What is AWG (American Wire Gage),
and Does it Matter?
...and when does it matter, and why?
Wire gage sizes are a bit confusing, and we get a lot of questions about them. Why does one 12 AWG speaker cable look smaller than another? Is wire gage a good indicator of cable quality? What is wire gage, anyhow, and when and why does it matter? Let's look at these issues.
What is AWG (American Wire Gage)?
Wire gage is an index which shows, indirectly (inversely and logarithmically), the cross-sectional area of a round wire. In the case of solid conductors, measurement of this area is pretty straightforward: the area is the radius of the wire squared, times pi, and for the sake of ease of expression, a measure called "Circular MIL area" is instead often used instead; one circular mil is the area of a circle with a diameter of one mil (1/1000 inch), and the circular mil area of a solid wire, consequently, is always the diameter of the wire, in mils, squared.
Stranded wire is another matter. For any given AWG size, a stranded wire will occupy more space than a solid wire, because the wire gage is measured by summing the cross-sectional area of the strands. Because there are air pockets between the strands, any given cross-sectional area of wire will take up more overall space in a stranded configuration than it will in a solid wire. Consequently, when we talk about "diameter" relative to wire gage, it's well to remember that diameter will vary not only with gage but also with stranding. In this article, when we talk about relative diameters, our examples are based on solid wire for the sake of simplicity.
The relationship of gage to wire size is, for a lot of people, counterintuitive. The larger the gage number, the smaller the wire. What's more, the relationship isn't linear, but logarithmic. Two 16 AWG wires, combined, amount to a 13 AWG conductor. If you're familiar with decibels (dB), this will make good sense. If we go up or down 10 gage sizes, we increase or decrease the area of the conductor by a factor of 10. If we go up or down 3 gage sizes, we increase or decrease the area by a factor of about 2. For some reason (we're not really sure why) the relationship isn't precise, but it's close enough, for most purposes, to a straight logarithmic formula. For example, a 40 AWG solid wire has a circular mil area, as specified by the National Bureau of Standards, of 9.61; a 30 AWG wire has a circular mil area of 100.5, a 20 AWG wire comes in at 1020, and a 10 AWG at 10380.
Incidentally, it's important to remember that it is the size of the WIRE, not the size of the wire with its insulation, that is measured in AWG. On occasion, we get a call from a customer who is convinced that our 12 AWG speaker cable cannot be 12 AWG, because it looks smaller than another 12 AWG cable he owns. Many speaker cables are jacketed in a very thick translucent PVC jacket, which not only makes the overall profile bulky, but also makes for something of a magnifying-glass effect, making the wire look a bit bigger than it really is.
What Does Wire Gage Have to Do with a Wire's Electrical Properties?
The most significant impact of Wire Gage upon the electrical properties of a wire is upon the wire's resistance. Any given wire material (copper, steel, aluminum, et cetera) has resistance, and DC resistance is inversely proportional to the circular mil area. If our wire is copper, that 40 AWG conductor, with a 9.61 area, has a resistance of 1080 ohms per 1000 feet; the 10 AWG, with approximately 1000 times the area, has a resistance of just about exactly one ohm.
Resistance is the property of a conductor which describes how current flowing through the conductor will be converted to heat. In a very low resistance conductor, relatively little energy will be lost to heat; as resistance increases, more and more will be converted to heat. How this affects electrical circuits varies with the type of circuit involved, however, and we'll get to that in a bit.
But Isn't that "DC Resistance"? Aren't Signals AC?
One of the most common misconceptions we run into, on the subject of resistance, is that resistance is somehow irrelevant to audio and video signals because those signals are alternating current (AC), and a wire's resistance is expressed as "DC resistance," which refers, of course, to direct current, not alternating current. So, we are often asked, if resistance is DC but the signal is AC, what could resistance have to do with anything?
Resistance acts upon both alternating current and direct current. The reason resistance is expressed as "DC resistance" on spec sheets is not that resistance is not applicable to alternating current. Rather, it's because of something called "skin effect." As the frequency of a signal increases, the current flow in a wire concentrates toward the outside, or "skin," of the conductor. This means that for any given wire, if we measure resistance at different frequencies, we'll find that the resistance increases with frequency. Resistance is expressed in spec sheets as "DC resistance" because the resistance value of one wire at DC can be meaningfully compared to the resistance of any other wire at DC. In theory, if one wanted to do so, one could specify the resistance of wires at any frequency; we could make up tables of "1 MHz resistance" instead of DC resistance. This isn't done because (1) there isn't any handy "reference" frequency which is broadly applicable to all uses of wire, and (2) it's harder to measure resistance properly at higher frequencies because it is difficult to separate out losses to other factors which become relevant as frequency increases, like capacitance, inductance, and return loss. But make no mistake: resistance converts electricity to heat in a wire regardless of whether the electricity is DC or AC. And, by the way: in the case of a stranded wire, the "skin" in question is still the outside of the bundle; it is not, as people often assume, the skin of each individual strand.
So, AWG relates to resistance. What does Resistance Mean to Signal Quality?
What does resistance have to do with signal quality? Well, that depends very much on the application. It's commonly assumed that AWG is a good indicator of cable quality, and this assumption goes back to the earliest days of marketing of "aftermarket" speaker cable; the sales pitch which launched the whole consumer aftermarket cable business was, in essence, "bigger wire is better." And this, as we'll see, is certainly true for speaker cable (within limits), but not necessarily for other applications.
Before we get into this, a couple of preliminaries. First, it's important to remember that what we are concerned with primarily is signal quality, not amplitude. If losses in a system are not frequency-dependent, it's very easy to adjust around them; for example, typical video input circuits will simply take weak signals and amplify them to a standard reference level for use in a display. In such a case, we want to be sure that the quality of signal is clean, but it doesn't matter--at least, it matters relatively little, within reasonable limits--whether the amplitude of the signal is high or low.
Second, to understand the following discussion, it's helpful to know a bit about something called Ohm's law. The German physicist Georg Ohm discovered a simple principle about resistances which is a fundamental idea underlying all manner of electrical circuits. If a circuit contains a series of resistances--that is, if current is going to flow through one resistor, then through another, and then another--the energy of the electrical flow will be absorbed by those resistors in proportion to their resistance (which, of course, we measure in Ohms, in honor of Georg Ohm's work). You will also, probably, be familiar with another use of "ohms": impedance. Impedance is a more complicated phenomenon than resistance, and there's a lot to be said about it; but for the purposes of the following examples, we can consider ohms of impedance to be equivalent to ohms of resistance, as though impedance and resistance were exactly the same thing.
So, to illustrate Ohm's law, let's consider a speaker circuit, and we'll suppose, for the sake of this example, that the installer has decided to use a dramatically undersized speaker cable. Each conductor of this cable has a resistance of four ohms, and the speaker has an impedance of eight ohms. Signal coming from one speaker terminal and traveling to the other will go through four ohms of speaker wire resistance, through an eight-ohm speaker, and then through another four ohms of speaker wire resistance. What does this mean? The total circuit resistance is 16 ohms (to simplify matters, we're going to assume a zero ohm "output impedance"; this isn't realistic, but is good enough to illustrate the principles at work here). So, of the energy being burned up in the circuit, one quarter (4 ohms over 16 ohms) is burned up on the way from the "plus" terminal to the speaker; one half (8 ohms over 16 ohms) is delivered to the speaker; and one quarter is burned up on the other side of the speaker cable, between the speaker and the "minus" terminal of the amp.
Obviously, that's a lot of energy being burned up in speaker cable. In our discussion below, we'll explain why that's a bad thing (beyond just being a waste of electricity). But before we talk about that, let's imagine another application. Let's say that we take a cable with the same resistance properties (4 ohms out, 4 ohms back) and hook it up to RCA connectors, and use it on a line-level analog audio connection between a source device (say, a CD player) and an amplifier. An amplifier input circuit will not have a low impedance like a speaker will; 10,000 ohms, rather than 8 ohms, is somewhere around typical. Now, when we hook this circuit up, what do we find? The total circuit resistance is 10,008 ohms. Of the energy being delivered by the source, 8/10008 of it--almost nothing--is burning up in the cable, and 10000/10008 of it is being delivered to the amp. The resistance that was horribly excessive in the speaker cable, and that was consuming half the energy delivered to the circuit, is insignificant in the interconnect.
The lesson here is that one application isn't like another. Wire gage is critically important if you're delivering power from a hydro plant to a city; it's critically important if you're driving a automobile starter; it's somewhat important if you're driving a speaker; and it's practically insignificant if you're interconnecting unbalanced line-level audio. Since we're not much concerned with hydro plants and Bendix gears here, let's go down a list of common audio and video applications and talk about what relevance wire gage has to these applications.
In speaker cable, barring some really odd construction practices, far and away the most important aspect of a cable is wire gage. Why? Well, think back a couple of paragraphs to that Ohm's law example. That's an extreme case, admittedly, but there, half the energy of the amp is burning up in speaker wire rather than being delivered to the speaker. Now, one might think, "what's the difference? The system will be a few dB quieter, but otherwise it'll sound the same." That would be true, but for one factor we didn't consider in our example. A speaker's impedance may nominally be eight ohms, but in reality it varies with frequency, starting high at low frequencies and falling. Consider what happens to our Ohm's law example now. If at one frequency the impedance is really six ohms, and at another it's ten, Ohm's law will distribute these different frequencies differently to the circuit. Where speaker impedance is low, more of the energy is absorbed by the cable; where speaker impedance is high, more of the energy is delivered to the speaker. The result is that excessive resistance in speaker cable will cause more loss of high-frequency than low-frequency signal; the system will sound differently from one wired with adequately-sized speaker cable.
Audio interconnects, as we've indicated, generally operate in very high-impedance circuits. Consequently, wire gage isn't really a meaningful factor in cable quality by itself. However, wire gage may have something to do with cable quality in an indirect sense--and that indirect sense points, somewhat counterintuitively, to a smaller rather than a larger conductor being desirable.
In high impedance circuits, capacitance becomes a significant factor in cable quality; capacitance is the tendency of the cable to store up a portion of the signal in itself and release it slowly, rather than deliver it immediately to the destination. Capacitance, in a cable with a single center conductor and an outer shield, will be determined by the outer diameter of the center conductor, the inner diameter of the shield, and the type of material (dielectric) that separates them. In an unbalanced audio interconnect, there are practical limits to what one can do to the inner diameter of the shield (cable needs to be of a size that's practical to attach RCA plugs to), and to the types of material that can be used as dielectric, and so the best way, at the margin, to diminish capacitance is to reduce the AWG of the center conductor. In our LC-1 Audio Cable, that's what we've done; the center conductor is 25 AWG, which is quite small, while remaining large enough to have good flex-life (i.e., not break when flexed) and to be susceptible to a solid crimp termination. We are sometimes asked why the AWG is so small, the unstated assumption being that a larger center conductor would be better; but even in a 50-foot run, the center conductor resistance is only 1.6 ohm, a vanishingly small value compared to a typical unbalanced audio input circuit impedance.
Analog Video, Serial Digital Video, and S/PDIF Digital Audio Interconnects:
Analog video interconnect circuits, whether they be modulated RF, composite, s-video, component, or RGB, are 75 ohm impedance circuits. Because all of these signals operate in the radio frequency range, skin effect increases the resistance of the wires in use, and because the cable lengths are often sufficient to make the cable's characteristic impedance (which is unrelated to its resistance--this is a function of the cable's capacitance and inductance) significant, the most important aspect of the cable design, from the standpoint of maintaining signal quality, is that the cable should have a 75 ohm characteristic impedance throughout the range of frequencies in use.
In long interconnection runs, the attenuation which results from, among other things, the resistance of the center conductor, will eventually become sufficient to harm signal quality; but for runs of moderate length, this is rarely a concern. Consequently, wire gage has some significance to signal quality, but is not the primary consideration. As with analog audio, however, there is a secondary sense in which wire gage is relevant to cable design; the cable's characteristic impedance is tied to its inductance and capacitance, and wire gage affects both of these because the center conductor must be in proper proportion to the other physical dimensions of the cable. If we stick a 16 AWG conductor into the center of an RG-6 cable where an 18 AWG conductor belongs, we wind up with our characteristic impedance too low; if we stuck a 20 AWG conductor in that same spot, characteristic impedance would be too high. So, while there may be no strong consideration affecting the specific choice of wire gage in most applications, it is nonetheless important that all of the cable's internal dimensions be in the right proportions to one another, and that includes the gage of the center conductor.
Parallel Digital Video (e.g., DVI and HDMI):
The dominant consumer digital video formats are HDMI and DVI. In HDMI and DVI, digital signals are run at bitrates which vary with resolution, and which can run quite high; currently, the highest HDMI resolution in common use is 1080p/60, which involves running signal at 1.485 Gbps. What does wire gage have to do with this sort of application?
As with analog video--and indeed, much more so, because of the very high frequencies involved--the really important attribute of a cable is its characteristic impedance. Here, we're not dealing with coaxial cable, but with twisted pairs, where characteristic impedance is much harder to control and is liable to change significantly from one inch to the next.
The frequencies in use here do an interesting thing to the significance of wire gage, which requires a bit of three-dimensional thinking to understand. In a 1.485 Gbps bitstream, our fundamental frequency is normally considered to be about half that bitrate, or 742.5 MHz, and because we're trying to convey some harmonics of that fundamental frequency to keep our bit edges from rounding off too much to be recognized by the receiving circuit, the bandwidth required to handle that is about three times that frequency, or 2.2275 GHz. Remember "skin effect"? Well, whether we're talking about 742 MHz or 2.2 GHz, skin effect at these frequencies is extreme. There is essentially no signal flowing through the middle of an HDMI cable conductor--it is all skimming the surface.
What that means to wire gage is that an increase in size is no longer as significant as it would be at lower frequencies, because the increase in wire surface area is proportional to diameter rather than to the square of diameter. Let's consider, say, the difference between a 24 and a 22 AWG cable. If we were buying 24 or 22 AWG wire for DC power, and wanted to know how much loss we'd see in a run, we'd be interested primarily in the cross-sectional area. A 24 AWG wire has a circular mil area of 404; a 22 AWG wire has a circular mil area of 640.4. Since DC resistance is inversely proportional to this area, this makes a big difference--the resistance of the 22 AWG wire is a bit less than 2/3 the resistance of the 24, for any given distance.
But if we're looking at skin effect, the picture changes. The cross-sectional area is practically irrelevant because the "skin depth" is next to nothing. Instead of cross-sectional area, loss to resistance is going to be inversely proportional to the amount of copper through which the signal actually passes--that is, it's going to be inversely proportional to the cable's surface area--or, speaking in cross-sectional terms, its perimeter. A 24 AWG wire has a diameter of .0201 inch, and a 22 AWG wire has a diameter of .0253 inch. Since the perimeters are simply these numbers each multiplied by pi, we can see the ratio of perimeters without doing that multiplication. The 22 AWG is "bigger" than the 24 by .0253/.0201, or a factor of 1.259. When we were concerned with area of the cross-section rather than perimeter, the ratio of circular mils was much steeper: 640.4/404, making the 22 AWG "bigger" by a factor of 1.585. Instead of the use of 22 AWG dropping resistance to about 63% of the 24 AWG wire's resistance, as happens at DC, it drops resistance only to about 80% of the 24 AWG's value.
Now, any reduction in resistance is good; the point here is simply to show that it isn't as good as one might expect. If all else were equal, one would expect 22 AWG HDMI cable to be useful for a distance of about 20% longer than a similar 24 AWG cable (this almost certainly overstates the advantage, because, of course, all else isn't equal. The longer run will show greater performance losses from other factors, including capacitance, crosstalk, skew and return loss).
The cable quality factors that really matter in HDMI cable are, primarily, impedance control on the TMDS pairs (which do the heavy lifting in the HDMI cable), and skew, which is a measure of the difference in electrical length of the conductors and pairs (by "electrical length," we mean the length of the wire, as measured by the time it takes a pulse to travel down the line; this may vary from physical length for a number of reasons, most but not all of which are related to impedance control). These parameters are notoriously difficult to control, and have nothing in particular to do with wire gage except insofar as it is sometimes easier to control tolerances in larger than in smaller cable. So, wire gage means something in HDMI cable; but it is not ordinarily the primary factor in measuring cable quality. A cable with superior return loss and skew can easily outperform a larger cable in a distance run.
Wire gage can be a meaningful factor in cable quality; but since it is very important for some applications, like speaker wire, only moderately meaningful for others like analog and digital video, and practically meaningless for still others, it's important to understand the demands of the application before making a judgment about cable quality based upon wire gage. When manufacturers fail to publish detailed specs on products, it can be a mistake to base relative quality judgments on whatever limited specs are provided, whether those be wire gage or something else.