What factors should be considered when choosing decoupling capacitors?

Click For Summary

Discussion Overview

The discussion revolves around the selection of decoupling capacitors in power rail applications, focusing on the factors influencing their effectiveness in filtering noise at different frequencies. Participants explore the roles of capacitance values, impedance characteristics, and physical properties of capacitors, as well as the implications of using multiple capacitors in parallel.

Discussion Character

  • Technical explanation
  • Debate/contested
  • Mathematical reasoning

Main Points Raised

  • Some participants suggest that lower value capacitors filter high frequency noise while higher value capacitors filter low frequency noise, though this is not universally agreed upon.
  • Mathematical relationships are presented, specifically the formula Xc = 1/(2 * pi * f * c), indicating that lower capacitance results in higher impedance.
  • One participant questions the effectiveness of using large capacitance to achieve a "perfect short" to ground for noise, implying potential limitations.
  • Concerns are raised regarding the equivalent series inductance (ESL) and equivalent series resistance (ESR) of large electrolytic capacitors, which may not provide low AC impedance suitable for bypassing.
  • A participant emphasizes the complexity of using different value capacitors, noting that they can create both series and parallel resonances, which complicates the filtering process.
  • Discussion includes the recommendation of using power integrity tools to assess the impact of capacitor choices, considering additional factors like plane-to-plane capacitance and trace inductance.
  • Some participants reference external resources, although one participant critiques the accuracy of a provided link, suggesting that the topic is more intricate than presented.

Areas of Agreement / Disagreement

Participants express differing views on the roles of various capacitance values and the implications of using multiple capacitors. The discussion remains unresolved regarding the best practices for selecting decoupling capacitors.

Contextual Notes

Limitations include potential missing assumptions about the circuit design and the specific application context. The discussion does not resolve the complexities introduced by resonances and the interplay of various capacitor properties.

d.arbitman
Messages
100
Reaction score
4
I have seen power rails bypassed with parallel capacitors which are at least an order of magnitude apart in capacitance value. From what I understand, one capacitor filters high frequency noise and the other filters low frequency noise.
1. Which capacitor does what? The low value capacitor filters low frequency noise?
2. What physical property of a capacitor is responsible for the difference in frequency response?
 
Engineering news on Phys.org
Do the math. Xc = 1/(2 * pi * f * c)
-
So, a lower capacitance offers a higher impedance compared to a higher capacitance.
 
Averagesupernova said:
Do the math. Xc = 1/(2 * pi * f * c)
-
So, a lower capacitance offers a higher impedance compared to a higher capacitance.

Yes, so why not just put a huge capacitance to create a "perfect short" to ground for noise?
 
Large electrolytic capacitors may hold a very large charge but that is not to say that have a small enough ESL ( equivalent series inductance) or ESR (equivalent series resistance) to provide a low AC impedance that is suitable for bypass.
 
The interfacebus.com link is inaccurate, the situation is actually way more complicated.

Using different value capacitors does create series resonances at staggered frequencies (as shown in figure at interfacebus.com), but they also introduce nasty parallel resonances (see figure 5 in attached link).

This is a topic that created a great deal of discussion in our industry about a decade ago. If you want to see the real impact of your capacitor choices you will need to use a power integrity tool. These tools include important factors such as plane to plane capacitance, and inductance of traces that feed the capacitors.

I have seen boards that used half dozen different value decoupling capacitors, then hooked them up with 100mil long 10mil traces.
 

Attachments

Similar threads

  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 15 ·
Replies
15
Views
5K
  • · Replies 15 ·
Replies
15
Views
7K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 1 ·
Replies
1
Views
4K
  • · Replies 2 ·
Replies
2
Views
2K
Replies
3
Views
3K
Replies
1
Views
3K
  • · Replies 41 ·
2
Replies
41
Views
7K