[time-nuts] time-nuts, frequency counters
Richard (Rick) Karlquist
richard at karlquist.com
Thu Sep 27 09:45:28 EDT 2007
>> The simple answer, is it introduces another place for drift
>> to occur. If the power supply that provides the current to
>> create the C-field drifts with temperature, component aging,
>> power line voltages, phases of the moon, ... your reference's
>> frequency will also drift.
>> -Chuck Harris
> I must be missing something. Aren't all those bad things going to happen
> whether you adjust
> the C-field or leave it untouched? Isn't the C-field required for the Rb
> to work dependably
> and on freq? I think the question pertains to Rb sources that already
> have a screw to make
> the adjustment. Maybe you are saying the screw is there to put the
> C-field back where it belongs.
> I would have no way of knowing how to tell if it was right or wrong
> other than frequency.
> Rick said it is not a good way to adjust the frequency. I don't really
> understand why exactly either.
There is an optimum range for the C-field. Too little will result in
insufficent Zeeman splitting and too much makes the frequency too
sensitive to magnetic field. Remember that frequency is proportional
to the SQUARE of the C-field intensity. The C field screw is not
designed to "put the C-field back where it belongs" but rather to
move it away from its comfort zone to pull the frequency. I'm not
an expert on Zeeman splitting, but I have heard people talking about
finding a setting to avoid any spurious lines, as if the setting was
like finding a clearing in a forest of lines.
It is true that an adjustable C-field has the potential to drift more
than a fixed C-field, but this is only one issue, that may or may not
be pivotal. Another issue with changing the C-field is that it may
affect the magnetic hysteresis in the magnetic shields.
With a synthesizer, you introduce no drift due to adjustability, you
always operate at the optimum C-field value, the adjustment is linear
and predictable, and the range of adjustment is not limited, so you
can introduce any offset you want, and it is remotely, digitally,
programmable. And with modern synthesizers, this capability may well
be virtually free. Even the HP10816 Rb standard, designed in the
1970's, had a synthesizer that we set to get on frequency, however this
synthesizer was not intended to be user programmable.
BTW, "C-field" is a bastardized term borrowed from cesium beam
technology where there were "A" and "B" magnets to select the state
and the C-field to bias the magnetic field in the interaction region.
Hence "A-field" "B-field" and "C-field". Of course, Rb standards have
no A or B fields. It is sort of logical in an "inside baseball" sense.
Rick Karlquist N6RK
More information about the time-nuts