> On Sun, 4 May 1997, Jon Lewis wrote:
>
> > On Sat, 3 May 1997, Al Hopper wrote:
> >
> > > On a point 2 point circuit (usually) the telco does not supply clock. In
> > > almost every other case the telco will supply the (sync) clock. Mr Jay is
> > > indeed correct. Also the circuit pairs should be flipped somewhere in the
> > > middle. Since you're seeing data the crossover must be OK.
> >
> > I keep getting conflicting advice on this...some say the telco provides
> > clock, some say they don't. Is it possible some CSU/DSU's deal with
> > incorrect settings better than others?...i.e. Say you have a PTP T1, with
> > CSU/DSU's at each end set to slave/network. Will some CSU/DSU's handle
> > this while others have a hard time?
>
> Most of them will handle it - for a while. We had this problem with a
> point-to-point MCI T-1 and Adtran TSUs. I made the assumption that the
> circuit was de-muxed from a larger sync pipe and clock supplied by telco.
> In fact, clock was not supplied by telco. Circuit came up, worked for
> hours or days, and then froze. Telco getting in and looping would bring
> it back, sometimes. Power-cycling one or both of the TSUs would usually
> also bring it back. Eventually got someone at MCI to tell us definitively
> that clocking was not supplied by them. Set one TSU to Internal, other to
> Network, and it's been purring fine for years.
>
One way to verify this is to measure the clock frequency using one of the
hi-end T-1 testers. What you will find is that the clock is off by 1% or
2% (YMMV). And (as per Jay above (our man "from the trenches")) the clock
frequency will drift over time. When you setup one of the CSU/DSUs to
supply the clock your T-1 tester will verify that the sync frequency is
very close to the required 1.5xxxMHz - depending on how accurate the
frequency source is on the CSU/DSU providing the clock. The actual
frequency may be a little off and not cause any problem, but it must not
drift!
If the telco provides the clock your tester will verify that the clock
frequency is *exactly* 1.5xxMHz, since the telco derives the clock from a
_very_ accurate frequency standard.
There is no way for a CSU/DSU to "auto sense" the clock situation and
automagically do the right thing. T-1 technology is too old for that!
So now you can rent a T-1 tester and prove to yourself that the telco is
not supplying clock on this point-to-point circuit!
Al Hopper Logical Approach Inc, Plano, TX. al@logical-approach.com
(972)-379-2133 or (972)-849-5765. Fax 972-379-2134