[time-nuts] FE-5680A heat sink

Poul-Henning Kamp phk at phk.freebsd.dk
Mon Jun 8 12:36:44 UTC 2009

In message <4A2CFCE2.5020107 at erols.com>, Chuck Harris writes:

>> I ran the device today for about half an hour, and used an infrared 
>> sensing thermometer to measure the external case temperature.

Be very careful about trusting this:  you need to do some tricky
calibrations to get anywhere near precise when you measure metal

The easy way, is to put a piece of duc[kt]tape on the metal surface
and make sure your thermometer can see only that surface.

Unfortunately, the tape will also act as insulation, so the result
you get is not precise even then.

>> It got up to 48 C externally in the physics package area.

That's quite normal.

Those small Rb's keep the internal temperature constant using
heaters, which can raise the temperature and by being able to dump
excess heat through their heat-sink to lower the temperature.

You shouldn't run your Rb too hot, as this decreases the electronics
lifetime and reduces the wiggle-room of the thermal management
inside the device.

On the other hand, cooling it too much will only increase the
power drain for the heaters and increase the thermal gradients
inside the unit, likely degrading thermal stability.


Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.

More information about the time-nuts mailing list