[time-nuts] Time syncing question

Bill Janssen billj at ieee.org
Thu Aug 31 00:51:37 UTC 2006


Hal Murray wrote:

>[telcos]
>  
>
>>They need to keep all the trunk signals in   perfect sync to keep from
>>loosing bits.
>>    
>>
>
>That's not quite right.
>
>You won't drop bits if the trunk is (slightly) faster than the source.  I 
>think the protocols for encapsulating T1s in T3s and such have a slot for 
>that so telco-X can transport bits for telco-Y.  Two telcos will, of course, 
>have independent clock empires.  (I haven't looked at the specs in many 
>years.)
>
>  
>
Yes T1 connected to T3 uses "stuffing bits" to get the T1 bit rate up to 
1/3 of the T3 rate. These bits
are then removed when the T3 is connected to three T1 lines. 
Intelligence is sent along with
the "stuffing" bits so they can be identified.

However T1 connected to a digital switch does not use that scheme. The 
digital switch has a buffer
that is used to buffer the digits until the correct time to send them on 
to an outgoing T1. If both T1's don't
run at the same speed the buffer will under run  and repeat digits or 
over run and drop digits.

Not too bad for a phone call but disastrous for a data connection.

Bill K7NOM

>That turns the problem into one of getting the receiver in sync with the 
>transmitter.
>
>How do cell phones derive their 8KHz clock?  Is that problem burried in the 
>compression?
>
>
>
>
>  
>






More information about the Time-nuts_lists.febo.com mailing list