[time-nuts] Thunderbolt? (re simple gpsdo.)

Charles P. Steinmetz charles_steinmetz at lavabit.com
Tue Jan 3 14:03:52 UTC 2012


Hal wrote:

>Does anybody know anything about the temperature coefficients of large caps?
>
>I'm not interested in the frequency shift of the filter as the temperature
>but the voltage shift due to a fixed charge as the capacitance changes.

The common rule of thumb is "thousands of ppm per 
degree C" through their normal operating range 
(i.e., neglecting low temperatures where they are much worse).

Here is what Cornell-Dubilier says about the 
temperature coefficient of capacitance of aluminum electrolytics :

>The capacitance varies with temperature. This variation itself is
>dependent to a small extent on the rated voltage and capacitor
>size. Capacitance increases less than 5% from 25 ºC to the high
>temperature limit. For devices rated –40 ºC capacitance declines
>up to 20% at –40 ºC for low-voltage units and up to 40% for high
>voltage units. Most of the decline is between –20 ºC and –40 ºC.
>For devices rated –55 ºC capacitance typically declines less than
>10% at –40 ºC and less than 20% at –55 ºC.

(see 
<http://www.google.com/url?sa=t&rct=j&q=aluminum+electrolytic+%22temperature+coefficient%22&source=web&cd=1&ved=0CEwQFjAA&url=http%3A%2F%2Fwww.cde.com%2Fcatalogs%2FAEappGUIDE.pdf&ei=MwMDT728Jcfo0QGggpm4Ag&usg=AFQjCNEZwwiXQYXxoXbC_on3M5MlOh3ypg> 
at p. 7)

5% is 50k ppm; an 85C cap would thus change 50k 
ppm over 60 degrees C, or a bit less than 1k ppm 
per degree C.  Based on the timing circuits I've 
seen implemented with aluminum electrolytics, I'd 
say CDE is being optimistic here.

For integrating in an environment where stability 
of parts in 10e10 or better is desired, you are 
likely to find that all of the other 
imperfections of electrolytic capacitors 
(leakage, noise, dielectric absorption, etc.) 
will stop you well short of the goal, never mind 
the tempco of capacitance.  You might get by with wet-slug tantalums.

Best regards,

Charles








More information about the time-nuts mailing list