SDSL/2B1Q uses the startup or activation sequence from HDSL. It begins with the HTU-C transmitting a 2-level signal (CS0), then the HTU-R responds with its own 2-level signal (RS0). Upon detecting RS0 the HTU-C switches to 4-level transmission (CS1), and upon detecting CS1 the HTU-R goes 4-level too (RS1). Following this protocol allows each terminal to train its receive section to extract the far end signal from the noise and the echo of the near end signal.

Many SDSL/2B1Q flavors however have extended this protocol by implementing a pre-activation phase that takes place before the standard activation sequence just described is started. At the very minimum, both terminals have to agree on the symbol rate (i.e., SDSL speed) in use before the standard HDSL-style activation sequence is started if this sequence is to have any chance of succeeding; one of the goals of pre-activation protocols is to negotiate this rate or at least to communicate it unidirectionally from the HTU-C to the HTU-R. And of course once unleashed, the creativity of engineers unrestrained by standards has no bounds.

We are currently aware of the following SDSL/2B1Q pre-activation protocols:

• Conexant/Mindspeed AutoBaud
• CM pulse train

We've also invented our own pre-activation protocol defined in the IFCTF SDSL specification. All listed pre-activation protocols are implemented in our SDCORE.