I'm using LTC3533 buck-boost regulator. The output capacitor is 100uF ceramic based. After checking the manufacturer's datasheet ,it turns out that at the switching frequency( 1MHz) the capacitor behaves as an inductor. Equivalent capacitance at the DC bias is around 48uF but will it continue to serve my purpose as a capacitor and filter out the high frequency ripples and how does it affect the design of compensation network?
2
-
1\$\begingroup\$ Data sheet links please. \$\endgroup\$– Andy akaCommented Jan 14, 2016 at 13:08
-
\$\begingroup\$ For compensation purposes it's a 48uF capacitor. For filtering, see @FakeMoustache's answer. \$\endgroup\$– Spehro 'speff' PefhanyCommented Jan 14, 2016 at 16:08
Add a comment
|
1 Answer
\$\begingroup\$
\$\endgroup\$
Does that 100uF cap meet the requirements listed in the LTC3533 datasheet ? I see a low ESR requirement. Note that an ideal 100 uF capacitor would have an equivalent impedance of 1.5 m ohms, that is 0.0015 ohms !!
I dare you to connect that ideal capacitor to the LTC3533 using wires with a resistance less than that 1.5 mohms. Indeed, that is practically impossible.
Conclusion: any 100 uF cap will not behave as 100 uF at 1 MHz, this is to be expected. Just use a capacitor as recommended by the manufacturer and than will be good enough.
If you're worried about 1 MHz (and higher) spurs on the supply then I suggest adding a series choke and a good quality, but much smaller than 100 uF, decoupling capacitor after that choke to filter the supply a bit more. I would use 2 caps like a 1 uF and a 100 pF cap in parallel after the choke.
The choke separates the 100 uF cap from the 2 decoupling caps so the feedback network will not be affected (which is what you want).
answered Jan 14, 2016 at 13:50