HP 5060A Cesium Beam Frequency Standard
© Brooke Clarke 2000 - 2005
|Fig 1 Front
|Fig 2 Rear
|Fig 3 Top
|Fig 4 Bottom
|Fig 6 Manual
|Fig 7 Power
Used Cesium Standards
Cesium frequency standards work by
squirting Cesium atoms from a reservoir that's heated to boil off some
atoms from a small piece of cesium liquid metal. These atoms are
processed is such a way that they become an atomic frequency
standard. The problem is that there is a finite life to the
Cesium source and when it's used up the whole physics package needs to
Another problem is that the physics package operates under a hard
vacuum. To maintain this vacuum there's an internal ion
pump. If the unit is left in storage without power the ions in
the physics package accumulate and when the unit is powered up some
time will be needed to pull them out. After some years of storage
it may not be possible to get the hard vacuum and the unit will be dead.
sells a universal
Cesium physics package for about $16,000
plus installation. Most of the used Cesium standards on the
market are there because they have either died or are about to die.
Prior to working Cesium beam frequency
standards the official meter was defined by an invar rod (i.e. a real
physical standard). Today a meter is defined based on the
distance that light travels in a specified time. The definition
of a second was also changed from being based on the rotation of the
Earth to the Cesium clock since the Earth speeds up and slows
down. Of all the things that can be measured time (or frequency,
time interval, etc.) is the one that can be measured with the most
The HP Cesium standards became the
world's common time standard. As of 2005 Agilent (formerly the
Test & Measurement division of HP) is selling off their Cesium
standard business to Symmetricom
1948 - Harold Lyons at NBS starts work
on an amonia based atomic clock
1955 - Louis Essen at the National Physical Lab in the UK comes up with
the idea for a Cesium clock and the first one is built. Note that
the NPL unit was a clock that kept Cesium based time for the three
years needed to do a correlation study with the USNO (and probably ran
longer than that). The U.S. standard was moved from the East
coast to Colorado and was more a frequency standard than a clock.
1964 - The HP 5060A, a cesium-beam standard instrument,
gains worldwide recognition as the "flying clock." International
standards for time
set by flying the HP 5060A around the world. Accuracy 2E-11, Precision
uses 5 MHz Xtal oscillator.
1967 - HP acquires the rights to make the Cesium Beam Tube from Varian,
1982 - FTS
microporcessor controlled Cesium standards
1987 - 5061B Cesium Beam Standard - Accuracy 4 or 6E-12, Settability
has 10 MHz output
5065A Rubidium Frequency Standard, 5 MHz output
105B Quartz Oscillator, 5 MHz
5089A 22 - 28 VDC Standby Power Supply 67# 7" high
1989, 1990, 1991, 1992 - same offering
1993 5071A Primary Frequency Standard introduced (first microprocessor
HP unit almost 10 years after FTS), Accuracy 1.5E-12
resolution 6.3E-15, RS-232 remote control. 10 MHz output ,
5065A Rb dropped
1995 - 58000A Time and Frequency Reference Distribution Module, 58503A
Time and Frequency Reference , 10811D/E Oscillators, E1750A 100 kHz to
MHz distribution amp in VXI format, E1752A 1PPS dist. amp in VXI
1997, 1998 - 55000 Ultra-Precise Time and Frequency Systems
2000 - 58502A 12 chan sine wave dist amp, 58503B GPS T&F Reference
58533A GPS T&F Reference box
Now that the 5060A has been running for a number of days
with the "Continuos operation" light always on, it just needed some
time to settle
When all is well you get the "continuos
C Field Adjustment
There is an adjustment for changing the magnetic field around the Cs
This has the effect of tweaking the output frequency.
The scale of the 10-turn knob is about 1E-9 per turn, or 1E-10
1/10 of a turn (big division), or 1E-11 for 1/100 of a turn (one tick
You can see that's it's not easy to get the Cs source set to within
This appears to be like the PRS10 which has
high Q but the actual frequency depends on a number of subtle factors.
It's questionable to me which is the better timekeeper, the old 5060A
spec) or the new PRS10 (1E-12 spec)?
The life of the PRS10 will be much longer than any Cs source but the
5071A will provide better stability.
The manual says that the C Field adjustment is to set the correct
hyperfine transition using a injected audio signal at the Zeeman
But the specifications clearly say that the fine frequency adjustment
1E-9 per revolution or 1E-10 per division (10 divisions per revolution).
26 July 2005 - Note even though the actual operating frequency can not
be set as accurately as newer standards, because of being a Cesium
standard the frequency will be constant (i.e. it will not drift).
By using a GPS timing receiver the actual frequency can be determined
and then backed out of measurements made. So for measurement
applications it's as good as a modern standard.
The two main cylinders inside on the top
are the Cs physics package and the 5 MHz crystal oscillator. This
oscillator is used in a number of other HP products.
It is based on the idea of having a
large thermal mass surrounded by a
oven that uses a proportional controller rather than a bang bang
The electronics are contained in metal boxes on the bottom side.
This is a used unit that has scratches and dings. The physics
package seems to be working, but the 1 Mc divider dies not start.
An adjustment may be needed or a repair. Since the 1 Mc divider
is not working then the 100 kc divider can't work.
Start Up Instructions
These instructions are intended for
starting a unit that has just been received or has not been used for
This is a quick check only. If the peaks are present while tuning
through 5 Mhz this is a good sign that the tube is operating, but you
would need to
run it at least 100 hours to make sure the levels stay stable. However
seeing them could be the result of other circuit failures.
- First take the top and bottom covers off and look to make sure
modules are there and cables connected inside
- Place the mode switch to osc. only.
- Connect AC power. (if you don't have a cord you can make a
cord with some matching pins)
- With circuit check 2 switch in supply verify a reading of about
If significantly hi or low remove ac power until supply is repaired.
If OK so far place the switch to ion pump I.
The meter might be pegged, if so leave unit on till it drops below 1.
may take up to 24 hours!)
- Then place the mode switch to loop open.
The ion pump I may rise and peg again.
This is OK it may cycle for up to 24 hours again.
The cesium oven reading should come up to above 4 when you switch to
It may also cycle Off and On until the ion pump I stops cycling.
the cycling stops the Cesium oven reading should drop to approx. 4 in
If it doesn't, turn power off until oven repaired.
- The in oven
oven readings should drop in about 4 and 2 hours.
If not turn power off until repaired.
- Now with mod switch off and
2 in beam I position connect a frequency counter to the 5 Mhz output.
- Slowly adjust the coarse oscillator adj. back and forth across 5
by about plus/minus 1 Hz while watching the beam I meter.
You should see three peaks on the meter with the one at 5 Mhz the
- Set the adj. for the central peak.
Put circuit check I to 2nd harmonic.
Turn mod to on.
The 2nd harmonic reading should rise.
- Now place mode to operate.
Wait about 30 seconds and then push the logic reset button.
The continuous operation lamp should light.
Note that once you have the 5060A locked, it can be unplugged and
later just plugged in and will just lock by it's self.
Anniversary of the First Accurate Cesium Atomic Clock - The NPL
Cesium standard was very stable
Lord…. Louis Essen D.Sc., F.R.S. - about the inventor of the Cesium
standard (Need IE not Netscape to see it.)
Atomic Frequency Standards: A Trip Through 20th Century Physics -
Time & Frequency -
scanned items of historical interest -
NPL - How an atomic clock works - practice
Essen 1908-1997 -
Fifty years of
atomic clocks -
Atomic beam frequency standard having RF chain with
frequency oscillator September 15, 1992 331/3;
Atomic clock system with improved servo system
September 8, 1992 331/3; 331/79
Back to Brooke's Time & Frequency,