[time-nuts] CS reservoir depletion

Tom Van Baak tvb at LeapSecond.com
Thu Jan 13 19:30:11 UTC 2011

> Really interesting thread. About a year ago on this thread the same
> discussion occurred.
> Several thoughts along these lines and frankly I know little accept for what
> I read here and online.
> I would agree the CS never runs out. Accept for one minor point. We tend to


> get these things way beyond HPs end dates by many years. So is it really
> true?
> This threads a bit different then the last one. The last one described
> opening the internals and various effects from trying to refill the kettle.
> This would be the first insight that I have in the fact that the systems
> polluted perhaps.
> Regards


What I've heard is that the lifetime of the 5071A high-performance
tube is around 7 years (when the cesium runs out) and the lifetime
of the standard tube is up to 20 years (when the EM fails). There
are other failure modes too so YMMV. Not sure about 5061A and
other tubes. There are variations within the 5071A tube as well
(including the short-lived turbo tube), so take this only as a rough
estimate (but do see the chart in paper10.pdf below).

As you may know the high-perf tubes have higher performance
simply because they use a much higher flux of atoms (beam
clock noise is a simple function of flux) and so they run out of
"gas" proportionally faster. Like a candle burning twice as bright,
or for you Neil fans, it's better to burn out than to fade away.

But there's lots of data on this subject since most of the national
timing laboratories use 5071A in quantity and they collect all
the internal performance data that the 5071A makes available
over the RS232 port. Some papers on the subject include:

End-of-life Indicators for NIMA's High-performance
Cesium Frequency Standards


Maintenance of HP 5071A Primary Frequency Standards

An Automated Alarm Program for HP5071A Frequency


More information about the time-nuts mailing list