Electrolytic capacitors are widely used in audio components for filtering, anti-coupling, high-frequency compensation, DC feedback, and more. However, a detailed explanation based on function, materials, and manufacturing methods is not easily achievable. Therefore, this article will focus solely on electrolytic capacitors, specifically aluminum electrolytic capacitors used for power smoothing and filtering. Every audio machine requires a power supply—except for passive preamplifiers. Since power is needed, filtering is essential. Don't argue that battery power eliminates the need for power smoothing and filtering. However, battery charging circuits also involve rectification and filtering, thus requiring filter capacitors. The formal name for the filter capacitors we commonly use today is aluminum foil dry electrolytic capacitor. From my observation, except for the Canadian Sonic Frontiers vacuum tube preamplifier, which used PP plastic capacitors for filtering in its high-voltage regulation circuit, all other models use aluminum foil dry electrolytic capacitors; therefore, it's necessary for readers to learn more about them. When you think of capacitors used for power smoothing and filtering in a power regulation circuit, what comes to mind first? —Capacitance? Voltage Rating? A capacitor's package always has a capacitance marking, referring to its electrostatic capacitance; it also always has a voltage rating marking, referring to its operating voltage or rated voltage. The operating voltage (WV) is the absolute safe value; the surge voltage (SV or Vs) is the surge voltage or breakdown voltage; exceeding this voltage will ensure the capacitor is overwhelmed by surges—be careful, the capacitor might explode! According to the international standard IEC 384-4, below 315V, Vs = 1.15 × Vr; above 315V, Vs = 1.1 × Vr. Vs is the surge voltage, and Vr is the rated voltage. The charge energy of a capacitor is expressed as Q = CV, where Q is coulombs, C is capacitance, and V is voltage; therefore, when the voltage remains constant, increasing the capacitance increases the charge energy. Please note that the unit of capacitor capacitance should be F (farad), but due to the measurement being too high, the value is often underestimated, so μF is commonly used instead, 1F = one million μF. In other countries, mF is sometimes used to represent μF. While mF isn't entirely accurate, mechanical typewriters don't have a μ key, so m represents micro. Having considered capacitance and operating voltage, what's the next thing you'd consider when buying capacitors? Intuitively, price. Yes, this parameter is important, and the lower the value, the better. Some people think of brand first and insist on never using Japanese products—is there some lingering sentimentality? American products are only second; Swedish or German-made capacitors are number one. Yes, this parameter is also important. But since we're talking about brands, we can't ignore the series and model; because a manufacturer produces many different series of products, and different series have different quality and prices. Okay, let's summarize the known parameters for power supply smoothing filter capacitors: capacitance, rated operating voltage, surge voltage, price, brand, and model series. There shouldn't be only a few; size is also important, because it relates to weight and lead type. Snap-in is for soldering onto PCBs, and screw is for screw-on. Regarding weight, two capacitors with the same capacitance and voltage rating but from different brands will definitely weigh differently; and the external dimensions are also related to the chassis design. Some capacitors are not perfectly round, but rather polygonal; Philips and BHC both have this type that looks quite high-end. Now let's summarize again, adding weight, external dimensions, and lead type—that makes nine parameters. Outer casing color? Who suggested this? Brilliant. Because white, black, and blue plastic packages are used by various manufacturers, it sometimes has certain meanings; for example, black on a gold background in Japanese specifications often indicates high-end audio capacitors. What else can you think of based on appearance alone? Manufacturing date; 9627 means it was manufactured in the 27th week of 1996; in recent years, Japanese capacitor manufacturers seem to have gradually omitted the manufacturing date marking. However, outer casing color and text printing are not directly related to quality, so only the manufacturing date parameter is included. Also, don't forget the applicable operating temperature, because 105°C is more suitable for vacuum tube amplifiers than 85°C. If the machine is to be placed in Antarctica, it's best to choose a type that can withstand -55°C. Don't overlook capacitance tolerance either. When using multiple capacitors in parallel, lower tolerance is always better to ensure uniform characteristics in each individual capacitor. Now, adding operating temperature and capacitance tolerance, we have 12 parameters, giving us at least a 30% understanding of capacitors. Please don't misunderstand; the capacitor's operating temperature doesn't refer to ambient or surface temperature—regardless of the degree, the plastic casing remains the same. It refers to the aluminum foil's operating temperature. Therefore, using an 85°C capacitor in the mounting machine is perfectly fine, as long as the capacitor is kept away from the tubes, it's safe. However, several crucial parameters related to capacitor quality only exist in the manufacturer's specifications and are never shown on the finished product's casing. These are the key parameters discussed in this article. Dissipation Factor – Loss Angle The dissipation factor (DF) exists in all capacitors, sometimes expressed as the loss angle tanδ. Think about it: a loss angle implies loss, so lower is always better. Plastic capacitors have very low loss angles, but aluminum electrolytic capacitors have considerably high ones. Whether a capacitor's dissipation factor (DF) value is high or low depends on factors such as temperature, capacitance, voltage, and frequency, even for capacitors of the same brand and series. When capacitance is the same, a higher voltage rating generally results in a lower DF value. For example, a 10000μF capacitor from the same brand and series with an 80V voltage rating will have a lower DF value than one with a 63V voltage rating. This explains why this publication often selects filter capacitors with higher voltage ratings. Furthermore, higher temperatures and frequencies also generally result in higher DF values. However, many capacitor manufacturers do not specify the DF value in their specifications because a high DF value is difficult to read. For instance, the blue PHE-420 series from RIFA (Sweden) is an MKP plastic capacitor with a DF value ranging from a minimum of 0.00005 to a maximum of 0.0008. However, the top-tier white PEH169 series aluminum electrolytic capacitors do not specify the DF value. If the DF value were specified, it would likely be 1.0000, with the decimal point after the 1. Leakage...leakage current! Leakage! Ideally, there wouldn't be any. But unfortunately, aluminum electrolytic capacitors will inevitably produce leakage current during operation. Leakage current should be low, and its calculation formula is roughly: I = K × CV. The unit of leakage current I is μA, and K is a constant, such as 0.01 or 0.03; each manufacturer will choose a different constant. However, regardless of the formula, the higher the capacitor's capacitance, the greater the leakage current. If you believe that a larger capacitance will result in better smoothing, please also consider this "leakage current." The calculation shows that the higher the rated voltage, the greater the leakage current; therefore, lowering the operating voltage can also reduce the leakage current. However, reducing capacitor leakage current is not easy. Low leakage current (LL series) capacitors are expensive. I once ordered a batch of low leakage current (LL series) capacitors from a domestic manufacturer, and the price was higher than many imported capacitors. In terms of leakage current specifications, aluminum electrolytic capacitors are much worse than tantalum electrolytic capacitors. Tantalum capacitors also come in dry and wet types, but their capacitance and voltage rating are lower. Aside from custom orders, for standard capacitors, reducing leakage current can be achieved by increasing the ratio of Vs to Vr. Vs is the surge voltage, which is naturally higher than the rated voltage Vr, but the applied voltage (the actual operating voltage) should be lower than Vr, for example, 90% of Vr; choosing high-voltage capacitors is entirely correct. Equivalent series resistance (ESR): A capacitor will have various impedances and inductive reactances due to its construction. The most important are ESR and ESL—this is the basis of capacitive reactance. Why would a capacitor need resistance when it provides capacitance? Therefore, ESR and ESL should be low; however, low ESR/low ESL are usually high-end series. The level of ESR is related to the capacitor's capacitance, voltage, frequency, and temperature. When the rated voltage is fixed, the larger the capacitance, the lower the ESR. Some people habitually connect multiple small capacitors in parallel to form a large capacitor to reduce impedance, theoretically because parallel resistors reduce resistance. However, considering the impedance of the capacitor leads, connecting small capacitors in parallel to a large one may not necessarily yield results. Conversely, when the capacitance is fixed, choosing a capacitor with a higher WV rated voltage can also reduce ESR; therefore, higher voltage ratings do indeed have many advantages. Frequency influence: ESR is higher at low frequencies and lower at high frequencies; of course, high temperatures also increase ESR. The unit for series equivalent resistance (ESR) is mΩ. High-end capacitors often use "low ESR" and "low ESL". Comparing low internal resistance and low leakage current, low internal resistance is easier to achieve, hence capacitors labeled "low ESR" are quite common. ESR is related to the loss angle: ESR = tanδ/(ω×Cs), where Cs is the capacitance. Sometimes capacitor specifications include "Z," which has a different meaning from ESR, but Z is related to ESR and also considers capacitive and inductive reactance; it represents the true internal resistance. The unit for capacitor ESR mentioned earlier is mΩ, which refers to large capacitors. For a small capacitor like 220μF, the ESR unit is Ω, not mΩ. Which type of capacitor has the lowest ESR? There's only one answer: Sanyo's OS organic semiconductor capacitors! The previously discussed dissipation factor (DF - loss angle tanδ), leakage current, and ESR (equivalent series resistance) are all better the lower they are. However, the ripple current, which we will discuss now, is better the higher it is. Especially now that power amplifiers are required to have high current output, the ripple current (Irac or Iac) of the power supply smoothing filter capacitor becomes particularly important. The ripple current (Irac) marking should at least have two specifications: one for low-frequency operation and one for high-frequency operation. The low-frequency standard is approximately 120Hz, and the high-frequency standard is approximately 10kHz, although there may be slight differences between different manufacturers. Ripple current is directly proportional to frequency, so the ripple current is lower at low frequencies. However, for audiophiles, the Irac value in the low-frequency range is more important. Therefore, the ripple current number is an extremely important factor when purchasing capacitors. Generally, within the same brand, screw-type capacitors usually have a higher ripple current than snap-in PCB capacitors. There was once a saying that RIFA's 10000μF capacitors were equivalent to other brands' 15000μF capacitors, because most Japanese capacitors have low ripple current, while RIFA's is exceptionally high, making it seem like one capacitor can do the work of two. German Siemens and British BHC capacitors also often outperform Japanese products in this Irac characteristic. To my knowledge, the largest Irac capacitor is the Siemens SIKOREL series, reaching a whopping 20A at 6800μF/63V! For smaller capacitance capacitors, the largest Irac capacitor is the Sanyo OS capacitor. Regarding the operation of power amplifiers, many people believe that low frequencies draw a lot of current. Here's a method to try: use the lowest DCV setting on a multimeter (preferably an analog meter) to measure the voltage drop across any emitter resistor. Play a record, turn up the preamp volume, and observe the meter's movement. You'll find that low frequencies do draw current, and even playing four guitars simultaneously draws a lot of current! What kind of music is best suited for a run-in power amplifier? Holst's *The Planets*, First Movement, *Mars*. Now you should understand more than 60% of it. Perhaps you're wondering: are there aluminum electrolytic capacitors that are small, have low leakage current, low ESR, low tanδ, low error, and low price, but high ripple current and a wide operating temperature range? Well… no! Regarding capacitance error, aluminum electrolytic capacitors have made considerable progress in recent years. Previously, it was -20% to +40%, but now it's mostly +/-20%. However, their capacitance often leans towards + rather than -, so a 10000μF measurement might be close to 12000μF. Accurately measuring the capacitance of large-capacity capacitors is something I've wanted to do for many years. Don't doubt it, such testing instruments are hard to find. The US once manufactured one that could measure up to 99999μF and simultaneously display DF and ESR values; moreover, the capacitance was the average of tests at three frequencies (not two) of 100Hz, 1kHz, and 10kHz. This instrument once appeared on the domestic market, selling for as little as NT$100,000—only lacking leakage current testing capabilities. My standard for the safety margin of rated operating voltage is at least 15%. For example, if a capacitor's rated voltage is 50V, although the surge voltage might be as high as 63V, I will only apply a maximum of 42V. Allowing the capacitor's rated voltage to have a larger margin reduces internal resistance, leakage current, loss angle, and increases lifespan—a win-win situation. I once saw a Japanese amplifier with a ±48V operating voltage paired with a 10000μF/50V filter capacitor; it wouldn't burn out in the short term, but over time, its lifespan might decrease, requiring replacement or a new amplifier. Therefore, Japanese products often have the fate of "it's time to go," and you can't accuse them of cutting corners; after all, businesses need to make a profit. If they could only sell to you once in their lifetime, how would they make money? Does higher capacitance mean lower hum? When assembling your own system, the most annoying thing is the persistent hum. Some people increase the filter capacitor, and the hum disappears. I'm not entirely convinced, as amplifier hum is usually caused by improper grounding, with minimal impact from capacitors. However, theoretically, higher capacitance results in better power smoothing, so many designers and DIY enthusiasts firmly believe in the importance of large capacitance. Consequently, many power amplifiers, especially American brands like Krell and Mark Levinson, favor large capacitors; Dynaudio from Denmark even uses tens of thousands of μF capacitors in their preamplifiers. AC and DC amplifiers also tend towards "large capacitance," but within reasonable limits. However, many renowned manufacturers opt for lower capacitance. For example, Amcron in the US has a 250W x 2 professional power amplifier, totaling 500W for both channels, using only two small 8200μF filter capacitors (which seem a bit small). Swiss brand Goldmund, considered a Hi-End brand, has sent its products to various magazines for auditions, and no reviewer dared to say it was bad; its large power amplifiers use small capacitors. Swiss FM Acoustics are ridiculously expensive; a single stereo power amplifier can cost as much as a Mercedes-Benz. Their 220W x 2 professional power amplifier boasts tens of amps of current output, and I personally witnessed it using only two 10000μF/100V filter capacitors. Large-capacity filtering and low-capacity filtering are essentially opposing theories, yet both exist in the audio industry. An amplifier designed with low-capacity capacitors can be completely hum-free, and its low-frequency performance is no worse than a "water tank" amplifier. The key is Irac ripple current. If you're still obsessed with large-capacity capacitors, you don't understand electrolytic capacitors! Here's a suggestion: when assembling a power amplifier using low-capacity filter capacitors, always use a high-power transformer. In other words, "thin capacitors, fat transformer"—this might be the secret to a good amplifier sound. Based on detailed observations over the past few years, using a high-power transformer is far more effective than using large-capacity filter capacitors for a good power amplifier sound. One large one? Several small ones? Okay, some people are still worried and insist on large μF filter capacitors—should they find one large one, or use ten or so small ones in parallel? Others say that using small capacitors in parallel not only reduces internal resistance but also increases response speed, transparency, and resolution. Mark Levinson and Krell power amplifiers don't use small capacitors in parallel with large ones, but does anyone think they have slow response speed or are opaque and foggy? I myself have been confused about this issue for a long time. From a chassis design perspective, using multiple small capacitors in parallel seems more ideal, and the price is cheaper with larger quantities. Even preamps, power amps, and integrated amplifiers can use the same type of capacitor. Imported and domestic amplifiers have different fates. When consumers see imported amplifiers costing hundreds of thousands of yuan using multiple small capacitors, they will explain to themselves: this makes perfect sense; but when facing domestic products, they may have a different, malicious argument: shoddy workmanship! In terms of sound quality, whether it's a large or small "watery" sound, or one large or multiple small capacitors, shouldn't be absolutely related. As Deng Xiaoping aptly put it, "It doesn't matter whether a cat is black or white, as long as it catches mice, it's a good cat." The manufacturer's brand also influences quality; as mentioned earlier, some people never use Japanese products in their entire lives. The US originally had two major capacitor brands, Mallory and Sprague. Now, Sprague is extinct because it was acquired by the Japanese company Nippon Chemi-Con, and the company name is registered as United Chemi-Con/abbreviated as UCC. However, as long as it's still manufactured in the US and the outer casing is marked "Made in USA," the trademark change should have no relation to manufacturing quality. However, there are rumors circulating: UCC is inferior to Sprague. How likely is this? Once a Japanese trading company takes over, marketing policies will naturally change drastically. To increase sales volume, prices must be lowered; but a decline in price will also lead to a decline in quality. Inquiries with the local distributor, Ruipu Company, revealed that UCC capacitor sales are lower than Sprague's, indicating that domestic manufacturers are rejecting UCC. Comparing the specifications of UCC and Sprague capacitors, they are indeed very Japanese – the size is significantly reduced, from 40mm x 80mm to 40mm x 50mm. The price may be lower, but ESR increases and Irac decreases – how can one not sigh? Have concerns about Japanese products? There's no way around it. Not only the US, but Germany also needs Japanese investment for German-Japanese cooperation. Siemens and Matsushita jointly produce S+M capacitors. This is the future trend, almost inevitable. RIFA was also acquired by EVOX long ago. EVOX is a large conglomerate with factories everywhere. The 1μF capacitor used in our SigEnd single-ended preamplifier is an EVOX brand capacitor; although imported from the US, it looks like a Taiwanese product. Storage and Operating Life Compared to semi-permanent components like resistors, ICs, transistors, and plastic capacitors, the lifespan of aluminum electrolytic capacitors is worth noting. Firstly, storage lifespan is naturally related to lifespan; 10-20 years should be no problem. Capacitors that have been stored for too long should not be used immediately. Use a power supply to "age" them first; clamp the terminals and slowly adjust the power supply voltage, from low to high, up to the capacitor's rated voltage. The working life is difficult to define precisely. A so-called long-life (LL) capacitor usually indicates stable ripple current (Irac). As mentioned earlier, a capacitor's Irac is related to operating temperature and frequency. For example, at 10kHz, it's 15A at 40°C and 9A at 85°C; 15A/9A = 1.67. This number is the capacitor's life factor (I came up with this on the spot). The higher the number, the shorter the life; the closer the number is to 1, the longer the life. If I remember correctly, 1.93 represents 100,000 hours, 1.85 represents 200,000 hours, so 1.67 would be at least 500,000 hours! However, the primary function of a capacitor is charging and discharging, so frequent rapid charging and discharging is not advisable. There are two ways to effectively extend capacitor life: one is to reduce the number of times the power is turned on and off, and the other is to try to reduce the instantaneous charging current when the power is turned on—do you understand? This publication has also noticed this issue, and has therefore maintained this practice for many years. Even so, it's difficult to answer the question: which type of capacitor has better sound quality? Basically, different brands and series of capacitors naturally produce different sound performances. Personally, I don't reject Japanese products outright; with proper handling, Japanese products can be just as good as European and American ones. Years ago, I used high-end ELNA Cerafine audio-grade capacitors. While their ESR was low, their Irac was also low. When installed in an amplifier, the low frequencies were very thick, but somewhat cloudy and lacked transparency. However, after adding a speed-up capacitor in parallel, the sound improved dramatically. Therefore, in actual assembly, remember to add a speed-up capacitor in parallel with the main filter capacitor; this will "at least" improve the high-frequency response. What values? Ideally, one large and one small capacitor, 1μF for the large and 0.1μF for the small, with MKP being the minimum requirement. Sometimes, adding a small capacitor in parallel may not provide much benefit; this might be because the small capacitor was not chosen correctly. For RIFA electrolytic and plastic capacitors, if you want to add them in parallel for speed-up, I advise against using WIMA. I suggest trying MIT's PPFX-S foil capacitors or RTX series 0.1μF capacitors instead. While writing this article, I also noticed advertisements in various magazines. New Hi-End power amplifiers from Krell (USA) and Class'e Audio (Canada) are using Japanese Nichicon capacitors for mains power smoothing and filtering! But which magazine reviewer dares to say they're bad?! Preamplifiers don't handle hundreds of mA of current, so filter capacitors are relatively easy to choose. High-wattage, high-output-current amplifiers are much harder to drive, and the Irac characteristics of the filter capacitors must be considered. Regarding electrolytic capacitors for filtering, there are a few points worth noting: 1. Generally speaking, Japanese products have lower Irac than European and American products; 2. Low leakage current is more important than low ESR; 3. Large filter capacitors should be connected in parallel with smaller capacitors; 4. Choose high-voltage capacitors whenever possible; 5. The highest-end capacitors often have low capacitance and voltage ratings, which explains why high-power amplifiers often sound rough. I don't recommend any one type of capacitor as the best, because any capacitor can produce good sound if used properly. As for those who insist on using only certain brands of capacitors, resistors, solder, and fuses, they are definitely laymen who don't understand circuit structure! Regarding the construction of aluminum electrolytic capacitors: Capacitors can be broadly classified according to their component construction as: 1. Wound type, 2. Multilayer type, and 3. Electrolytic type. Electrolytic capacitors are further divided into aluminum and tantalum types. Aluminum capacitors are further divided into liquid electrolyte and solid electrolyte types. It's incorrect to say that liquid electrolyte is aluminum foil wet type and solid electrolyte is aluminum foil dry type, because both aluminum foil dry and aluminum foil wet types are liquid electrolyte capacitors. Aluminum electrolytic capacitors use etched high-purity aluminum foil as the anode, an anodized film on its surface as the dielectric, and thin paper or cloth impregnated with electrolyte as the cathode. Because the electrolyte is absorbed, it is called an aluminum foil dry type electrolytic capacitor. What is aluminum foil wet type? Adding an electrolyte directly inside the capacitor—such as a mixture of ammonium borate and ethylene glycol—creates a sound like running water when shaken. The Swedish RIFA PEH169 series is an example of this. Even renowned European manufacturers don't produce their own aluminum foil for the anode; instead, it's supplied by a single company, much like how there are many watch factories in Switzerland, but only a few make their own oil movements.