterday and the old line is the ADSL up to 8 Meg product and I was told it cannot be moved (the new line is on ADSL2). Don't know why I didn't ask why it can't be moved! But would it make that much difference? Looking at the figures we are about 4 miles from the exchange and graphs I've seen for the speed differences between ADSL 1 & 2 seem to converge at about 4 miles.

Best wishes,
Jos

I rang TalkTalk yesterday and they adjusted the DSL settings. He said that this may lead to more drop-outs, but it doesn't seem to have been a problem so far. The line seems a bit faster. Before he changed the setting yesterday the speed had dropped to about 0.4 Mbps (even though the bandwidth figure in the stats remained unchanged). Here are the current stats - sorry the up times are different.

OLD LINE

Uptime: 0 days, 11:56:21
DSL Type: ITU-T G.992.1
Bandwidth (Up/Down) [kbps/kbps]: 896 / 2.560
Data Transferred (Sent/Received) [MB/MB]: 41,44 / 512,84
Output Power (Up/Down) [dBm]: 12,7 / 17,6
Line Attenuation (Up/Down) [dB]: 31,5 / 56,5
SN Margin (Up/Down) [dB]: 7,0 / 7,2
System Vendor ID (Local/Remote): TMMB / ----
Chipset Vendor ID (Local/Remote): BDCM / IFTN
Loss of Framing (Local/Remote): 0 / -
Loss of Signal (Local/Remote): 0 / -
Loss of Power (Local/Remote): 0 / -
Loss of Link (Remote): -
Error Seconds (Local/Remote): 301 / -
FEC Errors (Up/Down): 1.083 / 1.573.902
CRC Errors (Up/Down): 212 / 252
HEC Errors (Up/Down): -12 / 252 / 1.675


NEW LINE

Uptime: 0 days, 20:51:20
DSL Type: ITU-T G.992.3
Bandwidth (Up/Down) [kbps/kbps]: 904 / 3.339
Data Transferred (Sent/Received) [GB/GB]: 1,34 / 1,99
Output Power (Up/Down) [dBm]: 12,7 / 0,0
Line Attenuation (Up/Down) [dB]: 30,8 / 53,5
SN Margin (Up/Down) [dB]: 5,5 / 6,4
System Vendor ID (Local/Remote): TMMB / ----
Chipset Vendor ID (Local/Remote): BDCM / IFTN
Loss of Framing (Local/Remote): 9 / 0
Loss of Signal (Local/Remote): 1 / 0
Loss of Power (Local/Remote): 0 / 0
Loss of Link (Remote): -
Error Seconds (Local/Remote): 31.343 / 0
FEC Errors (Up/Down): 3.308 / 9.969
CRC Errors (Up/Down): 408 / 530
HEC Errors (Up/Down): 284 / 2.962


Thanks,
Jos

The speed dropping to 0.4Mb/s without the connection speed changing could suggest either errors on the line or congestion. It's possible that as it is on their old up to 8Mb/s service that it may use different parts of their network and they may be congested at times causing the slow downs.

The FEC errors are pretty high so that could be causing slowdowns whilst it is correcting them?

In reply to a post by Jos390:

(the new line is on ADSL2)

Do you mean ADSL2+? It is more normal that it is provisioned on ADSL2+ but happens to be running on ADSL2 as decided by the DLM.

In reply to a post by ian72:

The speed dropping to 0.4Mb/s without the connection speed changing could suggest either errors on the line or congestion. It's possible that as it is on their old up to 8Mb/s service that it may use different parts of their network and they may be congested at times causing the slow downs.

Thanks, I'll check again at busy times.

In reply to a post by ian72:

The FEC errors are pretty high so that could be causing slowdowns whilst it is correcting them?

The most recent stats were taken after TalkTalk changed the DSL settings. Could the increase in speed be causing the FEC errors?

Thanks,
Jos

In reply to a post by XRaySpeX:

In reply to a post by Jos390:

(the new line is on ADSL2)

Do you mean ADSL2+? It is more normal that it is provisioned on ADSL2+ but happens to be running on ADSL2 as decided by the DLM.

Yes, I think so.

Thanks,
Jos

Why do you keep changing the subject back to your nick? The Subject is supposed to be be informative about the matter being discussed which isn't yourself.

In reply to a post by XRaySpeX:

Why do you keep changing the subject back to your nick? The Subject is supposed to be be informative about the matter being discussed which isn't yourself.

Sorry about that. I haven't changed the subject back to my nick. When writing this post in reply to yours I had to change it from my nick, despite clicking reply to your post which has the correct title. I hadn't noticed this happening when replying previously or I would have changed those too. I thought I'd changed the title in my first post and had to edit the post once I noticed. This is probably why my nick is still automatically used in any replies.

Bws,
Jos

Edit: PS I've just replied to another message and Jos390 appeared automatically as the subject again! I changed it to the correct subject and posted it, but it changed it back to Jos390 and I had to go in and edit the post to correct it again. Every time I post, reply or edit a post the subject is always Jos390. Is this a common problem or a Firefox issue?

Edited by deleted (Wed 19-Nov-14 15:41:03)

Peculiar! I just replied to your post w/out changing Subject and it carries straight thro'.

I wonder if there's a special rule for OPs to carry forward Subject of OP? Never noticed that before.

EDIT: No, there isn't! Just tested it on 1 of my OPs.

Edited by XRaySpeX (Thu 20-Nov-14 00:22:08)

I think it's something to do with the automatic form filling 'feature' in Firefox. If I delete my nick from the subject and press the down arrow whilst the cursor is in the text box, all of the titles I've used so far are listed and available to select - but Jos390 is the top of the list and is automatically selected first. I'll just have to remember to check in future.

best wishes,
Jos

In reply to a post by Jos390:

and press the down arrow whilst the cursor is in the text box

You shouldn't be pressing any down arrow. You just need to delete or highlight what's already there and just start typing.