Introduction To Capacitors
"...has many omissions and contains much that is apocryphal, or at least wildly inaccurate..."
The Hitchhiker's Guide to the Galaxy
This site started out as a simple exercise in web
page making that sort of got out of control. Once I started writing, I found it hard to stop.
A few disclaimers: Some of the
information in this "FAQ" should be considered to be generalities. Material source, processing and capacitor construction can greatly affect a capacitor's electrical properties. Different manufacturers tend
to have different opinions on the characteristics of their dielectric materials. For example, most show polystyrene as having a slight negative temperature slope while at least two show it as being slightly
positive. When in doubt, contact the manufacturer to guarantee that the parameter you are interested in is good enough for your needs. This is especially true when using the various graphs. These
should be used more for comparing different capacitor types than for hard design data. That is best obtained from individual manufacturers. In some cases, construction variations so distort the basic
capabilities of the material that it seemed best not to include graphical data. When using charts of dielectic characteristics, it is best to look at a number of sources. One chart for example, shows the
dielectric absorption of mica to be around 0.003%, another shows it to be more like 0.5%. They canīt both be right (.5% is probably closer).
disclaimer: If the graphs sometimes look a little odd (less than optimal colors for example) blame Microsoft. Translating Excel 2000 graphs to GIFs does not seem to work as well as with earlier versions of
Excel. Iīll look into this someday. Drawings made in TurboCAD were another problem. No program I have can both import them correctly, and correctly export them to GIF format. The next version
should correct this.
Much of the available quantitative information of capacitors is not very complete. Dissipation factor vs. temperature is almost
always shown for 1 kHz. DF vs. frequency is normally shown for 20C or 25C. If you want to know the DF of polyester at 95C @ 20 kHz, you will probably be out of luck. Certainly the manufacturers have
some of this sort of information (the good ones anyhow), but it is rarely published. Itīs the rare hobbyist who needs to know that sort of thing however. If you are looking for information on just the
dielectric, as opposed to a finished capacitor, the material manufacturers often have very complete information.
On the other hand, as consumer electronics
moves into GHz frequencies, designers need information on dielectric behavior that may not yet be available.
Some capacitor parameters are difficult to
measure with good absolute accuracy. An instrument for measuring dissipation factor for example, may give good relative results for comparing different capacitors, but not agree with another brand of instrument on
the actual numbers. In any case, there are references to other sources for some of the things I omitted.
The information given here is as good as I can make it, which probably means it has its share of errors. There are also many
omissions; my main interest is in analog electronic applications so that is how this FAQ is oriented. There is some mention of high-power capacitors, but sort of thing is too specialized for good coverage
here. This is meant as an introduction to capacitors, not a book. This FAQ is aimed more at students, hobbyists, scientists, and new engineers than at electronics "insiders". It does not (yet) contain
a dictionary of terms, that has been done by other people. Not covered are some specialized types like feed-through capacitors and trimmers.
This is also a
good place to note that some limitations mentioned, like size, voltage ratings, packaging, temperature ratings, etc., will usually only apply to high-volume capacitor manufacturers. There are specialty
manufacturers who can make most anything you can think of.
What is a Capacitor?capacitor
is a energy storing device made up of two parallel conductive plates separated by an insulating (dielectric) material. When a voltage is applied across the plates, the electric field in the dielectric displaces electric charges, and thus stores energy. It is assumed that there are no free charges in the dielectric (at least in the ideal case), and that while they are displaced, they are not free to move around as in a conductor. The closest analogy in the mechanical world is probably energy stored by a spring. Dielectrics come in two types, "polar" and "nonpolar". Molecules where the "center-of-gravity" (as it were) of the negative and positive charges are at the same point are nonpolar. If this isnīt the case, the molecule is polar. H
2O, for example, is polar, but H2 is nonpolar.
The plates may be actual metal plates of various shapes but are most often in the form
of metal foil or a metal film deposited on the insulating material. Since the first capacitor was invented, the Leyden jar, almost every conceivable dielectric material and form has been tried by someone.
Wax, paper, plastics, ceramics, glass, oils, minerals, electrochemical films, air, etc., either alone or in combination such as paper/wax, paper/epoxy, plastic/plastic, paper/oil, plastic/oil have been used, to name
just a few. Some people seperate capacitors types into "electrostatic" and electrolytic". This is really an artificial distinction, more of a construction detail than a fundamental difference.
Characteristics leakage, ESL, and ESR
, plus special quirks all its own. This makes it tricky to select just the right capacitor for every application.
The reason for the many types of capacitors is cost and performance. Different
capacitor types are better at different things. No capacitor is just pure capacitance and there is no perfect capacitor at the perfect price. Every capacitor type has a number of parasitic elements
associated with it such as
Factors of concern to the user include (but are not limited to):
Aging rate, mostly of concern for ceramic capacitors although all
capacitors age, however little. Film capacitors typically age less than 1% over their life, while some ceramics may age >10%. With films at least, manufacturers tend to disagree over exact
numbers. This is probably due to differences in test methods.
Capacitance vs. frequency, determined by both the dielectric and construction details.
Capacitance vs. size (or "volumetric efficiency") . Mainly determined by the dielectric constant (K)
(sometimes called "permittivity") of the dielectric and its voltage breakdown vs. thickness. For a given dielectric, size tends to vary as the square of the voltage rating. This is a crude approximation at best. It assumes that the dielectric makes up most of the volume of the capacitor. In some capacitors, the packaging and associated materials actually makes up most of the volume. It also assumes that the breakdown voltage increases linearly with the thickness of the dielectric, also a dubious assumption.
Capacitance Drift vs. Temperature, normally given as ppm/deg
C. Every dielectric has a different temperature slope. This can be as low as +- 30 ppm/C for C0G ceramics
(even +-15 ppm/C for special porcelain RF capacitors) to thousands of ppm/C for electrolytics and some plastic films. An old trick for getting reduced drift is to parallel two capacitors which have opposite temperature drifts, such as polyester and polypropylene, or polystyrene and polycarbonate. The temperature drift for a given dielectric can vary from one manufacturer to another however, making this technique uncertain. Polycarbonate for example, is shown by some manufacturers to have a slightly positive slope, and slightly negative by others. Caps with mixed dielectrics exist, but are not common.
Capacitance vs. voltage, mostly of interest for Class 2-4 ceramic capacitors, and for some glass types.
Cost and availability.
Current-carrying capacity, of concern when the capacitor must handle high alternating current which would cause self heating. This a problem for electrolytics, and for film and ceramic capacitors used
in high frequency and/or high current applications. Manufacturors often have guidelines.
Dielectric absorption (or DA), sometimes called "soakage", "voltage retention", and other things. In high-voltage power cables itīs called "return voltage". A capacitor, once charged, stubbornly
retains part of the charge, even after being discharged (shorted for some number of seconds), as if it had "soaked" into the dielectric. DA is modeled
as an infinite series of RC networks in parallel with the primary capacitance, where the Rs are mostly very large and the Cs are smaller than the nominal capacitance. DA is a source of error in precision integrators and sample-and-hold circuits. Large high-voltage oil-filled capacitors (exceptionally high DA) can be shorted, yet retain enough charge to be dangerous. They are normally shipped with a short across the terminals that should not be removed until the part has been installed. High-voltage capacitors in TV sets are also still dangerous after being discharged, as is the parasitic capacitance in large oil-filled, high-voltage transformers and some high-voltage cables. There is a "standard" test for DA (actually, everyone seems to do it a little differently). The capacitor is charged to some voltage for one minute (often 100 volts, but others can be used), and is then shorted for two seconds. After one minute, the recovered voltage on the capacitor is read using a very high impedance meter. The DA is expressed as the ratio between the recovered voltage and the charging voltage, in percent. MIL-C-19978D calls for a 5 minute/5 second/1 minute sequence. These tests are useful for comparing dielectrics, but don't really tell you just how a capacitor will perform in your application. In particular, the difference in DA of various dielectrics as seen one minute after the short is removed does not necessarily precisely correlate with the difference in DA as seen in the first few 10s of ms. In general, Teflon, polystyrene, and polypropylene are the best (as low as 0.02%), while the electrolytics, high-k ceramics, and oil-filled are the worst (1% on up). Impregnated film caps tend to have a DA that reflects the impregnant more than the film. Given the RC model for DA, it would make sense that the dielectrics with the highest insulation resistance (which also tend to have the lowest dielectric constant) would have the lowest dielectric absorption. There is also a significant difference from part to part. However, while the RC model is usefull for predicting how DA will behave, it does not reflect the underlying physics. See
Space Charge. Also see: http://www.national.com/rap/Application/0,1570,28,00.html
Dissipation factor (DF), of concern for AC power applications. DF is usually expressed in % or in ratio to unity (1% = .01). DF is equal to ESR/Xc
. Actually, this is an approximation which ignores Xl, but it is usually close enough. Like most everything else about capacitors, it changes with time, frequency, and temperature.
DF is the result of three loss factors. Metal losses (the resistance of the leads, end terminations, and the metal foil or film), the insulation resistance (mostly very small), and the dielectric
losses. For film capacitors, film-foil types will have a lower dissipation factor than metallized film, especially at high frequency, because the metal losses will be lower.
Dissipation factor vs. applied voltage, mostly of concern for Class 2 (and higher) ceramic capacitors.
Dissipation factor vs. frequency.
Dissipation factor vs. temperature.
Equivalent series resistance
(ESR), mostly of concern in high-frequency power applications and for power-supply filtering in high-speed digital systems. ESR is measured at different frequencies, depending on how a capacitor will be used. For cheap electrolytics this will be 120 Hz, for high-quality tantalums, it will usually be 100 kHz. Although measured in ohms, it is not independent of frequency, and in electrolytics, is somewhat higher at low frequency than at high frequency.
Equivalent series inductance (ESL), seen as an inductance in series. Mostly of concern for power-supply filtering in high-speed digital systems, and in RF applications.
Frequency range. Some capacitors, like mica and C0G ceramic are usable to the GHz range. Others, like tantalum and polyester, become unusable somewhere in the 100 kHz to 1 MHz area
(depending on construction details) due to loss of capacitance and increase in dissipation factor.
Line filter capacitors. As a matter of safety, capacitors
used to filter line (mains) power for electronic equipment, must meet strict performance requirements. The Europeans have lead the way in this, while American and Canadian requirements have been more
Insulation resistance (IR), seen as a resistance in
parallel. This is referred to as "insulation resistance" when the leakage is very low (film capacitors), and as "leakage current" when the leakage is high (electrolytics). Electrolytic capacitors may
have the most leakage, but it is also of concern for film capacitors in some analog applications (integrators, sample-and-hold circuits). For films, the lower the dielectric constant, the higher the
insulation resistance tends to be. Data sheets normally state it in megohm-microfarads. To determine the IR for a given capacitor, divide the M-uF value by the actual capacitance.
Although film capacitors and C0G ceramics have the best IR, all things being equal, low-quality ceramics can still work in some leakage-sensitive applications if the value is low, the voltage rating
is high, and you can live with their other shortcomings.
Markings. It can sometimes be a challenge to determine the
characteristics of some capacitors from their markings. This partly due to size limitations, partly to a lack of universally followed standards, and partly to the failure to consistently follow the
standards that do exist.
Noise. Mostly of concern for electrolytic capacitors, due to leakage.
Packaging and construction. Construction details and
sealing methods can have a significant impact on performance and reliability. Most capacitor types require that air and moisture are sealed out to prevent degradation or contamination of the dielectric and
corrosion of the metal film. The best Military-grade capacitors may be hermetically sealed in metal and glass. Lesser types may be molded in plastic, dipped in epoxy, or inserted in a
plastic case and sealed with epoxy or urethane. The moisture absorption of the dielectric dictates the encapsulation methods allowed. As surface mounting of components replaces through-hole
construction, a new problem faces the designer: can you get the capacitor you need in a surface mount package? Some capacitor types, like ceramics, are well suited to surface mounting. Others, like
polystyrene and polypropylene, do not stand up to the heat of soldering and are not available for surface mounting.
Reliability. Reliability and expected lifespan are of
special concern in high-temperature, high-current, and high-voltage applications. However, capacitors used in more ordinary applications also have reliability issues.
Surge current capability. Of concern for high-current pulse applications. Manufacturers will specify a maximum current or a maximum dV/dT (in volts/usec.) for capacitors rated for high-current
Temperature limits. Some dielectrics are limited to as low as 85C operating temperature while others can go to the 200-400C region. I have seen references to aerospace capacitors that can work to
at least 1000C.
Voltage limits, both operating and surge.
You can roughly divide all capacitors into four groups: film, ceramic, electrolytic, and
Film includes a variety of polymers, such as polyester, polycarbonate, Teflon, polypropylene, and polystyrene. Traditional film capacitors were only available in modest
sizes, <10 uF. In recent years, film capacitors have sought to leverage their superior longevity compared to
electrolytics, to move into some applications that call for much larger parts, even to thousands of uF. Film
capacitors come in two broad categories, film-foil, and metallized film. Film-foil capacitors are made of
alternating layers of plastic film and metal foil, while metallized film capacitors have the metal vacuum deposited
directly on the film. In general, film-foil is better at handling high current, while metallized film caps are much better at self-healing. Various hybrid types can also be found.
Pros: The film capacitors mostly have reasonably well behaved electrical properties and offer many
tradeoffs of performance and cost for people with precise requirements. The main parameters of interest
include capacitance vs. temperature, dissipation factor, and dielectric absorption. Their main virtues include low leakage and low aging.
Cons: The main drawback of film capacitors is their low dielectric constants (K), which is only partly
offset by their relatively good breakdown voltages. That means that film capacitors are physically large for
their capacitance. Their Ks vary from a low of about 2.2 for Teflon to about 8 for PVDF (rarely used).
Unfortunately, the rule-of-thumb is that the higher the K (and therefore the smaller the size), the worse the
electrical properties tend to be. Film capacitors have not made an entirely graceful transition into the age of
surface mounting. While some film dielectrics are suitable for surface mounting, most can't withstand the heat
of soldering. Even polyester, the toughest of the traditional films, is barely good enough. However, capacitor
makers have responded by developing several new dielectrics. SMD film capacitors are not as widely second-sourced as other capacitors however.
Ceramic capacitors offer a broad range of size vs. performance tradeoffs and are easily the most
popular in numbers sold. Ceramic capacitors are available from < 1 pF to 1000s of uF.
Pros: The main virtue of ceramic capacitors are their relatively high dielectric constants. This can vary
from C0G with a K of up to 60, which has excellent electrical properties but is relatively large and expensive, to ceramics with Ks in the tens of thousands but with very poor electrical properties. Large-
value ceramics can replace electrolytic capacitors in high-frequency applications like switch-mode power supplies because of their
lower ESR. Ceramic capacitors are especially suitable for surface mounting due to their heat resistance,
mechanical integrity, and the ability to make them in very small packages at low cost, for portable equipment.
This has greatly added to their usage. To some extent ceramics are slowly displacing other types of capacitors.
Cons: Low breakdown voltage means that the low-K ceramics (Class 1), the ones with the good
electrical properties, have poor volumetric efficiency, and are usually found only in small values. High-K ceramics (Class 2 and higher) have poor electrical properties, which are highly dependent on temperature,
voltage, and frequency, plus a significant aging rate. Unlike many other capacitors, ceramics have no
self-healing mechanism. This means that manufacturers must maintain a high level of quality control over the
dielectric. Ceramics are most cost affective in small sizes at present. Very large ceramics are a bit of a challenge, especially in SMD.
"Electrolytic" means any capacitor that requires a conductive layer between the dielectric and one electrode. In the original electrolytic capacitor, the layer was an actual electrolyte, a conductive salt in a
solvent. Some electrolytic capacitors today donīt actually use an electrolyte, but the word is still commonly
used, to the annoyance of some. Electrolytic capacitors are made by growing a oxide film, the dielectric, on a metal, the anode, by electrochemical means. The films are very thin with fairly high Ks (roughly 10-25) which make for a lot of capacitance in a small package. The resulting devices pass current much better in one
direction than the other, making a rectifier of a sort. Because of this, the metals are sometimes called "valve" metals. The metals presently used are aluminum, tantalum, and niobium.
Pros: Electrolytic capacitors are best used when you need a lot of capacitance in a small space and at a
reasonable price, such as power supply filtering, or energy storage. They are available in sizes far beyond that
of other capacitors. Aluminum electrolytics are presently available from 0.1 uF to several F. I have no idea
why someone would use a 0.1 uF electrolytic capacitor however. Tantalum electrolytics are available from 0.1 uF to a few thousand uF.
Cons: Marginal electrical properties means that these capacitors must be applied with care. The
parameters to be watched include leakage, service life vs. temperature, ESR, ESL, and low-temperature performance. Unlike other capacitors, electrolytic capacitors are not inherently non-polar, but non-polar types
are available. Electrolytics are widely available in SMD packages, at least in moderate sizes, but users complain of more reliability problems than with through-hole styles.
Miscellaneous capacitors include materials like glass, mica, porcelain, and even gas and vacuum. A few
exotic dielectrics like silicon dioxide and sapphire are used in niche applications like microwave capacitors and trimmers. Some are available in surface-mount packages.
Pros: The electrical properties of the miscellaneous capacitors are generally most similar to film
capacitors. However, they all have electrical properties that make them useful in some special applications.
Cons: These materials also have Ks similar to plastic films so they have no advantage in size. Except
for mica, these capacitors are commonly available only in small sizes, <1 uF. They also tend to be more expensive than other capacitors of similar size.
Anyone who has corrections or additional information is more than welcome to contribute. If you have
an idea for additional topics, expansion of existing topics, if something unclear, if you see an outright error,
whatever, let me know. The site is still in its infancy, and I donīt really have a clear vision of what it will grow
into. I would especially like to hear from the capacitor engineers of the world, they are the people who should
have written this FAQ, but didn't. So far, my attempts to get these people involved has had limited success. It
would be nice if an expert with specific knowledge would offer additional or corrected information not available to me.
What is good or bad about the site?
What would you like to see added?
Is your interest mainly for personal or business reasons?
Were you looking for something in specific, and did you find it?
*: Research. Presently being treated for a long-standing capacitor fetish .
Dave D.: This was all his idea.
Schmuddy: Editing (real name withheld by request).
John Woodgate: For considerable help with the Line-Filter Capacitor section.
For some background material and graphics on stacked-film capacitors.
Oren Elliott Products: For background material and graphics on air-variable capacitors.
For background material and a graphic on vacuum and gas capacitors.
Eichhoff Electronics: For graphics of line-filter capacitors and the line-filter modules they are used in.
Copernic2001: A great tool for searching the internet.
Google: Another great search tool.
This site Copyright
Đ 1999-2010. All rights reserved.
Individuals may copy portions for personal use only.
All trademarks are property of their respective owners.
Links to other sites are not an endorsement of their information or products.
Date of last change: 9/14/09. Changed main font to 12, minor editing changes, changed or eliminated a few links.
* In a bit of bad news, my company was purchased by a conglomerate and at the end of July of `02 I became unemployed.
However, I have made my plans for getting my career back on track.