GPS Clock to DOM Hub Timing Distribution 2005
At the south pole, during the first deployment year, the clock fan-out
configuration is more complex than the ones used in test laboratories,
or DOM Freezer Labs at Wisconsin, Zeuthen, or
Uppsala.
The (temporary) counting house contains a Model
ET6000 GPS clock with Oven Controlled crystal oscillator option, plus
option daughter card (OP40) plus two 10 MHz shaper option cards
(OP21A).
The rear panel BNC outputs of the GPS clock will ultimately drive nine
DOM Hubs. In order to fan 8 GPS clock outputs to 18 inputs, a passive fan-out is required. Unequal
cables must be used to achieve
matching of the phases of the 10 MHz signals applied to all DSB cards.
As with the DFL and Test Lab installations, the delay between 1 Hz
_↑‾ to 10 MHz
_↑‾ timing, are measured at the monitor test point on the
end panel of
the DSB card.
Signal Sources available from the
BNC outputs on the back of the GPS clock are not carefully/precisely
timed, nor is the relative timing information documented in the
manual. The channel output to output skew can be quite
surprisingly large. Furthermore, the jitter from a divider output
with respect to the oscillator output is not well documented, or
explained.
At the pole, in Jan, '05, one instance was observed with the 10 MHz
signal rising 8ns before the 1
Hz signal. This timing has an 11ns safety margin with respect to
the forbidden timingsin the clock
skew page. (-8ns timing is the same as
+92ns, which is within a safe operating region)
The Presentation
Documention may be found in this powerpoint presentation.
You may download a powerpoint player (free) from the Microsoft web
site, or you may open powerpoint presentations with the tool available
in OpenOffice (www.openoffice.org).
Presentation
10/4/04 gtp LBNL