Shop OBEX P1 Docs P2 Docs Learn Events
The "Black Art" of choosing bypass capacitors — Parallax Forums

The "Black Art" of choosing bypass capacitors

Paul BakerPaul Baker Posts: 6,351
edited 2008-05-01 14:16 in General Discussion
Ever wondered whether the bypass capacitor you've chosen is adequate for your application, or always choose the same value but would like to know the science behind it so that you can use the minimum value needed in your design to reduce production cost? Here's an article that explains what you need to know to make an informed decision on your bypass caps: http://www.edn.com/contents/images/454638.pdf

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer

Parallax, Inc.

Comments

  • Lord SteveLord Steve Posts: 206
    edited 2008-04-24 23:22
    Um...any relation?
  • william chanwilliam chan Posts: 1,326
    edited 2008-04-25 13:09
    To me a 0.1uF capacitor or 0.47uF capacitor is about the same price, so there is no significant savings in choosing the smallest working capacitance.
    I'd just go for 0.47uF for all the bypass caps so I don't have to stock so many values.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    www.fd.com.my
    www.mercedes.com.my
  • Paul BakerPaul Baker Posts: 6,351
    edited 2008-04-25 21:52
    No relation. William, the reason I posted the article is that there's more to it than just picking a one size fits all, and I thought people would like to see the nitty gritty. Take a look at figure 1b, see how each value has a notch? If you optimize the capacitor value you can get that null to correspond with the switching frequency (like 80Mhz for the Propeller for instance). That 0 ohm path to ground short circuits the·switching frequency noise and doesn't allow the voltage ripple to affect the the rest of your circuit. Now for pure digital designs, doing this sort of optimization isn't going to make a difference in anything but >100 MHz designs, but if you are doing a mixed mode design (analog and digital), quieting your power supplies makes a world of difference in the accuracies you can achieve in your analog section, even when you have seperate supplies.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Paul Baker
    Propeller Applications Engineer

    Parallax, Inc.
  • william chanwilliam chan Posts: 1,326
    edited 2008-04-27 05:14
    The graph in Figure 1b seems a little odd.
    After the "touchdown" to zero ohms, the impedance seems to increase again with frequency.

    This goes against the formula for impedance calculation Z = 2pi C/F ( something to that effect ).
    Why is that so?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    www.fd.com.my
    www.mercedes.com.my
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2008-04-27 06:49
    Real-world capacitors are not purely capacitive. There are also parasitic resistive and inductive components that affect their impedance at various frequencies. The "touchdown" points you refer to occur at the capacitor's self-resonant frequency.

    -Phil
  • Beau SchwabeBeau Schwabe Posts: 6,568
    edited 2008-04-27 07:11
    william chan,

    The graph is perfectly normal


    Impedance, Z = √ (R^2 + X^2)

    Where:
    R = Resistance
    X = Reactance


    Total Reactance, X = XL - XC

    Where:
    XL = Inductive reactance
    XC = Capacitive reactance

    Inductive reactance, XL = 2pifL

    Where:
    f = frequency in hertz
    L = inductance in henrys

    Capacitive reactance, XC = 1 / 2pifC

    Where:
    f = frequency in hertz
    C = capacitance in Farads


    Note: The reactance is capacitive if the input frequency is above the frequency pole. (HIGH PASS FILTER)
    The reactance is inductive if the input frequency is below the frequency pole. (LOW PASS FILTER)

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Beau Schwabe

    IC Layout Engineer
    Parallax, Inc.
  • william chanwilliam chan Posts: 1,326
    edited 2008-04-27 13:35
    You guys are really smart.

    What about multilayer capacitors, would the frequency response be about the same as ceramic capacitors?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    www.fd.com.my
    www.mercedes.com.my
  • Bob Lawrence (VE1RLL)Bob Lawrence (VE1RLL) Posts: 1,720
    edited 2008-04-27 14:11
    Some of you digital guys and gals may find her book very useful. She mentions bypass caps in chapter 11.

    A Baker's Dozen: Real Analog Solutions for Digital Designers (Paperback)
    by Bonnie Baker (Author) "

    www.amazon.com/Bakers-Dozen-Solutions-Digital-Designers/dp/0750678194

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Aka: CosmicBob
  • mparkmpark Posts: 1,306
    edited 2008-05-01 09:40
    I hope this isn't too much of a threadjack. I have a D/A chip with VSS, VDD, DGND, and AGND. Where should I put the bypass capacitor(s)?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Michael Park

    PS, BTW, and FYI:
    To search the forum, use search.parallax.com (do not use the Search button).
    Check out the Propeller Wiki: propeller.wikispaces.com/
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2008-05-01 10:18
    Consider the source of your noise. If it is from the power Vss and Vdd are the choice for by-pass capacitors, the most common problem. Regarded the DGND and AGND, are you getting noise that needs to be removed? Where is it coming from? AGND is related to output, not input - so that pretty much excludes it. DGND? maybe...... Is there isolation involved?

    In principle, the by-pass capacitor is just a special case of using a capacitor as a low-pass filter.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    PLEASE CONSIDER the following:

    Do you want a quickly operational black box solution or the knowledge included therein?······
    ···················· Tropically,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan
  • Mike GreenMike Green Posts: 23,101
    edited 2008-05-01 14:16
    A good starting point is to follow the manufacturer's recommendations. They usually include one or more test circuits or sample circuits in their datasheets and/or application notes.
Sign In or Register to comment.