2
$\begingroup$

What is the purpose and benefit of the two capacitors on both the input and output pins of a LDO Regulator?

LDO Circuit

$\endgroup$
1
  • $\begingroup$ Note that we have a EE.SE site where specifically EE-related questions may get more specific answers than here. $\endgroup$ – Olin Lathrop Feb 21 '15 at 17:03
4
$\begingroup$

The capacitors serve two distinct functions

  • They serve as "reservoir" capacitors. If the regulator is somewhat remote from the energy source then sudden current peaks will cause $V_{in}$ to dip due to V = I x R drop in the wiring. $C_{in}$ serves to provide current for such peaks. $C_{out}$ acts similarly. It is charged to Vout_regulator in normal operation and if a heavy transient load exceeds the regulator's ability $C_{out}$ acts as a short term energy source. Both these roles are the same as for any conventional 3 terminal regulator.
  • They (usually just $C_{out}$) serve to provide "loop stability" or phase margin in some cases. A 3 terminal regulator is a high gain amplifier with negative feedback. Under certain circumstances the phase shift in the amplifier can become such that the feedback is converted to positive feedback and you have a quite capable oscillator.

Adding $C_{out}$ modifies the "transfer function" of the overall amplifier (that we see as a regulator).
The data sheet will specify capacitance values for $C_{out}$ which allow the overall system to be stable.
As well as a range of capacitance values $C_{out}$ often requires a specific ESR (equivalent series resistance) range. Too much or too little ESR and again you have positive feedback and oscillation. The "locus" of ESR and capacitance values is always given in the datasheet when it matters.

$C_{in}$ is not usually involved in stability calculations.

$\endgroup$
3
  • $\begingroup$ I'm not sure how in particular LDO are affected, but essentially (2) is the main reason for the output-side cap. What most regulators produce on output is not a steady DC voltage, but some kind of high-frequency noise that averages out to the target DC voltage. It's the capacitors that make it into neat, stable DC. $\endgroup$ – SF. Feb 19 '15 at 2:05
  • 1
    $\begingroup$ @SF Alas I have to disagree with several parts of what you say. (2) is seldom the reason for older regulators - it's modern LDOs that it applies mainly to AND it applies only for some - its related to design decisions and some LDOs are stable without the cap (and SOME would be unstable if some values of cap were added). | As for "main reason" - this may OFTEN be so but in some cases you may have a load with short very high peaks well over Imax_reg and 1. is then critical (An RF burst TX may be such). | Your comments on HF noise that average out is very misleading too , alas. .... $\endgroup$ – Russell McMahon Feb 19 '15 at 3:24
  • 1
    $\begingroup$ .... This is much more true of switch mode regulators, which is not what this is about. For linears LDOs while there will be HF noise and feedback loops 'hunting' and more it is not a major role of either Cin or Cout to deal with this. Look at datasheet for eg older LM340 / 7805v where stability does not complicate things and see what they say about the caps. From memory they say neither is essential but Cin helps when input wiring run is long and Cout helps with transient loads (which is what I said above AFAIR). $\endgroup$ – Russell McMahon Feb 19 '15 at 3:27
1
$\begingroup$

The ESR of the Output capacitor is important for setting up the a pole in internal closed-loop system of LDO Regulator. This close loop is responsible for controlling the current allowed depending on the feedback voltage and reference voltage.

The ESR of capacitor helps maintain stability of the LDO control loop. The output capacitor also can help manage rapid changes in load current (transient responses). Using a larger value helps improve transient response of the LDO but this also increases start-up time of the LDO.

Also changes in output capacitor has impact in the input capacitor too.

It is important to place a good capacitor based on the data sheets recommended capacitance and ESR values. Ceramic capacitors using X5R or X7R dielectric is a good choice. This is because they have good temperature stability and good low voltage co-efficiency.

Reference:

$\endgroup$
3
  • $\begingroup$ Ceramics are good BUT beware the potential affects of transients on Cin. In SOME cases where a very sharp input edge occurs ceramic caps will "ring" and may produce voltageslarge enough to destroy the regulator. This is not common - but that makes it all the more worth being aware of as otherwisethe reason for damage may be obscure. $\endgroup$ – Russell McMahon Feb 19 '15 at 3:31
  • $\begingroup$ Can you please define the sharp edge in terms of voltage. Do you consider a 20V a sharp edge. BTW, Most of the answer content is a summary from the reference. $\endgroup$ – Mahendra Gunawardena Feb 19 '15 at 12:35
  • $\begingroup$ Replying to your two year old question :-) - sorry I missed it at the time. It seems that the rate of change of voltage is more important than the absolute voltage level. I admit to not knowing the mechanism in any detail - but it certainly happens. It can probably be modelled by an LC resonant effect. I have read accounts of devices being destroyed by input transients within the s"safe" Vin range causing overvoltage. $\endgroup$ – Russell McMahon Jun 21 '20 at 21:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.