Introduction to Decoupling Capacitors
The DENPAFLUX Editorial Team is a group of writers, researchers, content creators and industry experts dedicated to delivering cutting-edge content in the field of electromagnetic compatibility (EMC) and related technologies. With a diverse range of backgrounds and specializations, our team collaborates to bring you insightful articles, in-depth analysis, and practical guidance on EMC topics and electronics industry
Decoupling capacitors, also known as bypass capacitors, are some of the simplest and most inexpensive components on a typical PCB. And yet, they are absolutely crucial: if omitted, the board may fail outright; if not properly implemented, the board may suffer from intermittent failures or impaired performance. One of the best ways to ensure that your board is resistant to electromagnetic interference (EMI) and prepared for electromagnetic compliance (EMC) is to learn about decoupling-capacitor best practices, so that you can optimize the decoupling techniques in your design.
The Need for Decoupling
One of the first steps in becoming a real-word PCB designer is remembering that every integrated circuit on your board needs at least one 0.1 µF ceramic capacitor between power and ground. This rule of thumb works surprisingly well in low-speed systems that don’t have demanding noise requirements and aren’t subject to serious voltage disturbances. But as clock frequencies increase, noise sources multiply, designs become more complex, and operating environments become more severe, there is greater need for informed, meticulous decoupling practices rather than rules of thumb.
Digital ICs always need decoupling capacitors because digital circuits create transient disturbances when they switch states. This is the fundamental problem that decoupling is intended to solve. When a digital circuit experiences an output transition, a quick burst of current is generated, and that quick burst of current is converted into a voltage spike when it flows through the output impedance of a power supply and through the parasitic inductance of wires and traces. Though purely analog ICs lack switching circuits and therefore don’t require these short bursts of current, they still need decoupling capacitors to protect them from noise that is generated by other components on the board and from conducted or radiated emissions.
All these spikes, along with the high-frequency harmonic energy that they create, can seriously degrade the power-supply voltages on a PCB. As power integrity diminishes, we see reductions also in the quality of nearby signals and in the performance of components that depend on this power-supply voltage. Analog waveforms become noisier, digital circuits become less reliable, and the board produces more EMI.
The Decoupling-Cap Solution
Decoupling capacitors prevent these harmful effects by functioning as low-resistance, low-inductance reservoirs of electric charge that can source or sink the bursts of current without creating voltage spikes. You can think about the process like this: When the board is first powered up, the supply voltage delivers the current required by the IC and charges up the decoupling capacitor. That capacitor now holds a certain amount of electric charge and acts like a very small, highly responsive battery located right next to the IC’s power pin. When digital circuitry inside the IC starts switching, the capacitor has the electric charge needed to supply the switching current, and it delivers this charge so efficiently that no serious voltage spikes are produced.
These noise-reducing, power-integrity-preserving components can be called either decoupling capacitors or bypass capacitors. A decoupling capacitor helps to make components or subcircuits more robust by decoupling them from other portions of the board, which might be generating noise or interference. A bypass capacitor helps noise to bypass the IC—in other words, noise flows through the capacitor instead of the IC because the capacitor provides a low-impedance path to ground. Both names make sense and suggest valid ways of thinking about the functionality of a decoupling capacitor.
Choosing Decoupling Capacitors
In general, the exact value of a decoupling capacitor is not critical, because standard choices such as 0.1 µF can hold more than enough charge for typical current spikes. The more important issue is how effectively a particular capacitor can bypass lower or higher frequencies. The basic relationship is that smaller capacitors are better for high-frequency noise, and larger capacitors are better for low-frequency noise.
This is why decoupling networks often include two capacitors: something in the 0.1 µF range (with a package size of 0805 or smaller) deals with higher-frequency disturbances, and something in the 10 µF range deals with lower-frequency power-supply variations. This is a good starting point for most designs, but lower-value capacitors—perhaps 0.01 µF, in a small package such as 0402—should be considered in high-frequency applications. Determining an optimal decoupling network for a particular IC is usually not something that will fit within a design team’s budget or schedule, but the IC’s manufacturer has more resources to invest in this task, so always check the datasheet for decoupling recommendations.
In addition to capacitance value, designers need to consider capacitor type and voltage ratings. For low-capacitance decoupling, ceramic is the best and most widely used capacitor type. For higher values, such as 10 µF, ceramic or tantalum can be used. The voltage rating should exceed the maximum expected circuit voltage by a safety margin of at least 20–30%; 50% is better if design constraints allow for it.
Layout Practices for Decoupling Capacitors
The key principle to keep in mind when laying out a decoupling network is the need for rapid bursts of current. Decoupling capacitors accomplish their task because they freely supply small bursts of charge to the IC, and the layout needs to help them do this by minimizing resistance and inductance.
What this means in practice is that decoupling capacitors should be as close as possible to their corresponding power pins and connected via traces that are short and reasonably wide. The diagram below shows the ideal case for a basic IC with one supply-voltage pin and one ground pin:
- The capacitor is very close to the VDD pin and connected to it by a short, wide trace.
- The other traces are also short and wide.
- The components receive power and ground through vias connected to internal plane layers.
Optimizing Your Decoupling Designs
If your decoupling issues are an obstacle to reliable performance and EMC certification, DENPAFLUX can help. Our AI-powered PCB-analysis software can automatically identify areas of concern, and we also provide a full range of EMC consulting services. Our experts can give you the personalized engineering support you need to overcome power-integrity challenges and achieve EMC compliance. Our goal is to make your EMC problems disappear—give us a call today, and let’s talk about how we can make that happen.