[time-nuts] HP/Agilent 5065A Rubidium Vapour frequency standardproblem

Bill Hawkins bill at iaxs.net
Thu May 8 15:56:12 EDT 2008


Found the book and looked at your symptoms.

A) The frequency output is drifting near the crystal rate, so the
control loop is open.

By "loop" I mean the path from the OXCO A10 to the synthesizer A1 and
multiplier A3 to the RVFR to the photo amplifier A7 to the phase
detector A8 to the integrator A9 to the frequency control of OXCO A10.
Something in that loop is bad.

B) The CONTROL reading is pegged because the loop is open. The
integrator A9 can't affect the measurement from A8.

C) Adjusting OXCO A10 doesn't work. Something after it in the loop is

D1) A8TP3 is a straight line. Assuming your scope is sensitive to 50
mv/cm, this indicates no measurement from A7. An Error reading of zero
on the meter no matter what you do to coarse frequency will confirm
this. Keep the frequency within 10E-7 of 5 MHz, though, or you will be
too far from RVFR resonance to get an error.

D2) A8TP2 is fine. This indicates that A8 is probably providing the 137
Hz phase modulation to A3.

The unit runs fine from a hot start with a cold OXCO. This indicates
that there are no fatal problems in the control loop.

CONTROL is pegged after 8 hours. This indicates a drift problem. It
could be the integrator capacitor C5 on A9, but the same symptoms occur
if the error signal from A8 stays at zero regardless of the OXCO error.

Another indication of low or no error signal comes from the 2nd Harmonic
reading on the meter. This is the second harmonic of the 137 Hz signal
sent to the multiplier A3. When the control loop is operating properly
(phase locked), the 137 Hz signal from the RVFR is very near zero,
leaving the 2nd harmonic dominant. Table 3-1 says the reading should be
20 to 50 (it will bounce around). My 5065A reads 11-12.

You can confirm that the integrator is OK without turning off power by
switching from OPER to OPEN LOOP. (Don't touch coarse or fine frequency
adjustments.) This shorts C5 and should drop the CONTROL reading to
zero. If ERROR stays at zero, the problem is most likely in the RVFR or
amplifier A7. Go back to OPER and CONTROL should drift away from zero.
The 5 MHz output of the 5065 should also drift as the CONTROL voltage
changes, since it comes directly from the OXCO.

Finally, the RVFR is old, with continuous use. They don't last forever.
In particular, the glass physics package can darken. I tried the
procedure in section 3-9 with the current set at 1 amp for 5 days. This
operates a cold trap that condenses excess rubidium from the glass
walls, allowing light to get to the photocell amplified by A7. It worked
well enough to regain control, but the 2nd Harmonic reading was low.

Bill Hawkins

More information about the time-nuts mailing list