Connect with us

Capacitor requirements?

Discussion in 'Power Electronics' started by Vagulus, Jun 5, 2018.

Scroll to continue with content
  1. Vagulus

    Vagulus

    26
    1
    May 6, 2018
  2. Harald Kapp

    Harald Kapp Moderator Moderator

    9,539
    1,967
    Nov 17, 2011
    These decoupling capacitors should be as near as possible to the regulators. C2 near REG1, C3 near Reg2.
    As the datasheet states on page 1: The input capacitor is only required if the regulator is far from the power source. Therefore C3 is not required if Reg2 is neqar Reg1. Using 1μF instead of 0.47 μF is no issue.
    Also C2 and C4 can easily be made greater than 22 μF or 33 μF, respectively. See footnote ** on page 1 of the regulator datasheet.
     
    hevans1944 and davenn like this.
  3. Vagulus

    Vagulus

    26
    1
    May 6, 2018
    I have to ask the age old question, "What is far?"
    Perhaps it's not only beauty that's in the eye of the beholder :rolleyes:
     
  4. Harald Kapp

    Harald Kapp Moderator Moderator

    9,539
    1,967
    Nov 17, 2011
    1 cm is near. 100 cm is definitely far.
    Anything in between? In my view it is good practice to provide the capacitors regardless of distance. Other parameters like trace width and thickness also play a roe.
     
  5. Vagulus

    Vagulus

    26
    1
    May 6, 2018
    Thanks Harald. I think I, too, will stick with putting the capacitors in (just for luck :D).
     
  6. elebish

    elebish

    177
    12
    Aug 16, 2013
    Generally, a regulator must have an unregulated input to be able to have a regulated output when the load changes. I suspect that both regulators should be fed by the same unregulated input as long as the 3.3 regulator can handle VDD with the current being drawn by the 3.3 volt regulator.
     
  7. elebish

    elebish

    177
    12
    Aug 16, 2013
    As for the caps, use the values called out. Using values too high will cause more current draw when first turned on! Those caps are for the purpose of maintaining a steady output under varying loads and to filter out fluctuations from the ac input after rectification.
     
  8. Harald Kapp

    Harald Kapp Moderator Moderator

    9,539
    1,967
    Nov 17, 2011
    I'm sorry but that is not true. With a linear regulator you can most easily regulate to a lower voltage from any higher voltage as long as the difference between input voltage and output voltage is sufficiently high. For low drop regulators a few 100 mV often suffice, It is not the regulation or non-regulation of the input voltage, it is the input-output voltage differential that counts.
    A valid point. However, a few microfarads more will do no harm. There may be even more and higher capacitances to be charged from the regulated 3.3 V and 5 V lines which we do not see here because they belong to the circuit that is attached to this regulator scheme.
    Kind of. These capacitors are required to keep the regulator's feedback loop stable. A little known fact is that the output capacitor often is required to have a certain minimum ESR, not less, for the regulator to operate stable, The theory is explained e.g here.
    Not those capacitors shown. For smooting the rectified AC you typically use much larger capacitors. As a rule of thumb 1000 μF per 1 A of output current - although more precise calculations can be made taking into account maximum allowed ripple voltage, minimum AC input voltage etc. Again the capacitor at the input is for stability purposes, see above,
     
    hevans1944 likes this.
  9. elebish

    elebish

    177
    12
    Aug 16, 2013
    If the input to the 3.3 regulator is only 5 volts regulated, is that sufficiently higher than the 3.3 volt output to give good regulation? Remember, the input can not change due to the first 5 volt regulated output! As a rule, I never use 2 regulators in series! As for the caps (as you mention about the feedback loop) the feedback loop is internal to the regulator which helps to keep the output stable! This is the second time I have been corrected on this site! I think I may be getting linear and pass regulators mixed up! Boy, my old age must be getting to me!
     
  10. Harald Kapp

    Harald Kapp Moderator Moderator

    9,539
    1,967
    Nov 17, 2011
    The datasheet of the LM3940 states a dropout voltage of max. 1 V at 1 A. Therefore an input voltage of 4.3 V would suffice, certainly 5 V will fulfill this condition.

    Right, but the output capacitor is part of the loop as it contributes to the phase shift, as explained in the TI article I linked.

    I'm sorry. Please let that not disencourage you.

    You may have other valuable assets up your sleeve then...
     
  11. BobK

    BobK

    7,671
    1,681
    Jan 5, 2010
    That's nonsense. A regulator will work fine with a regulated input as long as it is higher than the output voltage + the dropout voltage.

    Bob
     
    hevans1944 likes this.
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-