Substituting a tantalum for a ceramic type will not work, but they can work together very well.
The Capacitor Future:
The future for capacitors looks good.A constant search is going on by companies like Murata, Kemet, etc. Kemet in particular is researching a new type of adielectric substance called Niobium. Niobium Pentoxide (Nb2O5) offers a higher dielectric constant of 41 in comparison to Tantalum Pentoxide (Ta2O5) at 26. It implies that approximately 1.5 more CV (Capacitance x Voltage rating) can be obtained from the same amount of material, everything else being equal. What does this mean in plain english? Much smaller capacitors with larger capacity, especially important in surface mount technology. Recently, a new type capacitor with very high capacitance has been developed with capacitance designated in Farads! Yes, you read it well, Farads. This type of Electric Double Layer capacitor is known as a "Super Capacitor". I am sure we haven't heard the last of it about this type. At this particular time (2007) capacitors of 5 Farad are already available!
CAPACITOR NOMENCLATURE -- by Dean Huster
The questions of reading capacitor values, tolerances and their units of measure come up often on various Internet forums and have hit the old "Q & A" column that I wrote for Poptronics magazine (formerly Electronics Now and Popular Electronics magazines) during its last two years of existence. Here's an attempt at a global treatise on the subject.
The Farad, MicroFarad and PicoFarad
Capacitors have always had farad as the unit of measure, abbreviated "F". Since this is a very large unit of measure for most practical capacitors or for most uses of capacitance, you'll find that a millionth of a farad or a million-millionth of a farad are the more common units found on capacitors. Yes, these days we can find capacitors with ratings in the tens and hundreds of farads, but those are usually reserved for extremely high-current, low-voltage switching supplies or for a more frivolous use as energy-storage tanks for use with high-power automotive audio power amplifiers. This treatise is for "normal" capacitors.
In scientific notation, we would write 1 millionth of a farad as 1 x 10-6 farad. In electronics, since we deal with so many component values and circuit values on even the smallest schematic or product, the metric prefix form is used for an electronic shorthand to keep the scribbling to a minimum. That prefix form uses letter symbols to take the place of the scientific notation--or more accurately, the engineering notation--that would otherwise accompany a unit of measure. The metric prefix form replaces the engineering notation [Note 1] that would otherwise be used in front of the unit of measure. That list follows.
Metric Prefix Symbol Power of 10 (multiplier) giga [Note 2] G x 10^9 mega M x 10^6 kilo K x 10^3 (none) x 10^0 (same as 1 or unity) milli m x 10^-3 micro f x 10^-6 nano n x 10^-9 pico p x 10^-12 This list does extend farther in either direction, but those larger and smaller multipliers are not as commonly used in electronics. But using this list, you'll find that the common capacitor multipliers in the United States will be f (micro) and p (pico). A capacitor with a value of 3.3fF is the same as a capacitor with a value of 3.3 x 10^-6 farads or 0.0000033 farads. "f", by the way, is the lower-case Greek letter "mu", properly written as our Roman lower case "u" with a leading descender much as a "y" has a trailing descender.
Antique Capacitors and Schematic Diagrams
Prior to the 1960s, capacitors were called "condensers". Same part, same function, different name. You'll still hear the old name used by auto mechanics when they speak of the condenser in the ignition system on an older vehicle.
That earlier list of prefixes, although valid in the hayday of the vacuum tube, was used even less extensively back then. Mega (M) was the largest prefix commonly used and micro (f) was the smallest. To achieve larger or smaller multipliers, the prefixes were doubled up just like multiplying two numbers together. So, a small capacitor value might be 47ffF ("forty-seven micromicro farads") which was saying 47 x 10^-6 x 10^-6, which is the same as 47 x 10^-12, or 47pF. So, anytime you see "ff", just think "p" in your head. [Note 3] You'll also hear some old timers use the rather ridiculous term mickey-mikes instead of saying micromicrofarads. Makes me glad that we went to using pico.
To make it even worse, the symbol for micro, m, was not used often before 1970 except for expensive physics and engineering texts. It was not a common character to find in a print shop and impossible to find on a typewriter, unless you had the special "symbol" typeball for the IBM Selectric. So, the lower case "u" was usually used in its place. That means that on schematics of the 1950s, a capacitor that was listed as 68uF is the same as one today listed as 68fF. In addition, the components themselves rarely had even the "u" listed. The actual part could have any one of the following printed on it:
68 mF 68 MF 68 mfd 68 MFD It was always understood that "mF", "MF", "mfd" or "MFD" ALWAYS meant microfarad. Microfarad or micromicrofarad were the only units used for capacitors back then, so no one would ever even consider that "mfd" might mean "millifarad" or that "MFD" might mean "megafarad"! Even today, you'll still see "MFD" on capacitors, especially on motor start or motor run capacitors.
In addition, antique capacitors did not hold to today's standard value system. More on that later.
Even while we had nanoseconds and millihenries and other such measurements in the United States, capacitor values were only given in microfarads and picofarads (micromicrofarads) and that holds true even today. I have yet to ever find anyone or any source that can give the reason for that. I'm afraid that "because we've always done it that way" will be the only reason you'll ever hear. The practice thwarts the rules of engineering notation [Note 1], resulting in capacitor values written as 0.0068fF or 6800pF, both of which are the same value and either of which iscorrectly written in the engineering or metric prefix systems of notation. In addition, in this world of large computer-grade caps, you'll also find values of 180,000fF. Some companies preferred to use the latter of the two forms of notation just to keep the decimal point out of print so that there would be no confusion, preferring for instance, 6800pF over 0.0068fF
Meanwhile in Europe, and later Asia, the electronics folks there would have no problem at all referring to that same value of 0.0068 fF or 6800pF as 6.8nF (6.8 nanofarads). So, on diagrams from the U.K., don't be surprised to see both millifarads and nanofarads used extensively. And that's OK. It makes a lot more sense and it's easier to write. The only problem is that we Yanks have to learn to convert or at least get used to working with milli and nano when dealing with capacitors. The 180,000fF cap mentioned earlier would be more-correctly rendered as 180mF by someone in Germany.
Cryptic Capacitor Markings
In the 1960s, capacitors were marked with their value. Plain and simple. But nevertheless, it was a poor system as decimal points and zeros often got lost. Paper capacitors often had a wax coating on the outside that obscured the markings and these same markings could be accidentally "edited" when scraping the wax away for a better view.
Similarly, ceramic disc caps also often had a wax coating, possibly just from being next to waxy paper cap in a warm television set. A cap might be marked as ".0047MFD" or simply as ".22" and without that leading zero, there could be problems if the decimal point disappeared. You had to use a little common sense when reading these values. If the value was in decimal form (e.g., ".047"), you assumed the value to be in microfarads. If it was a whole number (e.g., "470"), you had to determine what type of capacitor you had. If it was an electrolytic type, you assumed the value to be in microfarads. If it was a small ceramic cap, it was in picofarads. You knew that a small ceramic or plastic cap couldn't possibly be as large as 470fF and that an electrolytic would nearly always have a value larger than 1fF.
During the 1970s, manufacturers began to mark the caps in a way similar to the way resistors were done, but without the color code. Rather than colors, they just used the actual printed digits, and the value was marked on the cap as two significant digits and a multiplier, the value being in picofarads. So, a cap marked "221" doesn't mean 221fF or 221pF. It means 22 + 1 zero or 220pF. "684" is decoded as 680,000 (68 + 4 zeros) pF or 0.68fF. Capacitor values smaller than 100pF might be printed as just two digits, such as "33" for 33pF or as two digits with a zero multiplier, as in 330 for 33pF.
This method of marking caps takes care of a lot of problems with disappearing decimal points and large numbers (e.g., .0047MFD) that would otherwise be difficult to print on small parts. But it did put in a bit of confusion when compared to the older way of marking. What did "100" mean? If it was an older cap, it meant 100pF. If it used the newer marking system, it meant 10pf (10 with no zeros added). So, you need to know which system is in use and that can be tricky.
That's why it's nice to have a digital capacitance meter or a capacitance bridge at hand. You can get a "ballpark" check of the value to determine where to go from there.
Later, manufacturers took a hint from the Europeans and begin to mark their capacitors like the Brits marked them on schematics. This took care of ALL the problems of marking as well as problems associated with schematic value markings. In the Euro system, you revert to the engineering form, but replace the decimal point with the metric prefix multiplier. So, the oldest "4.7MFD" capacitor that was updated to "4.7fF" and then to "475" (picofarad) with the later system was now marked as "4f7". Presto! The "f" shows the location of the decimal point AND gives the metric prefix, all at the same time! This is really the best system in the world, both for marking the components themselves and for marking the values on schematics. If a decimal point disappears because of wear or poor printing, it can mess things up in the old system. An entire character would have to disappear in the newest system.
You will find plenty of folks, especially here in the U.S. who don't like the Euro way of marking. But these same folks don't like change. As a matter of fact, a lot of these old clowns will still use a "33 em-em-eff condenser to set the frequency of their 4.5 megacycle" oscillator. Just ignore them!
Value tolerances used to be marked on the capacitor in percentage. An older cap might be marked as ".05 10%", meaning 0.05fF, 10%. Newer caps use a letter for the tolerance, and that will seem confusing right at first. A common value might be ".1M" which means 0.1fF, 20%. A little newer cap might be marked as "332K" and that drives some folks nuts. After all, "K" is a standard metric prefix multiplier and they automatically think they have a 332,000fF cap on their hands. In reality, they have a 3300pF (33 + 2 zeros, in pF) with a ±10% tolerance. The letters on these caps correspond to the following list. I've marked the tolerances that you'll find to be the most common with an asterisk.
B = ±0.1pF C = ±0.25pF D = ±0.5pF E = ±0.25% F = ±1.0% G = ±2% H = ±2.5% J = ±5% * K = ±10% * L = ±15% M = ±20% * N = ±30% P = -0, +100% S = -20, +50% W = -0, +200% X = -20, +40% Z = -20, +80% * You'll find that the "Z" tolerance of -20, +80% to be common for aluminum electrolytic caps and for disc ceramic caps that are used for what is known as "bulk capacitance" in applications such as power supply bypassing or filtering. These kinds of capacitors are used where it's OK for the value to be a lot larger than nominal, but they don't want it to go very far below that value.
If you do a lot of analog circuit design and building where you attempt to get frequency-dependent circuits to be as accurate as possible, you'll see or want to find Mylar caps with a "G" or "F" tolerances of ±2% or ±1% respectively. They're harder to find in catalogs, but if you watch the electronics surplus catalogs, you can find then on a sporadic basis. The tolerances of "B" through "E" are in pF vs. percent and are normally used on small caps of around 10pF or less.
When the numbers on a capacitor begin to rub off because of age, and you need to know the value, it may help to know that at least in more modern times, manufacturers have adhered to a set of standard values for capacitors. Typically, these will be the same list of standard values used for 20% tolerance fixed resistors, each decade having six possible values. This simply means that the significant digits of any capacitor value will be either 10, 15, 22, 33, 47 or 68. You'll find that 10, 22, 33 and 47 are the most-commonly used. This set of standard values applies whether the capacitor has a tolerance of 20%, 10% or 5%.
Why not have the full 24-per-decade range of values for the 5% capacitors just like we do for resistors? In most cases where tolerance might be a factor, a capacitor is used against a resistor for setting some type of timing or frequency of a circuit. There's no sense having all of those values for both components. So, since the capacitor is a lot more expensive to manufacture and is prone to larger tolerance variations, they limit the number of values to six per decade. A designer will then choose the value of capacitor most closely suited to the application and then "fine tune" the actual RC time constant by choosing one of the 24-per-decade resistor values. If the circuit is to be very accurate and stable, the designer will choose a 2% or 1% tolerance capacitor with a low temperature coefficient and then use a 1% metal film resistor for the resistive element where there are 96 possible values per decade.
All this is to say that if you're trying to decide if that capacitor you're trying to replace in a radio is marked "65" or "68" because the last digit is all scratched up, choose "68" since that's a standard value and is most likely the marked value. This is not to say that a capacitor will always hold to the standard value table. Older capacitors had values such as 0.03fF or 0.05fF. These were made before extensive standardization. And then there are specially made caps for the telephone industry. I've seen values such as 0.1746fF with a 2% tolerance. All those significant figures with a 2% tolerance is foolish since even the "4" can move up or down by almost three digits and still be within tolerance. And some manufacturers will feel the need to "fill in" the table by providing values that conform to the 12-per-decade table of 10% components, but that's OK.
Just don't expect it for all capacitors.
In addition to value and tolerance, a capacitor is often marked with a voltage rating. These may simply be noted as "50V" or "50VDC" or some such other voltage as appropriate. Voltage ratings are sometimes incorporated into a capacitors "coded description". For instance, the value code "2A104K" has a "2A" prefix which translates to a voltage rating of 100V. The "104K" part, as you now know, translates to 100,000pF or 0.1fF or 100nF with a tolerance of 10%. Voltage prefixes include:
1E 25V 1H 50V 2A 100V Since this seems to be European in nature, these voltage markings are new territory for me. I would appreciate more information on this so that I can flesh out this article and make it more accurate. My e-mail address appears in the "Wrapup" section following in case you would like to contact me with some of this information. I try to be accurate, so please make sure that you include source material rather than depending upon hand-me-down folklore!
Capacitors, most notably ceramic capacitors, have temperature coefficients ("tempco" or TC). That is, their value will change with a change in temperature. Some "bulk" ceramic capacitors (those "M" tolerance things) can change over 10 or 20 percent with a 20 degree shift in temperature, so are unsuitable for use in circuits that are frequency-dependent, such as oscillators or filters. Capacitance changes are not necessarily linear or even directly proportional at all times for a particular type of capacitor.
You'll see markings on a cap such as "Z5U". That's a temperature coefficient. The "Z" part has nothing to do with the tolerance. I have no intention of going into tempcos here since it has nothing to do with reading the value. I just wanted you to be aware of them so that you don't confuse them with the tolerance.
As usual, if you find any errors, however small, in this article or if you feel there's some information that needs to be added, please don't hesitate to contact me at: dhhuster at semo dot net (don't forget to use "@" for "at" and "." for "dot"). If everyone wants to see tempco information here, then I'll go ahead and add it.
Note 1: While scientific notation takes the form, n.nnnnn x 10^p where you have only one digit to the left of the decimal point and an appropriate power of 10 so that the number in scientific notation will be numerically equal to the original number in decimal form, engineering notation takes one of the three following forms, n.nnnnn x 10^3p, nn.nnnn x 10^3p, nnn.nnn x 10^3p, where there are either one, two or three digits to the left of the decimal point to achieve a power of ten that is an integer multiple of three, the final number being numerically equal to the original number in decimal form or scientific notation form. So, the only powers of 10 allowed in engineering notation are powers such as 15, 12, 9, 6, 3, 0, -3, -6, -9, -12, -15, etc. Engineering notation is handy in that it directly translates into the metric prefix form, which also uses powers of ten that are integer multiples of three. Some models of Casio scientific calculators are capable of displaying their results in engineering notation while some even have the capability of displaying in metric prefix form.
The metric prefix form follows the engineering form exactly with two common exceptions. The "bel", the unit of measure for a ratio of sound or signal amplitude levels, it too large to be practical.
One-tenth of a bel, or a decibel, is the normal unit used, abbreviated "dB", a term you have probably seen often. "Deci" is the same as "x 10^-1".
Another exception is the centimeter, often used in the metric system where we would use inches in the English or SAE system. A centimeter is 1/100 of a meter, abbreviated "cm". "Centi" is the same as "x 10^-2".
Note 2: "giga" is technically pronounced "jigga" beginning with a soft "g" as in the word "giant" with a hard "g" as in "gallup" the second time, and as is done by the character of Dr. Emmett Brown in the motion picture "Back to the Future" when he refers to "one point twenty-one gigawatts of power." However, you'll find that nearly everyone involved in electronics pronounces it beginning with a hard "g" as they would the "g" in the word "go".
Note 3: Other units of measure had similar doubling up. For instance, fast oscilloscope risetimes were noted as 3mfs or "3 millimicro seconds" which is the same as today's 3ns. At the other end of the spectrum, so to speak, microwave signal generators were marked as having frequencies of 1.5KMC or "1.5 kilomega cycles (per second)" which is 1.5GHz in today's vernacular. In addition, "cycles per second" as the unit of measure of frequency, which was usually shortened to the erroneous "cycles", was changed to Hertz, abbreviated Hz, in the 1960s.
Thanks Dean for your contribution, much appreciated!
Email note from Bas Viel
I received an email from Bas Viel, the Netherlands, see below.
Email in regards to the Engineering button [ENG} on a calculator, not every calculator has one.
There is a Table 1 for conversion at the capacitors page. No need for it!
Only remember REN or LEP
If you have a value what isnot in engineering notation and you want it to be, just move the decimal digit left or right, but... what is the new exponential value?
REN stands for move digit to the Right, Exponential goes (more) Negative!
LEP is the opposite for moving the digit to the left.
No calculator or Table needed!
"The Radio Amateur Handbook" from the American Radio Relay Leaque (ARRL). Good resource.
"The Capacitor Book". by Cletus J. Kaiser., C.J. Publishing. ISBN: 0-9628525-3-8
"Williamson Labs". Website about bypassing, decoupling, shielding, layout, etc. (http://williamson-labs.com)