James explores whether multilayer ceramic capacitors (MLCCs) can be replaced with polymer electrolytic capacitors, specifically KO-Caps, in a DC-to-DC converter. He compares the performance of both capacitor types, discussing the MLCC's DC Bias Effect, which reduces capacitance under load, and how KO-Caps avoid this issue. Through a series of tests on an evaluation board, James measures ripple and capacitance in various configurations and highlights the trade-offs between MLCCs and polymers in terms of performance, cost, and board space. Ultimately, the video offers insights into when it might make sense to switch from ceramics to polymers in power conversion designs.
Watch the Video to find out!
In this video, James explores whether multilayer ceramic capacitors (MLCCs) can be replaced with polymer electrolytic capacitors with the help of a DC-DC converter. He begins by discussing the characteristics of MLCCs and polymers. Using an off-the-shelf DC-to-DC converter evaluation board as the test setup, James demonstrates how the performance of both types of capacitors can be evaluated.
James explains that MLCCs comprise alternating layers of dielectric material (Barium Titanate) and metal (Nickel), forming tiny capacitors in parallel. The multiple layers give the overall device a relatively high capacitance. One of the advantages of MLCCs is their low equivalent series resistance (ESR), which is attributed to the parallel metal layers. However, he notes that class 2 and 3 MLCCs, such as those with X7R or Y5V ratings, experience a significant loss in capacitance when exposed to DC voltage, an effect commonly known as the "DC Bias Effect." He illustrates this effect by showing measurements of a 22-microfarad MLCC rated for 10 volts. When 50% of the rated voltage is applied, the capacitance drops to 6.7 microfarads, and at full voltage, the capacitance is reduced by over 85%!
In contrast, polymer electrolytic capacitors, or KO-Caps, as KEMET calls them, Are made with a tantalum anode, tantalum pentoxide dielectric, and a conductive polymer cathode (PEDOT). They do not suffer from the same DC Bias Effect. These capacitors offer high capacitance and low ESR without the voltage coefficient, which makes them attractive as replacements for MLCCs in some applications.
James provides a detailed overview of the test setup, which involves an evaluation board used in a DC-to-DC converter experiment. The capacitors are strategically positioned near the MOSFETs and output terminals. The equipment used in the test includes a low-noise oscilloscope, electronic load, linear power supply, and 10:1 oscilloscope probes.
James begins by testing the original configuration of four 100-microfarad MLCCs and measuring the ripple at different current loads. He finds that the ripple ranges from 132 to 151 millivolts, which becomes the baseline for the rest of the comparisons.
Next, he swaps the MLCCs for KEMET's KO-Caps, keeping the rated capacitance precisely the same and the physical size roughly the same as the originals. Running the same previous tests, James finds that the ripple in the polymer setup is only a few percent higher than the MLCC configuration, suggesting no immediate advantage in replacing MLCCs with KO-Caps.
However, James points out a critical difference: the MLCC configuration includes polymers on the back side of the board, near the output terminals, while the KO-Cap configuration does not. This subtle difference, he argues, could explain why the KO-Caps performed so similarly to the MLCCs.
James removed the bulk polymer capacitors from the MLCC setup to investigate further and reran the tests. Without these bulk polymers, the ripple skyrockets to several volts, demonstrating that the MLCCs alone are insufficient to stabilize the power supply. He concludes that while high-capacitance MLCCs can offer low ripple, they are not a standalone solution in this specific setup.
James then compares the cost of the different configurations. He notes that the price of high-capacitance ceramics is often comparable to that of polymer capacitors. However, the KO-Cap solution, despite having a (very) slightly higher ripple, costs nearly 60% less than the ceramic plus polymer solution!
Additionally, when James adds two bulk 470-microfarad polymers to the KO-Cap setup, the ripple drops by 45%. The total solution is still cheaper than the original MLCC-plus-polymer setup. This finding suggests that KO-Caps can provide a cost-effective alternative in some cases.
In the concluding section, James emphasizes that while polymer capacitors may offer advantages in specific applications, they come with trade-offs. MLCCs must be carefully evaluated for their DC Bias Effect, as their effective capacitance can drop significantly in real-world applications. Additionally, polymer capacitors tend to have larger footprints compared to MLCCs, so you should consider board space when making their selection. He also warns that polymers behave differently at higher frequencies and that their ESR changes as the frequency increases.
Overall, James's evaluation shows that while MLCCs and polymers have their respective strengths and weaknesses, polymers can serve as a viable replacement in specific designs, particularly when considering cost and board space constraints. The key takeaway is that you must carefully consider their requirements, such as ripple tolerance, size, and cost, when choosing between MLCCs and polymer capacitors.
Downloads and Links
- Learning About Polymer Capacitors -- The Learning Circuit 40
- Ceramic Capacitor Voltage Effect - Workbench Wednesdays 30
- Kemet KSIM
- WBW88 Additional Data - Screenshots and detailed outputs
Bill of Materials
| Product Name | Manufacturer | Quantity | Buy Kit |
|---|---|---|---|
| C1206C107M9PACTU SMD Multilayer Ceramic Capacitor, 100 µF, 6.3 V, 1206 [3216 Metric], ± 20%, X5R | YAGEO (KEMET) | 1 | Buy Now |
| T520A107M006ATE025 Tantalum Polymer Capacitor, 100 µF, ± 20%, 6.3 V, A, 0.025 ohm, 1206 [3216 Metric] | YAGEO (KEMET) | 1 | Buy Now |
