[R-390] Replacement Capacitors
Charles P. Steinmetz
charles_steinmetz at lavabit.com
Fri Apr 26 15:55:22 EDT 2013
Bob wrote:
>There is an interesting tradeoff with the otherwise "really really
>good idea" ceramic caps. Even if you go to the X7R dielectric (which
>you should), their leakage is higher than most plastic parts. It's
>not something I would worry about, there's also leakage from dust
>and that spider web there. I don't believe the insulation resistance
>leakage is going to mess anything up, either with the ceramics, or
>between varieties of plastic.
As you say, nothing to worry about. The starting leakage of ceramic
caps is much too high to have any practical effect in bypass
applications. As long as the leakage does not increase over time, as
it does with paper caps, there is no problem. All "Class 2" disk
ceramics (even those with lesser specifications than X7R, e.g. Z5U)
have a well-proven track record in this regard. If a radio has been
underwater for a long time (months to years), the ceramic caps may
get leaky enough to pose a problem. Otherwise, they are very reliable.
I agree that X7R is the dielectric of choice for supply bypass
capacitors, given a choice. The reason to prefer X7R over other
Class 2 ceramic caps is that they hold their capacitance value better
at high temperature, over time ("aging effect"), with higher applied
voltage, and at higher frequencies. For example, an X7R cap is
specified to be within 15% of its 25C value from -55C to 125C, while
a Z5U cap is specified to be within +22% and -56% of its 25C value
from 10C to 85C. An X7R cap may decrease in value by ~10% over time,
while a Z5U cap may decrease in value by ~20% over time. However,
for real-world use as a supply bypass in boatanchors, one can offset
these effects by using two Z5U caps in parallel in place of an
X7R. (One Z5U of double the value should also work, but will bring
the self-resonant frequency ("SRF") down by about an octave compared
to the two smaller caps in parallel.) The leakage of X7R and Z5U
caps of a given value and voltage rating is approximately the same,
and is entirely negligible in supply bypass applications.
Best regards,
Charles
More information about the R-390
mailing list