How VDSL2 and ADSL2+ signalling works in layman terms.
Posted: Fri Feb 10, 2017 5:30 am
Discrete Multitone (DMT) is an encoding method used within the broadband system to transfer bits of digital signals (bits per second, basically) along with a media in serially fashion around a twisted pair. The tones are modulated by a digital system onto a carrier frequency in tones or subcarriers, this method of encoding and modulating the digital signals is DMT.
The whole bandwidth for ADSL2+ is 2.07Mhz, that is a sine wave with 2 million 7000 thousand cycles going up and down in a second long, hence the data that can be pushed down that path is high depending on the distance from the exchange. The waves, that's cycles in a second are discrete along that complete frequency. So the first tone would be at a frequency of between 4.3135Khz to 8.827Khz, the next multitone would start above 8.627Khz to 12.9405Khz, etc...
The reason DMT is used is so those cycles within a subcarrier, which is a maximum of 4,3135 cycles a tone can be varied depending on noise at the tone frequency it is in along that 2 million 7000 thousand cycles. The number of cycles in discrete tone cycles is reduced when the noise occurs, so instead of say 4 thousand cycles in a discrete tone, it could be reduced to say 2000 cycles at the higher frequencies to enable those frequencies to be used even when the signal degrades as the frequency rises.
Much like a wave in the sea, the larger the wave the longer it travels even when it's discreet; so waves (cycles) at lower frequencies will travel further along the cable than shorter waves, higher frequencies experience more loss as the amount of ripples increases, as a very basic example that's easy to understand.
There are two bands in ADSL2, the upstream (US) and downstream (DS). Each set of multitone is 4.3135 kHz wide and the whole system consists of 256 * 4.3125Khz multi tones that are discrete for ADSL and 512 of the same bandwidth for discrete multitone on ADSL2+.
In ADSL2+ US band is 32 subcarriers or tones wide (138Khz width in total for the US), whereas the DS is 480 subcarriers or tones wide (2.07Mhz in frequency width).
The multitone have a variable number of bits on each set of discrete tones, and is in every decreasing bits as the frequency moves up the spectrum to 2.208Mhz plus 3.125Khz for voice with guard bands between voice, US and DS so they don't intrude on each other whilst being transmitted along the media in both upstream and downstream. Upstream is from the premises to the exchange and downstream from the exchange to the premises.
Some bits are used for error correction within the multitone, and the error correction bits are interleaved with the encoding of these bits onto DMT.
The actual Mbps that can be transferred along the DS is a threshold of 24Mbps in theory, but noise and error correction reduces the synchronisation speed to lower level as noise and loss increase. The threshold for upstream is 1.4Mbps, but the actual maximum usually data bits within the stream is usually lower.
The whole bandwidth for ADSL2+ is 2.07Mhz, that is a sine wave with 2 million 7000 thousand cycles going up and down in a second long, hence the data that can be pushed down that path is high depending on the distance from the exchange. The waves, that's cycles in a second are discrete along that complete frequency. So the first tone would be at a frequency of between 4.3135Khz to 8.827Khz, the next multitone would start above 8.627Khz to 12.9405Khz, etc...
The reason DMT is used is so those cycles within a subcarrier, which is a maximum of 4,3135 cycles a tone can be varied depending on noise at the tone frequency it is in along that 2 million 7000 thousand cycles. The number of cycles in discrete tone cycles is reduced when the noise occurs, so instead of say 4 thousand cycles in a discrete tone, it could be reduced to say 2000 cycles at the higher frequencies to enable those frequencies to be used even when the signal degrades as the frequency rises.
Much like a wave in the sea, the larger the wave the longer it travels even when it's discreet; so waves (cycles) at lower frequencies will travel further along the cable than shorter waves, higher frequencies experience more loss as the amount of ripples increases, as a very basic example that's easy to understand.
There are two bands in ADSL2, the upstream (US) and downstream (DS). Each set of multitone is 4.3135 kHz wide and the whole system consists of 256 * 4.3125Khz multi tones that are discrete for ADSL and 512 of the same bandwidth for discrete multitone on ADSL2+.
In ADSL2+ US band is 32 subcarriers or tones wide (138Khz width in total for the US), whereas the DS is 480 subcarriers or tones wide (2.07Mhz in frequency width).
The multitone have a variable number of bits on each set of discrete tones, and is in every decreasing bits as the frequency moves up the spectrum to 2.208Mhz plus 3.125Khz for voice with guard bands between voice, US and DS so they don't intrude on each other whilst being transmitted along the media in both upstream and downstream. Upstream is from the premises to the exchange and downstream from the exchange to the premises.
Some bits are used for error correction within the multitone, and the error correction bits are interleaved with the encoding of these bits onto DMT.
The actual Mbps that can be transferred along the DS is a threshold of 24Mbps in theory, but noise and error correction reduces the synchronisation speed to lower level as noise and loss increase. The threshold for upstream is 1.4Mbps, but the actual maximum usually data bits within the stream is usually lower.