Theory behind 31.25 MHz in 100BASE-TX?
Hi there,
100BASE-TX is said by a number of authors to have its fundamental frequency at 31.25 MHz, but the proofs given (if any) seem vague or questionable at best. E.g., I think I met a proof going as follows: There are four elements in the cycle through which the signal level loops in MLT-3, so the main clock of 125 MHz is divided by 4.An amateur in data communications, the only pseudo-scientific speculation I could put together is as follows: MLT-3 acts as a four-state line code, perhaps due to its cycle length: {-1, 0, +1, 0}. Therefore by Hartley's law its effective symbol rate in 100BASE-TX is 125 Mbit/s / log2(4) = 62.5 Mbaud. And now by the sampling theorem we get to the fundamental frequency of 62.5 / 2 = 31.25 MHz. What I cannot prove here is MLT-3's being a four-state code.
Do you think my speculation makes any sense? And if so, do you have any ideas on how to ground this apparent alphabet size of MLT-3? Thanks!
read more and respond »Posted 5 years ago by Yar Tikhiy in Ethernet LAN
