2
\$\begingroup\$

In LM2596 switching regulator datasheet suggested to use low ESR capacitors. since I'm going to use LM2596 as a pre-regulator for a linear regulator to reduce the voltage drop on linear regulator, is it still necessary to use low ESR capacitor?

\$\endgroup\$
  • \$\begingroup\$ What type of linear regulator? 150kHz switching frequency? \$\endgroup\$ – Justme Nov 1 '19 at 15:57
  • \$\begingroup\$ @Justme linear regulator is LM338 and LM2596 frequency is 150KHz. \$\endgroup\$ – ElectronSurf Nov 1 '19 at 15:59
3
\$\begingroup\$

Use Low ESR caps in the feedback loop and for the filter capacitor after the inductor. The input capacitor, is not as important, but the design might incur more loss and have slightly higher ripple with a higher ESR on the input capacitor. I believe the regulator will help with regulaton, especially with short traces. The best way to find out if the design will be tolerant to a low ESR cap is to simulate it in SPICE.

enter image description here
Source: https://www.onsemi.com/pub/Collateral/LM2596-D.PDF

| improve this answer | |
\$\endgroup\$
  • 1
    \$\begingroup\$ Good point simulate that ESR is not too low. \$\endgroup\$ – Justme Nov 1 '19 at 19:46
  • \$\begingroup\$ There are so many variables with switchers, might as well simulate. \$\endgroup\$ – Voltage Spike Nov 1 '19 at 19:58
4
\$\begingroup\$

You will note that the recommendation for a low ESR capacitor is to maintain stability of the control loop. Even if you are regulating after the LM2596, you would still want to maintain proper stability. Use the recommendation of the datasheet and go with the low ESR cap.

| improve this answer | |
\$\endgroup\$
  • 1
    \$\begingroup\$ Except that the datasheet specifies min and max ESR values allowed for stability, and many low ESR caps can be below this limit, so compatibility needs to be checked. \$\endgroup\$ – Justme Nov 1 '19 at 19:40
2
\$\begingroup\$

It is extremely important to use capacitor that is withing the allowed ESR limits, but as low as possible, because linear regulators can be bad at handling high frequency ripple so less ripple is better. If you look at LM338 datasheet, ripple rejection is very near 0dB at 150kHz and it can even amplify ripple at higher frequencies. Consider adding extra filtering between regulators, or consider if some other regulator is better.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Is This capacitors any good? \$\endgroup\$ – ElectronSurf Nov 1 '19 at 16:43
  • \$\begingroup\$ Which of them? That is a datasheet for a whole capacitor series. You have not told us how much capacitance you need, or rated voltage, or ripple current you need, so how can we answer if it suits your purposed. \$\endgroup\$ – Justme Nov 1 '19 at 16:50
  • \$\begingroup\$ 470uF 50V, it says high impedance without mentioning ESR though... \$\endgroup\$ – ElectronSurf Nov 1 '19 at 17:15
  • \$\begingroup\$ Each capacitor has it's own line of data. You look for the line with 470uF 50V capacitor. Then you look at the ESR. And here's the easy part: then you look if that is compatible with switching regulator limits, and see that no, it is not compatible. \$\endgroup\$ – Justme Nov 1 '19 at 19:37

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.