Output Capacitor Selection for Buck Converter

[white_devil_123]

Hey, I am trying to design a 24V/8V 8A Switching Power Supply at a 1MHz Switching Frequency with LM61480 Switching Regulator.

Here is the Datasheet to the regulator: https://www.ti.com/lit/ds/symlink/lm61480-q1.pdf

I have a doubt about the output capacitor selection. Can I use the usual formula, which is, "Cout = (peak-to-peak output current ripple)*(switching period)/(maximum allowable voltage output ripple)" for this switching regulator, too?

I did go through the datasheet but still confused.

[tszaboo]

TI has an online calculator called Webench, where you can design your SMPS.
You can create a design on the product page:
https://www.ti.com/product/LM61480
Try it!

[mtwieg]

Output ripple voltage will depend more on the capacitor ESR and ESL than capacitance. Also the capacitors must be rated for that ripple current.

Electrolytic capacitors will not be effective at filtering 1MHz ripple, you will need ceramic capacitors (electrolytic capacitors may also be used, but won't do much for 1MHz ripple).

[Faringdon]

I once did a 1MHz vicor module PSU for a co....they wanted the Cout  with a couple of 4.7uF MLCC's...i changed it for a few 1uF MLCCs...it was something like a quarter of the capacitance...and the output voltage ripple went  DOWN!.
(Because ESL was bigger in the higher capactiance MLCCs).
When  i reported it to them..they said i must have done it wrong , and fired me ...but i was explaining the situation as we walked along the corridor to the exit door...and they decided to look into it.....they took me back on.

[uer166]

, and fired me ....

Are you 10000% sure it was because of that issue, and that issue only?

[David Hess]

Output ripple voltage will depend more on the capacitor ESR and ESL than capacitance. Also the capacitors must be rated for that ripple current.

I agree; the impedance will be dominated by the ESR and ESL which depend more on factors other than capacitance, including the package.  Even with electrolytic capacitors at high frequencies, it is about impedance rather than capacitance.

[Faringdon]

Yes.
Please find the attached LTspice sim.....make out that for your capacitor bank.

simsurfing.com (by Murata?) for MLCCs is a free program which shows these charts

[eTobey]

...and they decided to look into it.....they took me back on.

I guess they did give you a rise?

[AnalogTodd]

Hey, I am trying to design a 24V/8V 8A Switching Power Supply at a 1MHz Switching Frequency with LM61480 Switching Regulator.

Here is the Datasheet to the regulator: https://www.ti.com/lit/ds/symlink/lm61480-q1.pdf

I have a doubt about the output capacitor selection. Can I use the usual formula, which is, "Cout = (peak-to-peak output current ripple)*(switching period)/(maximum allowable voltage output ripple)" for this switching regulator, too?

I did go through the datasheet but still confused.

As many people have pointed out, you want to look at impedance of your output capacitor network at that frequency instead of capacitance. Bulk electrolytics have higher ESR that becomes dominant at around a few hundred kHz. If you get a capacitor with a typical ESR of 50 milliohms, that will mean your ripple will likely be ~400mV as you will be in the range where ESR dominates over capacitance and ESL has not yet kicked in.

Ceramics have much lower impedance in that range, typically showing an ESR of 5 milliohms and an ESL below 1nH. You get roughly a 1MHz resonant frequency with a 10uF ceramic. You can get lower ripple with a number of small capacitors because the ESR and ESL effects in parallel drop the net impedance. You do need to watch out that your loop response works appropriately (most of these will have unity gain crossover in the 10's of kHz), so the ceramics may just be for ripple while a bulk electrolytic works for loop stability. If you want to do pure ceramic, watch for voltage coefficient of the caps as well.