Good point, Phil. I believe that the modern T1 spec states that a maximum or 45 zeros can be sent before the CSU will inject a '1'. I don't have a spec handy to verify it. Of course, with B8ZS, you'll never approach this long string of zeros on the line, so it's a moot point. B8ZS replaces eight zeros with 4/4 ones/zeros, in a known bipolar violation pattern. This will insure that ones' density is met on the line, and the circuit endpoints can replace the injected ones with eight zeros and restore the bit patterns of the modified DS0s. It's also the source of grief when a telco provisions one segment of a T1 circuit without B8ZS. When this non-B8ZS segment sees too many zeros, ones will be injected in the stream, corrupting the data. Of course, this NEVER happens anymore...right? ;-) -rb

From: Phillip Vandry <vandry@TZoNE.ORG> To: "Ron Buchalski" <rbuchals@hotmail.com> CC: nanog@merit.edu Subject: Re: Clear Channel on a T1 Date: Thu, 22 Mar 2001 18:04:46 -0500 (EST)

Please correct me if I'm wrong (because I'm not _that_ certain of this) but I think you still need the 1's density to maintain timing even in the absence of repeaters. If you have a whole bunch of zero bits in a row the voltage on the line will remain 0 for the length of the series. You get no indication of when one bit ends and another starts. If the string of zeroes is very long, the receiver might miscount the number of 0 bits that the sequence represents (more than intended if its clock is running faster than the sender's, less than intended otherwise). You don't have this problem with a long string of 1 bits because the voltage transitioning from + to - or back at every bit boundary.

_________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com