[time-nuts] Why a 10MHz sinewave output?
magnus at rubidium.dyndns.org
Thu Feb 9 01:00:34 UTC 2012
On 02/08/2012 06:15 PM, Bob Camp wrote:
> Some (but by no means all) gear actually looks at some of the data fields on
> the T1 before it will accept it as a reference. In most cases a bits clock
> does fine. Of course you do need a proper balanced line driver and all that
> stuff to get it running.
> Still not something that's readily available in my basement. At work - not a
> The simple / stupid way to do it is to use a framer chip. They are cheap
> these days and they have all the driver stuff built in. They will even pack
> the data fields with "hey, I'm a good clock - use me". Run a cheap PLL to
> generate the framer clock and you are up and running.
The T1/DS1 signal as well as the E1 signal has a way to indicate to
which standard level the delivered clock is traceable to. If you are in
luck, you get a PRS (ANSI top reference) or PRC (ETSI/ITU top reference)
indication, which would mean that you have a G.811 compatible clock
within 1E-11 in frequency. In their infinite wisdom the "1544 kHz" and
"2048 kHz" standards slightly out of tune with each other as they stem
from different advances in the respective PDH hierarchy development,
which later rippled into the SONET and SDH counter-parts.
Their SSM codes for Quality Level encoding has now rippled over to
Synchronous Ethernet, which is nothing but a very strange SDH interface
rate. The ITU-T G.781 standard is a place to get lost to understand
these messages, so is ANSI T1.101. Digging around the ITU-T G.810-813
and G.823-825 specs is recommended. There is also several good ETSI
specs and a good TR to read up on.
More information about the time-nuts