Let's take a look and highlight what have been some of the most useful power tips, relevant for power-supply design engineers even to this day...
In July 2008, Robert Kollman started a column to share common tips and tricks of power-supply design. At the time, the Texas Instruments (TI) Power Design Services team had a large library of reference designs that addressed many of the most common power-supply-design challenges. It was time to start sharing this knowledge with the community, and our aptly-named “Power Tips” became an instant hit.
Here we are 12 years later, with a monthly column that has reached 100 articles! To commemorate this milestone, I want to take a look back and highlight what I find have been some of the most useful power tips, relevant for power-supply design engineers even to this day.
Our initial topics comprised hints and basic equations to help simplify the design process. In Power Tips #1, Robert discussed how to pick the right switching frequency for a power converter, covering the relationship between the power-supply size (volume) and the switching frequency. This is usually the first step in a power-supply design and has a ripple effect all the way through the entire design process. Figure 1 shows the volume of the major parts of a synchronous buck converter as the frequency increases. Even though this topic is 12 years old, it remains the golden rule of power-supply design. Higher frequency equals smaller size.
What was interesting is that he didn’t stop with that point, publishing further discussions about which components start to dominate the size as frequency increases. There is also an effect on the design of the best semiconductor die size based on the frequency of operation, so choosing the right part is crucial.
Power Tips #10, recapped in this training video, is one of my favorites. I still use this technique today when selecting output capacitors for a design. This article explains how to simply estimate your transient response. A major challenge for power-supply designers is figuring out how much output capacitance is necessary based on changing load conditions. A quick way to get to a first-level approximation is to use the closed-loop output impedance and the load step variation.
If you know the maximum allowable output voltage deviation, you can simply take that voltage and divide it by the load current step to calculate the maximum converter output impedance. You can then use the bandwidth of the converter to determine the necessary ceramic output capacitance. You’ll need to account for the equivalent series resistance (ESR) of the capacitor, but it should be very simple to conduct a quick simulation to verify the results.
Power Tips #49, #50, and #51 explored the effects of different capacitor parasitics. For an engineer who designs power supplies every day, some of these may seem obvious, but for those who are part-time designers or new to the industry, these tips are gold. Power Tips #49 covers multilayer ceramic capacitors (MLCCs) and three things to keep in mind when selecting this type of capacitor. The first is understanding the three-digit code (for example, X7R, Y5V, and X6S) associated with the temperature and tolerance. Figure 2 shows an example of how the code can be broken down. These temperature ratings and tolerances are very important when selecting a capacitor. While the purchasing department might like the price of Y5V caps, trying to keep your loop stable with a +80 to -20% difference in capacitance will be a challenge.
Figure 2 Ceramic capacitor temperature and tolerance can be correlated to a three-digit code.
The second thing to keep in mind is the effect of DC bias. Capacitors have a maximum operating voltage and an initial capacitance. With MLCCs, as the voltage across the capacitor increases, the effective capacitance goes down. Sometimes this drop can be very significant, as much as 70 to 80%, so your 10-μF capacitor is actually only 3 μF!
The third pitfall with MLCCs has to do with reliability and cracking. Because the capacitors are made of many layers stacked and pressed together, there is a chance that the capacitor could crack as the circuit board moves or flexes. Because of this issue, power supply designers usually limit the case size of ceramic capacitors to a 1210 package.
Power Tips #50 covers electrolytic capacitors and some important factors to consider when using them in a power supply. For example, in applications designed to have long lifetimes, usually the first component to fail will be an aluminum electrolytic capacitor. Aluminum capacitors are rated for a maximum temperature and number of hours. A typical rating might be 1,000 hours at 105°C. For every 10°C decrease in temperature, the life expectancy doubles. If a product like an LED light bulb is supposed to last for 25 years (on for 4 hours a day), it needs to be able to operate for ~36,000 hours. That means that the capacitor needs to operate at a temperature less than 55°C, a tough task for something enclosed like a light bulb. The second major issue with aluminum capacitors is the variation of the ESR with temperature. The ESR of an aluminum capacitor could change by a factor of 20 across the operating conditions. This could create loop instabilities depending on the operating environment of the power supply.
Power Tips #51 examines different types of capacitors and how the parasitics affect voltage ripple. It is important to understand the type of capacitor used and how it will determine the voltage ripple. Capacitors not only have a capacitance associated with them, but also ESR and equivalent series inductance (ESL). Each of these parasitic elements contributes to a different component of the voltage ripple shape. Figure 3 shows the resultant waveform in a buck converter for each of these parasitic elements. The type of capacitor will have a large impact on how these parameters behave. As the switching frequencies of power converters increase, the parasitics start to play a larger role in the overall design.
Figure 3 Parasitic elements of a capacitor cause waveforms.
In August 2014, Robert Kollman wrote his last power tip before retiring. The series took a hiatus for a couple of years, but came back in the summer of 2016, with a number of engineers on TI’s Power Design Services team taking turns publishing the tips. As we climbed toward 100, we added a number of useful entries that built on the fundamentals of the original series.
In Power Tips #81, we covered the proper biasing of an optocoupler. For anyone who has tried to compensate an isolated power supply with opto feedback, this tip is very helpful. In Power Tips #89 and #92, a discussion of design considerations for high-frequency resonant converters provides a good base with which to get started on these very complicated designs. And for anyone trying to design a high-conversion ratio boost, Power Tips #90 shows a clever approach using a tapped inductor to provide more boost.
From all of us on the Power Design Services team at TI, we’re honored to reach the milestone of 100 Power Tips, and looking forward to the next 100. If you would like to see us cover something specific or share your favorite power tip, post a comment below.
Robert Taylor is an Applications Manager at Texas Instruments.