BT Internet - Connection Speed
Thread Starter
Join Date: Feb 2003
Location: Not Telling!
Posts: 162
Likes: 0
Received 0 Likes
on
0 Posts
BT Internet - Connection Speed
I've been experiencing poor internet rates lately and so I thought I'd have a play around with a few things. I'm currently with BT on an unlimited 8meg connection and ran the tests around 1800 local time (UK).
The starting point was with my PC connected to an extension socket upstairs and this is what my modem reported:
Data rate: Down - 3776, Up - 448
Noise margin: Down - 15.3, Up - 16.8
Output power: Down - 19.7, Up - 12.3
Attenuation: Down - 28.5, Up - 30.0
Running the BTspeedtest produced:
IP Profile - 2500
Up stream - 448
Downstream - 3776
Actual IP throughput - 2326
I then changed the filter & cable from the wall socket to the modem for brand new ones and ran the same tests. This time I got:
Data rate: Down - 6560, Up - 448
Noise margin: Down - 14.6, Up - 21.0
Output power: Down - 17.8, Up - 12.3
Attenuation: Down - 16.5, Up - 7.0
Running the BTspeedtest produced:
IP Profile - 2500
Up stream - 448
Downstream - 6560
Actual IP throughput - 2332
For the final test I connected to the BT master socket (inside the house's prime socket) with the new cable/filter and received these values:
Data rate: Down - 6624, Up - 448
Noise margin: Down - 14.6, Up - 24.0
Output power: Down - 15.8, Up - 12.4
Attenuation: Down - 15.0, Up - 6.5
Running the BTspeedtest produced:
IP Profile - 2500
Up stream - 448
Downstream - 6624
Actual IP throughput - 2367
Now out of all this it would appear that changing the cable/filter for new items produced a significant increase in the downstream connection rate (3776 upto 6560). There would appear to be a small loss by using an extension socket rather than the master BT socket.
The BT speedtests seem to confirm the modem reported downstream speeds in each case but, according to BT, the actual IP throughoughput never really changes. Does anyone know why? I assume it's to do with the IP Profile figure which, to my simple brain, seems to imply the line is capped at 2500.
If the line is capped, is this because I've been acheiving poor data rates around the 3500 mark? Now that I can connect at 6000+ will I see an automatic increase in the IP Profile or do I need to contact BT() to request a line recalibration?
I hope all the above makes sense
Finally, can anyone explain in layman's terms what noise margin, output power & attenuation are, as reported by my Belkin router, and whether high or low are better values.
=====================================================
Edited to add:
Now 1930 UK time and having used the BT speedtester again its reporting:
IP Profile - 500
Up stream - 448
Downstream - 6240
Actual IP throughput - 318
Is it possible that my connection has been throttled back from its previous level of 2500?
Last edited by Bo Nalls; 24th Oct 2007 at 18:33.
More bang for your buck
Join Date: Nov 2005
Location: land of the clanger
Age: 82
Posts: 3,512
Likes: 0
Received 0 Likes
on
0 Posts
Have a look here, some info that might be useful
http://www.kitz.co.uk/adsl/maxdsl2.htm
http://www.kitz.co.uk/adsl/maxdsl2.htm
Recidivist
Join Date: Jun 2005
Location: Essex, UK
Posts: 1,239
Likes: 0
Received 0 Likes
on
0 Posts
Showing my ignorance (again).
Having a DAC on the line used to be a problem with dialup. Does that apply to broadband, and if so, could there be one on your line?
Having a DAC on the line used to be a problem with dialup. Does that apply to broadband, and if so, could there be one on your line?
More bang for your buck
Join Date: Nov 2005
Location: land of the clanger
Age: 82
Posts: 3,512
Likes: 0
Received 0 Likes
on
0 Posts
GG, just to go off topic for a moment, did you benefit from debugging your system?
Official PPRuNe Chaplain
Join Date: Apr 2001
Location: Witnesham, Suffolk
Age: 80
Posts: 3,498
Likes: 0
Received 0 Likes
on
0 Posts
My daughter had a DACS unit on her phone, and they had to find another twisted pair before she could go onto ADSL. Apparently DACS = ADSL is impossible.
Hippopotomonstrosesquipidelian title
Join Date: Oct 2006
Location: is everything
Posts: 1,826
Likes: 0
Received 0 Likes
on
0 Posts
Now 1930 UK time
For an example of artificial load management, I've got 100Mbps fiber to my apartment on a "home" plan. Everyone else has 100Mbps fiber to their apartments on an "apartment" plan. I pay about $10/month more than the others do. Speed tests show we all have about 65Mbps up and down to the exchange. The difference is, I have 65Mbps to the box in the basement, and 65Mbps to the exchange on a 1Gbps connection. Everyone else has 65Mbps to the box, then share 65Mbps to the exchange. That's done by load management software. In reality, at peak times, my connection seems to be about a bazillion times faster than everyone else.
(My other feed is 1 Gbps fiber to the building on a competing carrier, with 67Mbps VDSL to me. It feels much, much faster than my fiber but is asymmetric.)
Since you seem to be a competently inquisitive person, see what happens when you disconnect every other device from the line. The parent's place couldn't Slingbox TV to save its life, until I found the cheapo handset that was killing the performance, even through its filter.
Last edited by Bushfiva; 25th Oct 2007 at 12:14. Reason: Put a month's worth of typos in one posting. Shall save some for later.
Spoon PPRuNerist & Mad Inistrator
If you are using an internet "speed test" site, make sure you find and try half a dozen or so.
There seems to be massive variation in the results, possibly due to differences in methods, algorithms and inferences - not to speak of congestion at THEIR end if they are timing the upload of files to your PC! Add to this the possibility of caching along the route...
For example, 3 sites report my (cable) connection speed at about 800 kbps, whereas another 3 report it as around 3800kbps.
Yet another couple of sites put it at about 1200 Kbps.
It is nominally 4Mbps, btw
Given the grouping around these numbers I suspect the methods used are different - but I have no real idea of how the testing is done (apart from timing the download of files).
SD
There seems to be massive variation in the results, possibly due to differences in methods, algorithms and inferences - not to speak of congestion at THEIR end if they are timing the upload of files to your PC! Add to this the possibility of caching along the route...
For example, 3 sites report my (cable) connection speed at about 800 kbps, whereas another 3 report it as around 3800kbps.
Yet another couple of sites put it at about 1200 Kbps.
It is nominally 4Mbps, btw
Given the grouping around these numbers I suspect the methods used are different - but I have no real idea of how the testing is done (apart from timing the download of files).
SD
Beware of anything BT say. I know that is a given but it doesn't seem to stop them trying it on. I recently received a call from BT informing me that I had been 'specially selected to receive their 20mb/s service. We had several minutes conversation basically on how they were going to achieve that rate and what discounts they offered if they failed to achieve 20m. I pointed out that I was miles from the exchange and over copper wires 20m wasn't just optimistic it was pure fantasy! I also pointed out that my local cable company had a fibre cable not 10 yard from my front door. Of course BT are laying fibre everywhere I was told, not round here dear. Finally got them to admit that when they tested the line the max throughput was 6mb/s. Bunch of crooks, get them to test your line before committing to Broadband service from them or the others that resell their services, particularly Sky.
pm575
pm575