Talk:Robbed-bit signaling

From the article: "Although only one bit out of 48 is robbed, there is no way to know which frames will be robbed by the various T1 connections in a phone conversation, so the signal to noise ratio will be somewhere between 31 and 37 dB." I thought that bit-robbing stamped on the LSB of every 6th sample in every DS0 channel carried in the T1 (I think that's correct; I'm not a telephone engineer, can anyone check this?), so it should be possible to calculate the SNR exactly from this. -- The Anome 11:10, 12 October 2005 (UTC)

Comment on Above
The robbed bit is indeed every 6th sample. However, that 6th sample is not aligned with the other "6th sample" from other T-1s arriving at a central office. For switching the T-1s arriving at a central office, all the timeslots must be aligned, that is, all the incoming T-1s are buffered so that all the TS1 arrive together. However, this does not insure that all the 6th samples are aligned. As a result, in the outgoing digital stream, the new signaling bits will rob the 6th sample of whatever is defined as 6th outgoing, which is not necessary the same as the incoming 6th.

In the beginning of T-1 system design in 1958, the objective was to be good enough as an "exchange carrier", that is, voice transmission within a city. This required only 7 bits per channel. The 8th bit is dedicated to signaling. Intercity transmission was still by analog carriers. Later, when T-1 was adapted to intercity long haul use, 8 bits were required. Thus the robbed bit signaling method was invented, which achieved almost 8-bit performance. When compared to analog systems, where signal to noise would degrade with distance, T-1 with robbed bit signaling was acceptable. Its signal-to-noise degrades with distance, but slower compared to analog and tops at 6 dB. LoopTel 07:24, 30 November 2006 (UTC)