Re: gEDA-user: Capacitors

[sdb@xxxxxxxxxx (Stuart Brorson)]

> The only thing that bothers me about electrolytics is that they are only 
> guaranteed to work for 2000 hours, typically.  But if you choose higher 
> temperature and voltage ratings than needed, then you can get much 
> longer life.

The lifetime is specified at the upper limit temp.  That is, an 85C
rated electrolytic is gauranteed to live for its rated lifetime when
run under bias at 85C.  It may live longer ...

The lifetime follows an Arrhenius activation law, so running the caps
at lower temps dramatically increases the lifetime.  That is, the
lifetime increases by orders of magnitude as the operating temperature
decreases.  

You can get different temp ratings for the caps.   For cheap consumer
junk, 85C is a not uncommon rating.  But for industrial equipment,
you're better off getting electrolytics rated 105C or better.  Due to
the above activation law, the 105C cap run at room temperature will
have *much* longer lifetime than the 85C cap run at room temperature. 

> > They are horrible over temperature, especially cold.  Got into an 
> > argument with older engineer about this.  He said electrolytics lose 
> > 50% value when cold (-40C).  Checked it out, and they don't.  It's 
> > more like 10%-15%.  He was remembering from 30 years ago, maybe.  
> > Point is, they've gotten better (but still suck :-)

Many caps change their value as a function of temperature.  If you
need temperature stability (i.e. you actually care about the cap's
value -- it's not just for decoupling), NPO ceramic caps (for lower
values) or Polypropylene caps are the way to go.   

Stuart