[time-nuts] HP5061B Cold Start

Skip Withrow skip.withrow at gmail.com
Sun Apr 9 11:39:15 EDT 2017

Hello Donald & Nets,

I have very much enjoyed the reports on your 5061 trials and tribulations.
Many other time-nuts that have been fortunate enough to have a 5061A or
5061B have gone through many of the same exercises in bringing old units to

I agree with your assessment that running the unit for as short a time as
possible will extend the tube life as much as possible.  I'm not quite sure
that I agree that after one hour you have the full stability of the unit as
it takes many hours (8-24) for it to reach thermal equilibrium.

This doesn't mean that I run my cesium 24/7 either.  I wish I had an
infinite supply of tubes to do so.  So, when I feel the use of the cesium
is appropriate, I fire it up, wait a day or so if possible, then test
away.  Usually it runs a week or two, then I shut it down.

The real purpose of the post though is to point out that using your cesium
for only an hour a day is pointless.  There are LOTS of alternative sources
that should be used, that have better ADEV, at taus less than 5000 seconds.

If you are making measurements less than 10-100 seconds in length use a
good undisciplined crystal oscillator.  In the 100 to 1000 second range,
consider a good rubidium oscillator (they are dirt cheap compared to a
cesium).  It's really only those long tau measurements that your cesium
brings something to the party, and this is what makes turning on your
cesium so painful.  If you are using it correctly, you are contributing to
its eventual death.

Another factor is the GPSDO.  A good one is the best of both worlds, short
term crystal stability, with long term GPS (steered cesium/rubidium)
stability.  I find that my cesium is most often used when I need to make
oscillator measurements without the several nanosecond hits that the GPSDO
takes from constellation switching, multipath, and other issues.  For the
most part, a GPSDO is my go to reference (but doesn't mean that I don't
want my cesium handy).

Keep up the good work.  You have definitely caught the disease

Skip Withrow


We performed an experiment with our working HP5061B which had been
cold overnight.  This was to better determine lock time from cold
ovens.  We found that the crystal oscillator oven heats up about twice
as fast as does the cesium oven.  Therefore the lock time is not much
different whether the crystal oscillator was cold or hot.  Our crystal
oven has been running 34 when hot.  The cesium oven has been running
about 14 when hot.  Here is our data as a function of measurement time
in minutes versus meter readings.  "+" denotes a pegged meter.

The instrument was powered on at 0 minutes except for the cesium
heater.  The cesium heater was turned on at 13 minutes. The crystal
oven had nearly stabilized by that time.  Oscillator heater current
was stable 15 minutes after power up and undershoots a bit at about 23

Note that first signs of beam current occur 7 minutes after cesium
oven was powered up.   First signs of 2nd harmonic occurred 9 minutes
after cesium oven was on.  We conclude that lock is feasible 23
minutes after a cold start.  If the crystal oven is left on, lock can
likely be achieved 20 minutes after the cesium oven is powered up.
The ion pump preserves beam tube vacuum if the crystal oven is left
on.  If it is left off for long periods, high ion pump current may
delay the application of power to the cesium oven until the vacuum is
pumped down.

For those who only occasionally need the stability of a cesium clock,
you can have full performance 23 minutes after a cold start, or 20
minutes if the crystal and the ion pump are hot.  If you use cesium
for an hour a day, tube lifetime will be extended by 24 to 1 over
continuous operation.  If a beam tube lasted five years in continuous
operation, it will last 120 years if used an hour per day.  A similar
benefit will occur if you use it continually for 2 weeks once a year.
It does no good to waste cesium unless you are using the instrument.

Time Ces Osc Beam 2nd

 0   +  0  0  0
 7   +  0  0  0
11 55  0  0  0
13 40  + 0  0
15 42 34 0 0
20 42 34 2 0
22 42 32 6 2
23 25 32 10 5 no lock
24 19 32 12 9 lock (free run back to OPR)
28 19 34 20 34 normal beam current

More information about the time-nuts mailing list