[time-nuts] chip scale atomic clock

[Magnus Danielson]

Dear all,

After very mild amount of homework, I think a followup was due.

Magnus Danielson wrote:
> Peter Vince wrote:
>> 2009/12/26 Robert Lutwak <Lutwak at alum.mit.edu>:
>>> ...
>>> CSAC is intended for portable battery-powered operation. Surely your
>>> basement has the space and wallplug power to support an LPRO. (p.s. 
>>> don't
>>> cool the damn thing, heat it).
>>> ...
>>
>> Hi Robert,
>>
>>      Do I understand you are suggesting heating an LPRO, not cooling
>> it?  That seems to go against what I understood, that greater cooling
>> leads to increased life.
> 
> While not directed to me, these are my understandings:
> 
> Besides the power applied to heat the Rb lamp, the physical package 
> needs to be at the sweet-spot in temperature, so heating is performed.

Looking in the LPRO manual, as found in say:
http://www.ham-radio.com/sbms/LPRO-101.pdf
and the LPRO repair-guide:
http://www.radcomms.net/EFRATOM%20LPRO%20101%20Repair%20Guide.pdf

The Rb lamp heats to 110 C and the physical package to 78 C.
Notice also figure 1.3 displaying power dissapation as a funciton of 
baseplate temperature. The simplified model for power consumption in 
chapter 3.2.3 gives a good hint about what is going on.
For 20 degrees the RF lamp consumes about 1,7 W where as for 70 degrees 
it consumed about 750 mW. Similarly, for 20 degrees the physical package 
heating consumes about 3,8 W where as for 70 degrees it constumed about 
520 mW. Thus, allowing the increase of baseplate temperature from 20 
degrees to 70 degrees reduces the Rubidium assembly heating from a total 
of 5,5 W to 1,3 W. Looking at figure 1.3 and the equation again, we see 
that about 280 mA derives from the other electronics and that a lower 
(18 V) supply has significant shift in power. Thus, by paying attention 
to supplied power and baseplate temperature and cooling (which becomes 
more important to maintain baseplate below 70 degrees) less power 
dissapation can be achieved. With that in hand, both passive and active 
ovenizing could be considered. 5-6 W is significantly lower than 12-13 W 
and should allow for simpler solutions.

There is also hints about how to temperature compensate the LPRO by 
steering the C-field from a temperature sensor. A sensible ovenization 
should reduce the need of such approaches, even if possible. Boxing one 
up similar to that of Thunderbolts may be the way to go.

Cheers,
Magnus