More cap questions I couldn't answer . . .

[Mark Amundson]

media at mail1.nai.net wrote:
> ... A few weeks ago, I had added the following questions as notes to myself in
> this entry in the original FAQ.  Since then I haven't found an answer.
> 
> > POWER SUPPLY CAPS - For the cap between the diodes and the regulator,
> > use a cap rated to 105 degrees C. They are only slightly more expensive
> > (pennies) than the 85 deg. C types, but much higher quality.  For the
> > output caps on 3-terminal regulators, use a 1, 2.2 or 3.3uf tantalum
> > [**WHY TANTALUM??  WHY NOT ELECTROLYTIC OR SOMETHING ELSE??**] in
> > parallel with a 0.1uf ceramic.  [**AGAIN, WHY CERAMIC??**] MARK

If the large electrolytic caps are used in high hear situations like
power amplifiers or in stagnent air near 3 terminal regulators, then the
+105C rated maybe justified. I would rather see a thermal design for a
project that can assure a +85C or less rating so that other thermally
sensitive signal processing circuits do not act up.

The output capacitors on three terminal regulators should present a low
Effective Series Resistance (ESR) for a bandwidth beyond 100kHz for
regulator stability. If you use only one capacitor, then low value (1uf
to 10uf) tantalum or Panasonic HF/HFU type electrolytic will suffice. I
prefer to use a .1uf ceramic near the regulator input pin, a 10uf
tantalum near the output pin, and sprinkle .1uf ceramics near critical
signal processing circuits. I save the polypropylene for signal
decoupling or eltra high-end power amplifiers were I can bypass large
electrolytics with a couple of different size polyprops each.

The reason for ceramic over film type caps is that most film caps are
wound instead of stacked. Their parasitic inductance of film winding
reduces the benefit at higher frequencies.

Mark Amundson,