Maker Pro
Maker Pro

DCDC buck converter output capacitor placement

I have a microprocessor with built in DCDC converter. We need to provide inductor and output capacitors externally at the respective pins of the processor. The regulated o/p from the internal DCDC converter is for internal core and IO as well as for external peripherals. The manufacturer recommends using 0.01uf, 0.1uf, 1uf and 33uf at the DCDC output pin. What should be the order of placement of these o/p capacitors on the board? 0.01uf close to the pin and 33uf away or the other way? The current from the o/P pin is both ways. i.e goes internal to the IC as well as goes out of the pin to the other peripherals on the board.

-bjnk
 
J

Joerg

Jan 1, 1970
0
I have a microprocessor with built in DCDC converter. We need to
provide inductor and output capacitors externally at the respective
pins of the processor. The regulated o/p from the internal DCDC
converter is for internal core and IO as well as for external
peripherals. The manufacturer recommends using 0.01uf, 0.1uf, 1uf and
33uf at the DCDC output pin. What should be the order of placement of
these o/p capacitors on the board? 0.01uf close to the pin and 33uf
away or the other way? The current from the o/P pin is both ways. i.e
goes internal to the IC as well as goes out of the pin to the other
peripherals on the board.

Smallest capacitance closest to the pin. But nowadays a 0.1uF in 0603
package is good enough, there is usually no need for an extra 0.01uF.
The 1uF is debatable, may not be needed if the 33uF is a ceramic.
 
That sounds like overkill. I'd go for 2x15uF ceramics. The node should
be a big copper pour, adjacent to the ground plane. Another cap, 330
nF or some such, near each peripherial chip wouldn't hurt, but don't
overdo it.

What uP is it? Built-in buck converters are uncommon.

I wonder why?

I remember working a zigbee chip years ago, it had an on-chip LDO
since it the chip was 1.8V and usually it would run on a ~3V battery

everyone looked at me funny when I suggested adding a buck converter,
it was really only a matter of adding a switch on chip and an external
coil
and diode it didn't have to be very accurate and in sleep mode where
the chip only used uA the switch could just be permanently on since it
would
be feeding the LDO

I know it got put in, but I don't know how well it worked


-Lasse
 
Top