how important is a regulator output capacitor?

[darron]

i just made a multi-isolated pose supply for myself but i want to sell it as a prototype to make a new one better. it got me thinking, what if someone plugs DC/AC into one of the DC outputs? it could cook the 22uF caps on the output of the regulator (LM317 style).

started thinking, how important is that output cap?

the data sheets don't really show an output cap often, except where it shows how to put a protection diode in if there is an output cap. then, it shows a non-polarised cap. at one point it labels it as 1uF. here's the data sheet i'm looking at btw: http://docs-asia.electrocomponents.com/webdocs/0780/0900766b807801e4.pdf




i pulled apart a Boss regulated power supply pack for repair once. those power supplies used to be regarded as "the sh*t", the best. from memory, the i think configuration was only TRANSFORMER > RECTIFIER > BIG FILTER CAP > REGULATOR > OUT and they never have problems


most of the PSU schematics out there put a big cap on the output. i only put a smallish electro there for good measure, figuring it wasn't as important as the 1,000uF ones i've seen.


what about just a decent 1uF MKT cap? would there be ANY difference in output ripple compared to say a 100uF electro (+small NP cap)?



just after some opinions.

[merlinb]

With the common reg ICs you only want a *small* output cap. I never use more than 10uF. The very large values you sometimes see actually tend to make transient response worse! I think you'd be fine with a 1uF non-polar. Or you could put a protection diode in parallel with the cap?

Or, if someone plugs the wrong thing into the wrong bit you can just say it's their fault for being careless. Can't wrap the world in cotton wool...

[Seljer]

I thought the issue here is that the regulators want some capacitance on the output to ensure stable operation under varying conditions (i.e. it doesn't start oscillating by itself and such). Ideally the AC ripple should already be taken out the regulators circuitry. It doesn't need be be much....and good practice when building effects is for each stompbox you plug into your power supply to have its own power supply capacitor

I believe this section of the datasheet is relevant

EXTERNAL CAPACITORS

An input bypass capacitor is recommended in case the
regulator is more than 6 inches away from the usual large
filter capacitor. A 0.1μF disc or 1μF solid tantalum on the
input is suitable input bypassing for almost all applications.
The device is more sensitive to the absence of input bypassing
when adjustment or output capacitors are used, but the
above values will eliminate the possibility of problems.

The adjustment terminal can be bypassed to ground on the
LM317L to improve ripple rejection and noise. This bypass
capacitor prevents ripple and noise from being amplified as
the output voltage is increased. With a 10μF bypass capacitor
80dB ripple rejection is obtainable at any output level.
Increases over 10μF do not appreciably improve the ripple
rejection at frequencies above 120Hz. If the bypass capacitor
is used, it is sometimes necessary to include protection
diodes to prevent the capacitor from discharging through
internal low current paths and damaging the device.

In general, the best type of capacitors to use is solid tantalum.
Solid tantalum capacitors have low impedance even at
high frequencies. Depending upon capacitor construction, it
takes about 25μF in aluminum electrolytic to equal 1μF solid
tantalum at high frequencies. Ceramic capacitors are also
good at high frequencies; but some types have a large
decrease in capacitance at frequencies around 0.5MHz. For
this reason, a 0.01μF disc may seem to work better than a
0.1μF disc as a bypass.

Although the LM317L is stable with no output capacitors, like
any feedback circuit, certain values of external capacitance
can cause excessive ringing. This occurs with values between
500pF and 5000pF. A 1μF solid tantalum (or 25μF
aluminum electrolytic) on the output swamps this effect and
insures stability.

The schematic on page seven shows how to add a capacitor on the output and on the adjust pin with the appropriate protection diodes (if you unhook the input side and the on the output side the capacitors still retain a charge, this charge runs over the diodes instead of through the regulator).

Regarding plugging stuff into the output, you could put a zener diode on the output, which would make any current from a voltage too high or from a negative voltage run over the diode instead of over the chip.

[darron]

thanks guys. i think i'll make them 1uF MKTs then.. they should last forever too!

i used a protection diode across the regulator, but nothing to like a parallel protection diode.



i used a switch mode power supply input with 2.1mm DC sock, negative outside, so opposite polarity to the standard outputs. could easily stuff that up. used different coloured sockets on opposite sides, and will definitely label them well.



here's a pretty picture of what got slapped together... i had a spare enclosure, i know the arrangement isn't ideal: