[time-nuts] The Big Picture

[Tom Van Baak]

> I've been doing a lot of reading recently on time-related topics. 
> There's a lot of good info out there on time theory, the physics of 
> atomic clocks, time transfer, etc, but little on practical precision 

And that's sort of why we started this list! Welcome.

> time measurement, a topic on which there are numerous experts on this 
> list. Therefore I'd like to pick the collective brains of the group 
> on the practical matter of setting up a time reference.
> 
> Let's say that I pick up a used 5071a. What would it take to set it 
> up as an accurate standard? What other equipment would I need? How do 

It very much depends on what you want to do with it. But
first let me point out that there are two different kinds of
standards: time, and frequency. The atomic clocks you
buy, new or surplus, are typically frequency standards.

These have both an accuracy and a stability spec. If you
are a repair shop that needs perhaps 1 part in 10 to the
9th accuracy then almost any rubidium will do. On the
other hand, if you are a cal lab that needs 1 part in 10 to
the 12th accuracy you should be looking for a cesium. A
GPSDO will also work as a substitute in many cases.

A time standard is harder; it's a frequency standard that
also counts seconds and so it not only has to be accurate
in frequency but you also have to set it to the correct time,
at least once. A GPSDO takes care of both of these to
high levels of precision. Depending on the level of accuracy
desired, you can set the time of an atomic clock using GPS,
WWVB, Loran-C, NTP, a portable clock, etc.

> you sync it to UTC? What if I wanted to start with a less expensive 
> rubidium reference? Are there pearls and pitfalls on interconnecting 
> equipment? How about monitoring techniques?

I don't know if there's a cookbook about all this. If you read
the time-nuts archives you'll see many examples of the kind
of stuff we amateurs do. Most monitoring is based on having
a second (or third) clock or a GPS receiver along with a good
time interval, frequency, or "universal" counter.

I would recommend getting a used rubidium first. They are
cheaper, more plentiful, and more likely to be operational
even if purchased surplus, than a cesium.

> If I've missed some good references out there on the nuts and bolts 
> of the hobby, please point me in the right direction.
> 
> Keith

See also the links here:

    http://www.leapsecond.com/time-nuts.htm

I started with one old nixie tube HP 5245L counter and a
quartz oscillator. Then I picked up a M-100 rubidium. Then
another counter (one with RS232 output). Then a WWVB
receiver. Then a FTS 4050 cesium, ...

If anyone else on the list has a story on how you started
down the time-nuts road, let Keith know.

Thanks,
/tvb