Question about noise margins

I'm just curious about why there are usually two different figures for noise margin and line attenuation for download and upload? Shouldn't the figures be the same up and down (as it's still coming through a single phone line).

The 2 channels, down and up, are separate.

Edited by deleted (Tue 26-Jan-16 09:56:12)

The main causes are they use different frequency ranges and the signal strength is supplied from the sending end.

On ADSLx the upstream frequencies are the low ones, and a narrow range. These attenuate (lose signal strength) gradually. The downstream frequencies are immediately above the upstream, and higher frequencies attenuate faster. Which is why the speeds fall off at greater distances as the highest frequencies become unusable.

On FTTC the principle is the same though the frequencies come in a number of groups alternating so it works a bit differently. However, VDSL2 signals attenuate much faster than ADSLx ones - I don't recall the ranges but of course the larger number of high ones cause a faster drop off with distance.

Which is why over a few hundred metres FTTC is so fast, but at long distances is no better and often worse than ADSLx.

Very broad brush there with a lot of detail missing, but I hope gives the general idea.

But surely the physical wire that's carrying the data up is the same as the one carrying the data down. Shouldn't they get affected by noise and attenuation equally?

Attenuation is a factor of the medium the data is travelling in, in this case usually copper and the lower frequencies of the upstream side on ADSL are stopped less by the copper than the higher frequencies of the download side.


Same physics apply in the atmosphere, low frequency sounds travel further than high frequency sounds, hence why whales communicate at such low frequencies over massive distances.

It's just tones of sound being transmitted and received, just like you can have a 2-way phone conversation.

Attenuation is a property of the copper ... but the important thing to understand is that the attenuation values change depending what frequency you measure with.

If you measure attenuation using a 300Hz signal, you will find less attenuation than at 1MHz. Which will be less than 10MHz.

When you read "attenuation" from a modem GUI, you don't know what frequency it used to measure that attenuation.

Because the upstream frequencies are different from the downstream frequencies - true of ADSL, ADSL2, ADSL2+ and VDSL2 - they will be attenuated differently on the same physical wire.

Thank everyone. Makes more sense now

If you run DSLstats under <Telnet Data> select the <Attenuation Log> and you will see the line and signal attenuation in the various up and down bands. These are "averaged" across the tones in the band

If you follow through: U0 D1 U1 D2 U2 D3 you will see the attenuation increasing. U0 uses the very lowest frequency tones whilst D3 uses those at the very top