[time-nuts] Airflow around oscillators.
billj at ieee.org
Sun May 22 11:19:57 EDT 2005
David Kirkby wrote:
> I am building a GPS locked oscillator and will have in the box the
> 1) GPS receiver
> 2) 24 V linear power supply I have had sitting here for ages - a use
> for it at last!!
> 3) PRS10 rubidium
> 4) 10811-60111 quartz
> 5) 2 x 7 Ah 12V sealed lead acid batteries.
> 6) Shera board, to lock the 10811 to either the rubidium or GPS
> (interesting to see what gives best performance)
> 7) A couple of DC-DC converters to give other voltages necessary for
> the Shera board. This seems more sensible than separate power
> supplies, due to my desire to have battery backup. I don't want a
> battery for +5V, another for -5V etc.
> 8) A small amount of control logic, to sequence the ovens on the
> quarts and rubidium, as I can't have both heating up at the same time,
> although they can both run together later.
> I need to keep the batteries cool, so are planning on blowing cool air
> over them and around to the power supply.
> I'm wondering if its best to try to reduce the airflow around the
> 10811-60111 and PRS10 (rubidium) to a minimum necessary to keep the
> things within temperature, or is airflow over them not going to be an
> issue? That would make life a bit easier, if I can ignore that.
> I'm also thinking of using a thermostatically controlled fan, in an
> attempt to keep the temperature inside the box a constant. The unit
> will *not* be run in a temperature controlled environment, but one in
> which there are quite large changes in temperature.
> Whilst I don't mind experimenting with such things
> a) Whether the theromostaticaly controlled fan gives smaller or larger
> changes in the rubidiums base temperature (it has a sensor which can
> be read.
> b) What gives the best performance (10811 or PRS10)
> c) What it is best to lock the 10811 to (GPS or PRS10)
> I am not too keen on treating this lot as mechano and trying every
> possible combination of component placement. Once placed, they will
> stay there, with the exception of perhaps some thermal insulation and
> the location of a temperature sensor for the fan if used.
> Any thoughts?
My opinion (such as it is) is to minimize the temperature changes around
the 10811 and PRS10.
Especially I would minimize sudden changes of temperature that may
exceed the speed that the
oven control can compensate for. Even then I wouldn't expect that the
10811 (or PRS10) can
completely stabilize the internal temperature in the face of extreme
external temperature changes.
My opinion (again) is that the GPS will have better long term stability
(days) than the PRS10.
Therefor I would lock the 10811 to GPS. That way the 10811 will clean up
the short term
jitter from the GPS and the combination can be used to discipline the PRS10.
Don't have a PRS10 and have no experience with it so can't give fist
More information about the time-nuts