|
|
I would appreciate some opinions on the likely cause of the following set of results.
First, a little background.
I live about a mile and a half from my local exchange.
I usually sync at around 3500-4000Kbps and have a 50db down attenuation.
I have been with a well known and, I think, reliable reseller for a number of years and the service has been faultless.
A month or so ago I was moved to the 21CN system and almost immediately I synced at a much higher rate, over 5000Kbps and my throughput down and up increased a lot.
This was great for a few weeks until suddenly I noticed the download rate falling of markedly in the evenings and weekends. I raised the matter with my ISP who, after a week of having me run test after test has now decided that I have a local issue and want to send a BT engineer to check my equipment at a possible cost of £200 or more.
Before I risk what is not an insignificant sum of money to me, perhaps those among the forum could suggest if it is likely to be local. I also fear that the BT engineer will turn up at 10am on a weekday when all is working well.
My adsl router is a Thomson Speedtouch ST585
It is never turned off, except on occasional and rare lockup or power cut, is presently synced at 736 / 5,440 and my profile a week ago was 4000 but is presently 3500.
Here are my results for the last seven days.
(BT) is BT speedtest
(ST) is Speedtest.net
Any opinions would be welcome.
Sorry about the crazy formatting but I hope it is comprehensible.
Date Time Down Up
Monday
17th Jan 10:26am (ST) 4.81Mb/s 0.59Mb/s
18:26pm (ST) 0.72Mb/s 0.59Mb/s
Tuesday
18th Jan am (BT) 4576 Kbps Not Tested
17:50pm (BT) 545 Kbps Not Tested
17:45pm (ST) 0.67Mb/s 0.59Mb/s
Wednesday
19th Jan 12:45am (ST) 4.62Mb/s 0.56Mb/s
12:50am (BT) 4482Kbps Not Tested
17:20pm (BT) 186Kbps Not Tested
17:21pm (ST) 0.51Mb/s 0.60M/bps
19:33pm (ST) 0.55Mb/s 0.58M/bps
Thursday
20th Jan 14:40pm (ST) 2.33Mb/s 0.60Mb/s
16:05pm (ST) 2.60Mb/s 0.60Mb/s
16:15pm (BT) 3400Kbps Not Tested
16:35pm (ST) 0.85Mb/s 0.58Mb/s
17:39pm (BT) 405Kbps Not Tested
Friday
21st Jan 14:42pm (ST) 4.82Mb/s 0.56Mb/s
18:30pm (ST) 0.27Mb/s 0.59Mb/s
18:35pm (BT) 157Kbps Not Tested
Saturday
22nd Jan 10:04am (ST) 2.01Mb/s 0.60Mb/s
10:09am (BT) 1850Kbps Not Tested
14:50am (BT) 553Kbps Not Tested
15:50pm (Broadbandchoices)
0.5Mb 0.6Mb
18:05am (ST) 0.78Mb/s 0.59Mb/s
Sunday
23rd Jan 10:14am (ST) 1.31Mb/s 0.60Mb/s
13:04pm (ST) 1.79Mb/s 0.58Mb/s
17:12pm (ST) 0.40Mb/s 0.61Mb/s
17:18pm (BT) 247Kbps 585Kb/s
Monday
24th Jan 09:46am (ST) 4.27Mb/s 0.58Mb/s
09:54am (BT) 3459Kbps Not Tested
19:10pm (ST) 0.43Mb/s 0.57Mb/s
19:16pm (BT) 294Kbps Not Tested
The ping is consistently around 85-100ms
Edited by deleted (Mon 24-Jan-11 23:48:42)
|
|
|
Can you post your full router stats? and possibly the full result of a BT speedtest?
Matt
|
|
|
Also at what time of the night does it improve, if around 1am, then its congestion based and probably nothing to do with the connection.
You should only get the £200 charge if the engineer finds a fault that they fix on your side of things, e.g. a phone plugged in without a microfilter.
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
Register (or login) on our website and you will not see this ad.
|
|
|
Can you post your full router stats? and possibly the full result of a BT speedtest?
Matt
Sure Matt,
Router Info--------------------
Link Information
Uptime: 3 days, 2:58:39
Modulation: G.992.1 annex A
Bandwidth (Up/Down) [kbps/kbps]: 736 / 5,440
Data Transferred (Sent/Received) [MB/GB]: 140.20 / 1.75
Output Power (Up/Down) [dBm]: 12.5 / 19.5
Line Attenuation (Up/Down) [dB]: 27.0 / 50.0
SN Margin (Up/Down) [dB]: 11.0 / 5.5
Vendor ID (Local/Remote): TMMB / 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): 0
Error Seconds (Local/Remote): 2 / 0
FEC Errors (Up/Down): 0 / 1,153,387
CRC Errors (Up/Down): 0 / 1,598
HEC Errors (Up/Down): 0 / 1,449
------------------------------------
BT test 23:45 24th -------------------
Results Image not loaded
Test1 comprises of Best Effort Test: -provides background information.
Download Speed
2953 Kbps
0 Kbps 4000 Kbps
Max Achievable Speed
Download speed achieved during the test was - 2953 Kbps
For your connection, the acceptable range of speeds is 1200-4000 Kbps.
IP Profile for your line is - 4500 Kbps
The throughput of Best Efforts (BE) classes achieved during the test is - 11.86:24.24:63.9 (SBE:NBE:PBE)
These figures represent the ratio while sententiously passing Sub BE, Normal BE and Priority BE marked traffic.
The results of this test will vary depending on the way your ISP has decided to use these traffic classes.
If you wish to discuss these results please contact your ISP.
If you are experiencing problems with specific applications, servers or websites please contact your ISP for assistance.
Your test has completed please close this window to exit the performance tester.
Note: An Upstream test was not conducted on this line due to a technical issue. If your concern is related to upstream performance then please retry
the performance test again in 1 hours time.
--------------------------------
I'd be happy with this result if it was consistent.
Larry
|
|
|
Some errors on the downstream, which can indicate an issue.
Could you explain how the router is connected to the phone socket and if this is the main socket, an extension or the test socket?
Matt
|
|
|
Also at what time of the night does it improve, if around 1am, then its congestion based and probably nothing to do with the connection.
You should only get the £200 charge if the engineer finds a fault that they fix on your side of things, e.g. a phone plugged in without a microfilter.
I have actually just done a BT test for Matt, (results posted) and this is the first time I have checked it later than around 22:00.
It is pretty good for my location and much improved over earlier tests this evening, (see my first post).
Larry
|
|
|
Some errors on the downstream, which can indicate an issue.
Could you explain how the router is connected to the phone socket and if this is the main socket, an extension or the test socket?
Matt
It is connected directly into the BT socket with a filter which also has a BT phone attached.
The BT socket is maybe too old to have a test socket, circa 1980, but I may need to look.
I might also add that the Line Attenuation down often drops to 4.5db which may account for some of the errors but even with this poor LA I have often achieved around the 4000 on the 21CN setup whilst anything less than 6db would cause trouble on the old exchange setup.
Larry
Edited by deleted (Tue 25-Jan-11 00:00:17)
|
|
|
Do you have any extensions in the properly or is this the only socket?
Matt
|
|
|
Do you have any extensions in the properly or is this the only socket?
Matt
No, the router and the phone are the only connections and the PC is connected by ethernet from the router.
Larry
|
|
|
I'd still have a look for a test socket, in case the faceplate has become faulty.
Have you also tried a different equipment?
Sorry for asking the obvious.
Matt
|
|
|
Loss of Framing (Local/Remote): 0 / 0
Loss of Signal (Local/Remote): 0 / 0
Loss of Link (Remote): 0
Error Seconds (Local/Remote): 2 / 0
Uptime: 3 days, 2:58:39
I don't think there's anything wrong with your link to the exchange, TBH, and I don't see the need for an engineer visit given the above.
Phil
MaxDSL - goes as fast as it can and doesn't read the line checker first.
MaxDSL diagnostics
Are your kids pirates ? Limewire, Bearshare, Kazaa, BitTorrent, eMule are all tools of the trade.
|
|
|
I'd still have a look for a test socket, in case the faceplate has become faulty.
Have you also tried a different equipment?
Sorry for asking the obvious.
Matt
I have been known to overlook the obvious in the past
I haven't tried another router but I might have a BT Voyager 2100 knocking around. Would that handle 21CN?
Larry
|
|
|
Loss of Framing (Local/Remote): 0 / 0
Loss of Signal (Local/Remote): 0 / 0
Loss of Link (Remote): 0
Error Seconds (Local/Remote): 2 / 0
Uptime: 3 days, 2:58:39
I don't think there's anything wrong with your link to the exchange, TBH, and I don't see the need for an engineer visit given the above.
Yes, I'm tending to think it is congestion at the exchange caused by all the little darlings coming home from school after 3:30pm and turning on all the internet connected gadgets they were bought for Christmas.
This crashes the speed for me from around 4000Kbps to three figures every evening and at weekends.
This evening when the speed inevitably dips, I plan to turn off the wireless on the router, reboot it, connect it and it alone to the master socket with and without a filter, (not needed if only the router is attached, I believe?) and connect a netbook directly to the router by ethernet cable.
If the symptoms remain I may have a go with a BT Voyager 2100 if that can handle 21CN.
If that shows the same results and speeds pickup again around midnight I will put it down to exchange bottleneck and await BTs upgrade to fibre which has been promised for our town .......
Is that a plan?
Larry
|
|
|
As others are saying your stats don't look too bad
I don't think you've said who your ISP is yet? Sorry if I've missed it.
Lots of ISPs have capacity issues - more likely than a congested exchange I would have thought. Let us know who you're with and it might shed some light on things.
EDIT: Just read your 'plan' to possibly re-boot this evening. Syncing after dark will always give you a lower connection speed...
Edited by b4dger (Tue 25-Jan-11 10:38:11)
|
|
|
Modulation: G.992.1 annex A
you're getting a plain old ADSL1 connection, 21CN or not, so any router will do.
If you're on a business product the contention may go up when you get to the end of the office priority period eg 8pm.
I wouldn't assume it's the exchange, it could be the ISP contention.
Phil
MaxDSL - goes as fast as it can and doesn't read the line checker first.
MaxDSL diagnostics
Are your kids pirates ? Limewire, Bearshare, Kazaa, BitTorrent, eMule are all tools of the trade.
|
|
|
As others are saying your stats don't look too bad 
I don't think you've said who your ISP is yet? Sorry if I've missed it.
Lots of ISPs have capacity issues - more likely than a congested exchange I would have thought. Let us know who you're with and it might shed some light on things.
EDIT: Just read your 'plan' to possibly re-boot this evening. Syncing after dark will always give you a lower connection speed...
No, I didn't name my ISP as I trust them to have done all the checks. They claim not to throttle and I tend to believe them.
They are ADSL24 and they resell Murphx.
The connection has been rock solid over the years I've been with them and although the Entanet days had their moments the throughput was generally good.
I was moved to Murphx a good few months ago and until just after Christmas the new connection has been faster than ever 24/7.
This morning a speedtest gave a down speed of 4.33Mb/s at 9:00am and at 12:00 midday it had dropped to 3.57Mb/s.
The connection never drops and the Up speed and pings stays constant.
I realise that the SN will increase after dark but if I connect at a reasonable speed and achieve a comparable download speed I will feel that I have proved something is amiss with my equipment and within my control.
If the down speed stays drastically below the sync speed then I feel the problem is somewhere beyond my BT phone box.
Larry
|
|
|
Modulation: G.992.1 annex A
you're getting a plain old ADSL1 connection, 21CN or not, so any router will do.
If you're on a business product the contention may go up when you get to the end of the office priority period eg 8pm.
I wouldn't assume it's the exchange, it could be the ISP contention.
Thanks for clearing that up. I thought there was some change to the Exchange equipment that allowed the ISP to control much of what BT used to do wrt sync speeds and profiles and that may have changed the modem requirements.
Larry
|
|
|
Claiming not to throttle is very different to telling you that they have checked all the links from the exchange to leaving their network (is it just leaving MurphX) and that none of them are getting close to maximum utilisation.
With some links you can see bandwidth reducing even before they hit the maximum in any one direction, i.e. routers can be CPU load limited sometimes, so bandwidth checks look good, but the extra delay in handling packets slows down some protocols. Very old BT Centrals (years ago) used to suffer from this, e.g. would support 120Mbps of traffic, with a nomimal 100Meg downstream capability, but if the upstream hit 40Mbps, then downstream limit would be 80Mbps.
In short its a complex issue, and the evidence you've posted suggests more to do with ISP. What you need to look at now is how other ADSL24 and Murphx customers are doing.
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
I think as others have covered, router should be fine.
Have you run the second/third BT speedtest? (different login details). This will show if the congestion is within the ISP network, or within the BT network (before it gets to the ISP), i.e congested VP).
Matt
|
|
|
I think as others have covered, router should be fine.
Have you run the second/third BT speedtest? (different login details). This will show if the congestion is within the ISP network, or within the BT network (before it gets to the ISP), i.e congested VP).
Matt
The BT test has been a little problematic in that only once has it completed the Up stream test. it usually finishes with -
"Note: An Upstream test was not conducted on this line due to a technical issue. If your concern is related to upstream performance then please retry
the performance test again in 1 hours time."
Only once in the last week has it suggested a further test and it asked for me to change the router login to a different user (BT something or other) and leave the password blank.
However, the router failed to login at all so I couldn't run the test.
Larry
|
|
|
Was the username...
startup_user@startup_domain?
Matt
|
|
|
Claiming not to throttle is very different to telling you that they have checked all the links from the exchange to leaving their network (is it just leaving MurphX) and that none of them are getting close to maximum utilisation.
In short its a complex issue, and the evidence you've posted suggests more to do with ISP. What you need to look at now is how other ADSL24 and Murphx customers are doing.
The ADSL24 support passed all the test results etc up to what I suspect are Murphx tech support and they cogitated (tested?) for a week and then came back with the news that they had found no problem on their network and asked that i call a BT engineer, which is where I came here.
The ADSL24 customer forum, whilst still in existence, is not advertised or supported by ADSL24 staff so hardly anyone know they are there. There can only be one or two posts a month but I did find this one from yesterday.
http://adsl24.co.uk/forum/showthread.php?t=8835&page=2
There doesn't appear to be a Murphx or ADSL24 dedicated forum on this site and there is nothing I could find in the resellers forum, so I honestly have no way of knowing if this is widespread other than if an ADSL24 Murphx customer happens by this thread.
Larry
|
|
|
Was the username...
startup_user@startup_domain?
Matt
I really can't remember so it might have been.
Larry
|
|
|
If you can, try that with no password.
You should then be able to do a speedtest. If the results are in line with what you're getting normally, it indicates congestion in the BT network, anything else and it is likely the ISP.
Matt
|
|
|
If you can, try that with no password.
You should then be able to do a speedtest. If the results are in line with what you're getting normally, it indicates congestion in the BT network, anything else and it is likely the ISP.
Matt
Thanks, I'll do that in an hour or two when I expect the next dip in performance. I ran a BT test at 13:49 and achieved 2980Kbps which is pretty good for this address.
I appreciate the help you are giving and others on this forum.
Larry
|
|
|
There doesn't appear to be a Murphx or ADSL24 dedicated forum on this site and there is nothing I could find in the resellers forum...
Were you looking here? http://forums.thinkbroadband.com/otherisp.html
That's where most of the Murphx/ADSL24 user 'hang out'
|
|
|
There doesn't appear to be a Murphx or ADSL24 dedicated forum on this site and there is nothing I could find in the resellers forum...
Were you looking here? http://forums.thinkbroadband.com/otherisp.html
That's where most of the Murphx/ADSL24 user 'hang out' 
Yes, thanks, I had a look but no-one seems to be reporting a problem such as mine.
Larry
|
|
|
bt_test@startup_domain ?
Phil
MaxDSL - goes as fast as it can and doesn't read the line checker first.
MaxDSL diagnostics
Are your kids pirates ? Limewire, Bearshare, Kazaa, BitTorrent, eMule are all tools of the trade.
|
|
|
Asking you to call for a BT engineer is NOT your job, unless it is the voice part of your phone that is playing up and you pay voice line rental to BT.
For an engineer to arrive who will have ADSL testing kit, THE ISP must book this.
Again what some ISP's consider a problem and what you consider a problem may differ, e.g. if ADSL24 pays for 20Mbps (random number) of peak capacity, and it is hiting this sometimes, then Murphx will see no fault. ADSL24 may see no fault as its finances mean that that is the amount of capacity it can afford for the number of customers it has.
Andrew
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
If you can, try that with no password.
You should then be able to do a speedtest. If the results are in line with what you're getting normally, it indicates congestion in the BT network, anything else and it is likely the ISP.
Matt
I just tried that as well as bt_test@startup_domain.
I am supposed to enter these into the router setup as the user_name just as written above and with a blank password entry?
with both I get "Connection could not be established: Authentication failed" upon trying to connect.
Larry
|
|
|
Asking you to call for a BT engineer is NOT your job, unless it is the voice part of your phone that is playing up and you pay voice line rental to BT.
For an engineer to arrive who will have ADSL testing kit, THE ISP must book this.
Again what some ISP's consider a problem and what you consider a problem may differ, e.g. if ADSL24 pays for 20Mbps (random number) of peak capacity, and it is hiting this sometimes, then Murphx will see no fault. ADSL24 may see no fault as its finances mean that that is the amount of capacity it can afford for the number of customers it has.
Andrew
To be fair to ADSL24, they were going to book the engineer but required my permission as the risk that the problem lay within my premises would leave me open to a bill of £200+.
I have rebooted the router at 16:45 today and have synced at 736/5568 which is remarkably good for my distance from the exchange, (1.5 miles).
I have also been taking readings today.
9:37 - 4.33Mb/s - 0.56Mb/s
12:02 - 3.57Mb/s - 0.58Mb/s
13:49 - 2.98Mb/s- BT not tested
15:13 - 2.16Mb/s - 0.60Mb/s
Kids released from school @ 15:15
16:31 - 0.93Mb/s- 0.60Mb/s
16:50 - 0.44Mb/s - 0.61Mb/s
Now, it is meal time for the youngsters so
18:00 0.68Mb/s 0.57Mb/s
I suspect that when they get back to their dens I will get speeds around 0.30Mb/s for the rest of the evening.
Got to be some sort of log jam somewhere but I seem unable to get the BT test logon to work on my router.
Larry
Edited by deleted (Tue 25-Jan-11 18:12:37)
|
|
|
Someone has to post first 
How do you know others aren't in your position?
If I was you I would try a post asking if others are seeing 'throughput problems with Murphx' just in case
|
|
|
I have rebooted the router at 16:45 today and have synced at 736/5568 which is remarkably good for my distance from the exchange, (1.5 miles). FYI My attenuation (line length/quality) is also 50db (sometimes 51) and I'm sync'd at 5728kbps so nothing strange about your sync. NB. I'm on plain old ADSL as well...
|
|
|
Someone has to post first 
How do you know others aren't in your position?
If I was you I would try a post asking if others are seeing 'throughput problems with Murphx' just in case 
A very good point.
I am off to the 'otherisp' now.
Larry
|
|
|
I have rebooted the router at 16:45 today and have synced at 736/5568 which is remarkably good for my distance from the exchange, (1.5 miles). FYI My attenuation (line length/quality) is also 50db (sometimes 51) and I'm sync'd at 5728kbps so nothing strange about your sync. NB. I'm on plain old ADSL as well...
I have occasionally synced at 5000+ before being moved to Murphx 21CN system but it always dropped connection during the night and resynced at around the 3500 mark.
It is only in the last few weeks that it has held sync at around 5000+ and allowed a profile of 4000 to give me unheard of, (for me), speeds.
That is why I am hoping that I don't have to leave for another ISP.
It is not a case of cost, I am willing to pay for a reliable and speedy service, albeit with a month to month contract.
Larry
Edited by deleted (Tue 25-Jan-11 19:20:50)
|
|
|
|
I have just managed to run a BT speedtest all the way through including the third test that got me to log on the router at speedtest@speedtest_domain.
The result was broadly the same speed as the first two tests produced, namely, 239Kbps down and 574Kbps up on a profile of 4500Kbps down and 736Kbps up.
It finished with the statement -
" Your service was found to be performing poorly.
Please raise a fault with your service provider, stating that the BT Performance Tester tool indicates poor service throughput performance after the third test. "
Can any one explain the difference between the tests 1 and 2 and the 3rd and what, if anything, getting the same sort of poor throughput on both might indicate?
I am on a 21CN service.
Thanks all.
Larry
|
|
|
the slow 3rd test means the problem is not specific to your ISP.
traffic for that test is routed differently, so it means the link to the exchange is overloaded (or possibly elsewhere in the BT system, but not between BT and your ISP).
Phil
MaxDSL - goes as fast as it can and doesn't read the line checker first.
MaxDSL diagnostics
Are your kids pirates ? Limewire, Bearshare, Kazaa, BitTorrent, eMule are all tools of the trade.
|
|
|
the slow 3rd test means the problem is not specific to your ISP.
traffic for that test is routed differently, so it means the link to the exchange is overloaded (or possibly elsewhere in the BT system, but not between BT and your ISP).
Ah, thanks.
In a way I am relieved as I like it where I am as regards ISP.
I will just have to wait for BT to increase its capacity I suppose.
Larry
|