Author Topic: Using an MLCC above its self resonant frequency for decoupling  (Read 1372 times)

0 Members and 1 Guest are viewing this topic.

Offline hazukiTopic starter

  • Contributor
  • Posts: 39
I'm looking at using a 3MHz switcher, Analog Devices' ADP2108 https://www.analog.com/media/en/technical-documentation/data-sheets/ADP2108.pdf

In the datasheet they recommend some output capacitors on page 14. These are 10uF 0603s. It just so happens that at 3MHz these capacitors are past their self resonant frequency and appear more like inductors! Check out the datasheet of one of the recommended caps here: https://product.tdk.com/en/search/capacitor/ceramic/mlcc/info?part_no=C1608JB0J106K080AB

What would happen if I use one of the caps they recommend? Since we are above the SRF it no longer makes sense to talk about the MLCCs having a capacitance at these frequencies. These should still filter out the switching noise though since it's really the impedance at frequency that matters. Does this all sound correct?

Thanks for the help.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Using an MLCC above its self resonant frequency for decoupling
« Reply #1 on: May 16, 2020, 12:08:48 am »
What do you think will happen?  How could you figure it out?

SPICE simulation is a good exercise here. :)

Keep in mind the SRF isn't a very meaningful quantity, as it will be considerably lower in a real circuit that has trace lengths.  You must approximate these impedances as well.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 7454
  • Country: pl
Re: Using an MLCC above its self resonant frequency for decoupling
« Reply #2 on: May 16, 2020, 07:43:37 am »
It literally means that the voltage drop required to change the direction of current through the capacitor will be larger than the normal ripple voltage. And that their phase relation will be different.
« Last Edit: May 16, 2020, 07:45:09 am by magic »
 

Offline KE5FX

  • Super Contributor
  • ***
  • Posts: 2096
  • Country: us
    • KE5FX.COM
Re: Using an MLCC above its self resonant frequency for decoupling
« Reply #3 on: May 16, 2020, 08:08:14 am »
They appear like really awful inductors, which means they still exhibit a usefully low impedance well beyond their SRF.  (Notice how SRF curves almost always appear on log-X plots.)

It's more instructive to think about the capacitors' physical construction than it is to run them through SPICE and call it a day.  An MLCC is just that -- a multilayered construction in which numerous smaller capacitors, each with a somewhat-higher individual SRF, are effectively paralleled.  A SPICE model that looks like a capacitor in series with an inductor and a resistor isn't entirely adequate to capture all of the subtleties.
 
The following users thanked this post: cdev, hazuki, ogden, SiliconWizard

Offline AndyC_772

  • Super Contributor
  • ***
  • Posts: 4315
  • Country: gb
  • Professional design engineer
    • Cawte Engineering | Reliable Electronics
Re: Using an MLCC above its self resonant frequency for decoupling
« Reply #4 on: May 16, 2020, 08:17:57 am »
You might find this article useful.

 
The following users thanked this post: hazuki

Offline chuckb

  • Frequent Contributor
  • **
  • Posts: 345
  • Country: us
Re: Using an MLCC above its self resonant frequency for decoupling
« Reply #5 on: May 16, 2020, 03:14:16 pm »
Capacitor impedance plots are drawn with a pure resistive source. They are not attenuation plots. In the real world with source resistance and inductance you can develop attenuation plots that are totally different and better! This is a good SPICE exercise.

Filtering capacitors work fine above the SRF as long as the rest of the circuit is designed correctly. In the output stage of a switching power supply you have say a 1000nH source impedance. The filter capacitor has an inductance of say 1nH and some capacitance. Above the SRF the output filter will look like a capacitively coupled inductive divider. With low cap ESR you will have a 1000:1 divider.

This is also why you want to extract your load current directly from the capacitor terminals.
« Last Edit: May 16, 2020, 07:36:58 pm by chuckb »
 
The following users thanked this post: hazuki

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: Using an MLCC above its self resonant frequency for decoupling
« Reply #6 on: May 16, 2020, 03:25:50 pm »
In the datasheet they recommend some output capacitors on page 14. These are 10uF 0603s. It just so happens that at 3MHz these capacitors are past their self resonant frequency and appear more like inductors!
It is working indeed, otherwise they would not sell eval board. Main question here - will it's performance fit your requirements like output ripple, conducted and radiated EMI. I would either buy eval board or copy-paste it, test and see what needs to be improved, if anything.
 
The following users thanked this post: hazuki


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf