Best 15 quid

Just thought I share my experience as I'm well chuffed.

I was running an old DSL-504. Working ok and had been for years but the other day I put my phone number into a site and it said I should get 7M which I wasn't. So I risked 15 quid on a TP-Link TD-8840T to replace the DSL-504. Sync speed on DSL-504 was 6624 Kbps, IP Profile 5500 kbps. On 8840T sync is 7616 kpbs and after a week my IP profile is now 6500 kbps. On speedtest.net the best I used to get was 5.05M, today I got 6.20M. That's 23% faster for 15 quid .

I've also asked my ISP to turn off interleaving to see if I can get 8128 and thus a higher profile still. The exchange in not ADSL2 so until (or if) that happens that as best as I can do.

That's good. It's a bit of a lottery though, it could have produced a lower sync.

Perhaps your new router has the same chipset manufacturer as the exchange end, that is quite often believed to give better results.

I was expecting the same with a small chance of bit better. I don't think it is down to chip set. I reckon it's down to the age of the DSL-504 because:

1. Some components have deteriorated with age (such electrolytic caps)
and/or
2. Firmware/hardware has improved (so its now faster/smarter)

I'm inclined to think it is both. The DSL-504 never connected as fast even when new and looking back there has been a small deterioration over the last few years. Processing power has certainly improved over the years so faster/better hardware seems a reasonable theory and electrolytic caps do deteriorate with age.

When I was checking out the TP-Link adsl router I came across a post but someone who had replaced a DSL-504 too and he also got a speed increase.

I have seen reports on other forums of the TP-LINK giving intermittent connections.

In reply to a post by john2007:

Perhaps your new router has the same chipset manufacturer as the exchange end, that is quite often believed to give better results.

That's not what I see. My exchange end is TI but using a TI chipset router is no good for me, a Broadcom chipset works much better.

I think line length/quality is a big factor on what chipset will be best for your line

In reply to a post by b4dger:

In reply to a post by john2007:

Perhaps your new router has the same chipset manufacturer as the exchange end, that is quite often believed to give better results.

That's not what I see. My exchange end is TI but using a TI chipset router is no good for me, a Broadcom chipset works much better.

You mean the TI-AR7?

Mine seems rock solid. I'm running RouterStats-Lite to record its stats. It's the only one I could find that works with TP-Link. The stats are a bit minimal in the router interface. No graph like the DSL-504 had.

The numbers before and after are odd. DSL-504 interface said Atten = 18dB, SNR=3dB, Sync 6624. The 8840T has atten = 24, SNRM = 8, sync 7616. Graphing SNRM it ranges from 6-9 with the odd dip down to 5. The DSL-504 said I had less attenuation before but it connected slower which contradicts that.

In reply to a post by bearded_malc:

The DSL-504 said I had less attenuation before but it connected slower which contradicts that.

No, it makes sense- you're reversing cause and effect.

The modem connects at rate determined largely by the SNRM and then measures the attenuation- lower speeds give lower attenuation (ignoring differences due to different ways of measuring it).

Edited by billford (Sat 11-Jun-11 10:09:32)

I see what you're getting at and that fits with my experience but it contradicts:

http://www.kitz.co.uk/adsl/max_speed_calc.php

In that calculator the lower the atten figure the faster it goes.

In reply to a post by bearded_malc:

In that calculator the lower the atten figure the faster it goes.

True, but:

The figures used are based on attenuation measured at 300Khz for adsl1
.
.
The calculator is based on a default Target SNR of 6dB

In essence it's just saying that the longer your line is (ie the higher the attenuation) the slower your maximum speed will be and giving a guide, but it's not a simple calculation because the actual attenuation varies with the speed rather than being the value for the fixed 300kHz that Kitz uses.

If the target SNRM were increased, the speed would go down and so would the actual attenuation.

The attenuation as reported by the router isn't really a single figure, it's a sort of average over the frequency range being used and can only be measured after that frequency range has been determined ie when the sync has been determined by the SNR.

There are several inter-dependent variables involved, which makes it a pig of a non-linear calculation to get right!

Edited by billford (Sat 11-Jun-11 10:25:54)

Right, Thanks for the info. So really you can only rely on trial and error.

If you just want a rough idea it's perfectly reasonable to make some approximations and assumptions, as Kitz's (and others') calculators do, but if you want an accurate answer then yes, it's suck it and see.

And I haven't even mentioned external interference- if your line is overhead and you live near a radio transmitter for example, then the theory goes straight out the window

In reply to a post by billford:

lower speeds give lower attenuation (ignoring differences due to different ways of measuring it).

How can that be true? The attenuation is a function of the line length and the frequency it's measured at, regardless of speed(? I guess you mean bandwidth?). If the frequency is fixed as per Kitz, then the length of the line must be varying

In reply to a post by BatBoy:

(? I guess you mean bandwidth?)

No, I meant speed- reported sync speed.

If we're in pedant mode, I suppose I should have said "maximum data rate" as it's reported in Kbps (or Mbps). Bandwidth is reported in kHz (or MHz).

In reply to a post by billford:

In reply to a post by BatBoy:

(? I guess you mean bandwidth?)

No, I meant speed- reported sync speed.

If we're in pedant mode, I suppose I should have said "maximum data rate" as it's reported in Kbps (or Mbps). Bandwidth is reported in kHz (or MHz).

Bandwidth has several related meanings:

Bandwidth (signal processing) or analog bandwidth, frequency bandwidth or radio bandwidth: a measure of the width of a range of frequencies, measured in hertz

Bandwidth (computing) or digital bandwidth: a rate of data transfer, bit rate or throughput, measured in bits per second (bps)

I think we all mean Bandwidth (computing)

Note that in textbooks on wireless communications, modem data transmission, digital communications, electronics, etc., bandwidth refers to analog signal bandwidth measured in hertz�the original meaning of the term. Some computer networking authors prefer less ambiguous terms such as bit rate, channel capacity and throughput rather than bandwidth in bit/s, to avoid this confusion.

Edited by billford (Sat 11-Jun-11 13:34:48)

In reply to a post by billford:

Note that in textbooks on wireless communications, modem data transmission, digital communications, electronics, etc., bandwidth refers to analog signal bandwidth measured in hertz�the original meaning of the term. Some computer networking authors prefer less ambiguous terms such as bit rate, channel capacity and throughput rather than bandwidth in bit/s, to avoid this confusion.

Yeah, I understand it was the original meaning, however those days have gone.

Now, back to the point - how can the speed affect the attenuation?

In reply to a post by BatBoy:

Yeah, I understand it was the original meaning, however those days have gone.

But the ambiguity remains... so if you want to try pedantry, do it properly.

Now, back to the point - how can the speed affect the attenuation?

Higher data rates require the inclusion of higher frequencies (cf Fourier). In ordinary cable such as used in telephone lines the higher frequencies are attenuated more than the lower ones for a given cable length. Electromagnetic radiation, capacitive/inductive losses, skin effect... it's not rocket science, just basic physics.

So higher data rates result in a higher overall attenuation.

Edited by billford (Sat 11-Jun-11 13:47:33)

In reply to a post by billford:

In reply to a post by BatBoy:

Yeah, I understand it was the original meaning, however those days have gone.

But the ambiguity remains... so if you want to try pedantry, do it properly.

Now, back to the point - how can the speed affect the attenuation?

Higher data rates require the inclusion of higher frequencies (cf Fourier). In ordinary cable such as used in telephone lines the higher frequencies are attenuated more than the lower ones for a given cable length. Electromagnetic radiation, capacitive/inductive losses, skin effect... it's not rocket science, just basic physics.

So higher data rates result in a higher overall attenuation.

Oh I get it, you're off into a theoretical discussion. I was talking about the OP's experience.

In reply to a post by BatBoy:

Oh I get it

No you don't.

In reply to a post by billford:

In reply to a post by BatBoy:

Oh I get it

No you don't.

Yes I do. This is the post I was referring to, not some theoretical discussion involving Fourier

In reply to a post by billford:

lower speeds give lower attenuation

- which is totally wrong.

In reply to a post by BatBoy:

In reply to a post by billford:

lower speeds give lower attenuation

- which is totally wrong.

In reference to www.kitz.co.uk/adsl/linestats.htm#attenuation:

Because the router measures against the frequencies available, some users may notice a very slight increase in attenuation if say moving from a fixed rate 512 kb connection up to 2Mb.
Higher frequencies such as those used to transmit faster speeds are more likely to be attenuated (higher frequencies = higher attenuation).
ADSL 2+ has an increased frequency spectrum, therefore an increased attenuation of around 3-4dB is not unusual. As a very rough guide a speed increase of 4Mb is said to increase the attenuation by 1dB.

In reply to a post by billford:

ADSL 2+ has an increased frequency spectrum, therefore an increased attenuation of around 3-4dB is not unusual

This is because the frequency range is doubled therefore the attenuation is also doubled.

In reply to a post by BatBoy:

This is because the frequency range is doubled therefore the attenuation is also doubled.

Ah, I see... so increasing the frequencies in use leads to higher attenuation, but lowering the frequencies in use doesn't lead to lower attenuation.

I'm most grateful for your insight.

In reply to a post by billford:

In reply to a post by BatBoy:

This is because the frequency range is doubled therefore the attenuation is also doubled.

Ah, I see... so increasing the frequencies in use leads to higher attenuation, but lowering the frequencies in use doesn't lead to lower attenuation.

I'm most grateful for your insight.

The frequencies are fixed, changed only by switching from ADSL to ADSL2+. Therefore the attenuation is fixed for ADSL and it is fixed for ADSL2+.

In reply to a post by BatBoy:

The frequencies are fixed, changed only by switching from ADSL to ADSL2+. Therefore the attenuation is fixed for ADSL and it is fixed for ADSL2+.

The available frequencies are fixed, which ones are actually used depends on the negotiated sync speed.

In reply to a post by billford:

In reply to a post by BatBoy:

The frequencies are fixed, changed only by switching from ADSL to ADSL2+. Therefore the attenuation is fixed for ADSL and it is fixed for ADSL2+.

The available frequencies are fixed, which ones are actually used depends on the negotiated sync speed.

The available frequencies are used to calculate the attenuation.

In reply to a post by BatBoy:

The available frequencies are used to calculate the attenuation.

No, the actual frequencies are used to measure the attenuation.

Calculation of the attenuation of a phone line possibly several kilometers long, with unknown properties, over a range of frequencies is not a practical proposition.

In reply to a post by billford:

In reply to a post by BatBoy:

The available frequencies are used to calculate the attenuation.

No, the actual frequencies are used to measure the attenuation.

Calculation of the attenuation of a phone line possibly several kilometers long, with unknown properties, over a range of frequencies is not a practical proposition.

The reported attenuation is a result of a calculation.

In reply to a post by BatBoy:

The reported attenuation is a result of a calculation.

Of course it is.

The attenuation at a range of frequencies is measured, a (probably weighted) average is calculated and the result given.

If certain frequencies, or a range thereof, aren't used then they don't provide an attenuation figure and don't contribute to the reported average.

The important point is that the individual attenuations for each frequency are measured, not calculated.

I moved from a Draytek to a TP-LINK and have not looked back. Mine was a present so cost me £0. now where is ebay?

In reply to a post by billford:

In reply to a post by BatBoy:

The reported attenuation is a result of a calculation.

Of course it is.

The attenuation at a range of frequencies is measured, a (probably weighted) average is calculated and the result given.

If certain frequencies, or a range thereof, aren't used then they don't provide an attenuation figure and don't contribute to the reported average.

If that were true, then attenuation would change instead of being fixed. Attenuation only changes when the equipment is changed or the modulation changes.

In reply to a post by BatBoy:

If that were true, then attenuation would change instead of being fixed. Attenuation only changes when the equipment is changed or the modulation changes.

You're not so far off.
The first sentence above is correct - the attenuation DOES change.

The second sentence is wrong though. The fact that you believe it's true probably means that (a) you are lucky enough to have a relatively low attenuation figure, and (b) that your router rounds the reported attenuation to the nearest dB (like mine). Some report values rounded to the nearest tenth of a dB...

You said in an earlier post: "The attenuation is a function of the line length and the frequency it's measured at". This was absolutely correct. The bit that you are missing and which billford attempted (semi-patiently ) to explain is that the frequencies IN USE can change (subtly) with every resync. On my line (atten > 50dB, ouch) I can see varying speeds AND varying attenuations. It's always in the range 50-52dB though.
I think all of this stuff is explained on the kitz site (or links from it) but it's been a while since I read it.

Thanks for creating another username just to make that post. Unfortunately, it does nothing for your credibility

Hilarious!
I'm not Bill though. (You think a moderator would make a new username? )
And did you actually disagree with what I wrote?

In reply to a post by spudski:

Hilarious!
I'm not Bill though. (You think a moderator would make a new username? )
And did you actually disagree with what I wrote?

I don't think you read any of my posts. On the bright side, you now have a post-count of 2.

In reply to a post by BatBoy:

Thanks for creating another username just to make that post. Unfortunately, it does nothing for your credibility

Seb would take an extremely dim view of a moderator posing as an ordinary member just to demonstrate support for his own arguments.

I feel this is a matter you should report to [email protected] for further investigation.

Batboy's allegation just shows him for what he is...a)wrong and b) in need of medical attention.

DrT

In reply to a post by billford:

In reply to a post by BatBoy:

Thanks for creating another username just to make that post. Unfortunately, it does nothing for your credibility

Seb would take an extremely dim view of a moderator posing as an ordinary member just to demonstrate support for his own arguments.

I feel this is a matter you should report to [email protected] for further investigation.

Don't let the new poster trick you into thinking I meant you, Bill. As I see it we were having a sensible and fruitful discussion. But I think you'll agree, he's not a newbie, and as far as I'm concerned not worthy of entering into a discussion with, as it's clearly a troll post.

In reply to a post by BatBoy:

But I think you'll agree, he's not a newbie

On the evidence available to me- no, I wouldn't agree with that.

In reply to a post by billford:

In reply to a post by BatBoy:

But I think you'll agree, he's not a newbie

On the evidence available to me- no, I wouldn't agree with that.

Fairy nuff. Let's see if he gets past 2

Why ignore what he said though?

Edited by deleted (Sun 12-Jun-11 15:32:16)

Because it's a troll post by a newbie.

If you say so.

IMO obviously. You did ask...

This sub-thread is way off topic and performing no useful function any more.

Closed.

Edited by billford (Sun 12-Jun-11 16:08:27)

In reply to a post by billford:

In reply to a post by BatBoy:

Because it's a troll post by a newbie.

You were a newbie once...

and I will be again, no doubt.

I'm just wondering about your motivation for labelling him a such. nm.

I'm shutting shop on this out respect to the OP. He can try again if he still retains the will to live.

Most of the technical stuff goes over my head but but I certainly take exception to a new member being accused of using a false username without any basis other than it was a first post that I can see, and it dared to challenge the views of a fellow member, what cheek!

Batboy, if you can substantiate that accusation.... do so, so.... Put up or Shut Up!. As I know you cannot put up option two is the one to go for. I think it shows that when you cannot win an argument you try to divert it to a personal attack.....not a very noble trait I'm afraid.

Your assertion that the post in question, which was polite, and well argued and very much to the point was a troll post has an incredible irony that I think most members will see, what is that in your sig about attacking the post and not the poster?...... made my day anyway.