Black Lives Matter - Action and Equality. ... Adafruit is open and shipping.
0

Sizing Ripple Capacitors
Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.

Sizing Ripple Capacitors

by stinkbutt on Thu Oct 14, 2010 5:20 pm

OK, I understand the principle that you throw filter capacitors between your supply voltages and your ground in order to reduce power supply ripple. But what I don't understand is this:

How big should they be? How do you even determine the order of magnitude of how big they should be? Couldn't you just pick the biggest one you can conveniently find, throw it in there, and be done with it? Are there situations where a large one is bad? I've heard rumblings that electrolytics are nice and big, but are bad at higher frequencies, but not other sensible advice on choosing a ripple filter cap. I don't want to just copy what I see in other people's work: I hope to understand some of the maths.

(Even if it is just hand-waving, back-of-a-napkin math.)
Red M&M, Blue M&M: They all wind up the same color

stinkbutt
 
Posts: 593
Joined: Wed Feb 17, 2010 2:40 am

Re: Sizing Ripple Capacitors

by zener on Fri Oct 15, 2010 12:04 am

There is a formula for determining ripple voltage. I studied it last year, so I don't remember what it is, but thankfully there is Wikipedia...

But usually it is done iteratively (trial and error). Electrolytics have poor ESR compared to ceramics. So usually a mix is used. The electrolytics do the heavy lifting and the ceramics handle the high frequencies. So that is why you see ceramics next to chips and regulators, and electrolytics on the main power supply.

And with electroilytics there is the issue of ripple current, which multipled by the ESR gives you heat, and that is what kills the cap, or at least shortens its life.

zener
 
Posts: 4567
Joined: Sat Feb 21, 2009 2:38 am

Re: Sizing Ripple Capacitors

by uoip on Fri Oct 15, 2010 7:05 pm

You need to know the frequency of the ripple you're trying to supress. It might be 120 Hz if you've got an old fashioned transformer->rectifier->filter type supply, but it could be many kilohertz if you're working with a switching supply. You also need to know how much you need to attenuate the ripple. Usually this is expressed in dB.

For the math, see here:

http://www.electronics-tutorials.ws/fil ... ter_2.html

There's another reason for caps, distinct from filtering power supply ripple. Many chips can suddenly attempt to draw spikes of current, or suddenly stop drawing current. These spikes in current demand can cause their own ripples, unrelated to the ripple inherent in the power supply. For this problem, you normally want to put a capacitor right across the chip's power supply leads, as close to the chip as you can. The chip's data sheet generally provides a recommendation as to what size capacitor is appropriate.
uoip
 
Posts: 127
Joined: Wed Sep 23, 2009 5:48 pm

Re: Sizing Ripple Capacitors

by uoip on Fri Oct 15, 2010 7:17 pm

stinkbutt wrote:Couldn't you just pick the biggest one you can conveniently find, throw it in there, and be done with it? Are there situations where a large one is bad?


Should have addressed this in my previous post, but yes, too much can be bad. When you initially switch the device on, the capacitors will be discharged. If you try and put a voltage across a discharged large-value cap, it will initially want a huge rush of current. This can blow a fuse, burn up a power supply, cause a delay in reaching the specified voltage, or have other bad effects.

Also, there are more subtle effects that can happen if you put too much or the wrong kind of capacitor on the output of certain regulators. You can induce oscillations.

http://www.national.com/nationaledge/ju ... icle2.html
uoip
 
Posts: 127
Joined: Wed Sep 23, 2009 5:48 pm

Re: Sizing Ripple Capacitors

by fat16lib on Fri Oct 15, 2010 9:30 pm

An approximate formula is

C = I*t/V

C - capacitor in Farads
I - max current supplied
V - allowable ripple voltage
t - time between charging pulses (1/120 sec for a bridge rectifier @ 60 Hz)

The idea is that the cap supplies power between charging pulses.

I = C (dV/dt) for a capacitor. dV is ripple dt is the time between charging pulses.

This is for a simple unregulated supply. http://my.integritynet.com.au/purdic/power1.html

Real world supply design for high current power supplies is a real art.

Here is an interesting link http://www.smpstech.com/.
fat16lib
 
Posts: 593
Joined: Wed Dec 24, 2008 1:54 pm

Please be positive and constructive with your questions and comments.