By Jacqui Cheng | Published: September 30, 2008 – 11:30PM CT
Count to five
"More bars in more places," reads the latest slogan from one of the largest mobile carriers in the US. It's plastered everywhere, from print ads to billboards to TV commercials. At least here in the US, it's pounded into our heads early on that bars = signal level = fewer problems overall. And it's kind of true. Generally speaking, the number of bars being displayed on your cell phone give a general idea of whether you have a good chance of making a call.
However, there are a number of variables that the phone takes into account when figuring out exactly how many bars to show you in the first place, and in fact, those variables can be (and often are) different among cell carriers, manufacturers, or even different devices from the same manufacturer. Put simply, the number of bars displayed on your phone does not necessarily equal signal level; rather, it would be more accurate to say that the bars are "correlated" with signal level.
Under most normal circumstances, the fact that bars are somewhat disconnected from signal level doesn't really matter. Most average people tend to use the same handset for years, and as long as it works, they're not paying attention to whether one particular phone on one particular carrier displays one more bar than another phone on the same or a different carrier. However, we geeks are not average people (usually). We pay attention to those tiny details, and we use that information to make judgments on how good or how crappy a phone or a carrier really is.
The topic of bars and signal strength has recently come to a head with the introduction of the iPhone 3G. Some of you have written into us expressing dissatisfaction with your iPhone 3G's number of bars when compared to other handsets you have used, only to discover how to put the phone into field test mode and see that it's reading the same level of signal that you're used to. Others have complained about experiencing dropped calls while having a full set of bars. As a result, we took a close look at what makes up all the links in the chain that connect signal level with bars, and why bars are not as reliable an indicator as you might think. Here's what we found:
Signal, cell towers, and you
Wireless engineers talk about signal strength in terms of power: dBm, or decibels to the milliwatt. At the very top of a cell tower, signal is transmitted at roughly +43dBm (20 watts) right at the antenna. However, that power diminishes significantly as you move further away, because the power density of a radio frequency (RF) signal in a vacuum follows the inverse square law in free space. Or, to put it in plain English, every time a signal doubles its distance away from the antenna, its power drops by a factor of four. However, cell towers don't really transmit signal in a vacuum. And, on top of that, it's being transmitted from an average of 100 to 300 feet above the ground, or 200ft to 5,000ft away, as the crow flies to your phone. Therefore, the problem of dropped power is magnified significantly.
By the time a signal transmitted from the tower hits your cell phone, it will be in the µW (microwatt) or nW (nanowatt) range. Since decibels are based on a logarithmic scale (they merely indicate a change between two numbers), the actual power change between 0dBm and 3dBm is a power of two—the same between 3dBm and 6dBm, and so on. A dBm of zero (which is next to impossible unless you suddenly find yourself climbing halfway up a tower) would be one milliwatt of power, and for every 3 dBm going down into the negatives, that power is being cut in half. So, for example, a dBm of -3 is about a half a milliwatt, -6 is a quarter, and so on. Once you get to -90dBm, you've reached the lower threshold of allowable signal level to connect to a tower.
You can see the number of dBm you're getting by putting your phone into field test mode. The method for doing this varies by phone and manufacturer, but generally speaking, it's not difficult to find with a little digging on the Internet. For the iPhone, for example, entering *3001#12345#* will let you enter field test mode and see your numerical signal level.
Notice how the bars in the upper left-hand
corner have been replaced by a number
There are a couple of reasons why the user isn't shown these numbers outright, though. For one, this can fluctuate by seemingly huge amounts with relatively little physical movement, which would cause confusion among users. This is due, in part, to a phenomenon called multipath signal fading.
A cell tower signal can be reflected by any sort of material that is partially reflective to electromagnetic radiation (i.e., inorganic material, like certain types of stone, metal, glass, or water.) Sometimes, a certain material can absorb a small part of a signal and reflect the rest of it off in another direction; this reflection means that all these signals are bouncing around and taking different physical paths to your cell phone's antenna. Different signals taking different paths means that some take longer paths than others. Since they all are traveling at the speed of light, multiple signal paths mean that some signals will arrive at your cell phone's antenna at different times. These signals interact with each other, canceling some parts of the signal out and amplifying others. The typical result is that the power density of signals in these multipath environments will fade in and out. In this respect, a typical urban environment plays hell with RF signals.
But multipath signal fading isn't the only thing that affects your ability to make clear phone calls.
In addition to signal, there's another variable that factors into your cell phone's reception: signal-to-noise ratio (SNR). SNR is affected by the environmental interference factors that we just discussed, i.e., how much interference there is in your raw signal, regardless of the strength of the signal. If you're in the downtown area of a city, and signal is bouncing off buildings, or if you're near a large body of water that reflects signal back, this can affect your call quality significantly. These are situations when you might experience fading while making a phone call on what otherwise seems to be a good signal.
This signal-to-noise value can be analyzed on the handset, and it's used in conjunction with the phone's reported signal level to formulate another scaled value—a call quality estimate—that determines the overall status of the phone and whether it's okay to connect to a tower. If this number is low enough, the tower won't let you connect, and hopefully, your phone will attempt to find a better tower within range. If the phone can't find a closer tower, then you're completely out of luck, and no tower will let you connect. At this point, most phones display a message like "No Signal" or "No Service," which isn't entirely accurate. It's more like "No Usable Signal according to the thresholds specified by the carrier," but that doesn't fit quite as easily into a ten-pixel-wide area in the corner of your screen.
This introduces a new conundrum for the cell carriers in charge of these towers. The carriers are able to set a floor for how low the call quality has to be before the tower won't let you connect. If they set the threshold too high, they run the risk of cutting off huge numbers of customers who may be just outside of the preferred range, and those customers will eventually complain about having no signal at their homes or workplaces. At the same time, however, setting the call quality threshold higher will also mean that there will be no dropped calls. The lower and lower this number is set, the more (and farther away) customers will be able to connect to towers to make calls. However, the level of signal and interference at that distance may cause them to have poorer call quality and more dropped calls.
Where the bars come in
The five-point scale, known as "bars," comes in when the cell phone itself looks at all these values being communicated from the tower and at the thresholds the tower would like the handset to obey. These values are ultimately what the phone and the tower use to determine whether the phone has a good enough chance of 1) connecting to the tower, and 2) maintaining a decent enough connection to conduct a phone call.
Note: There are a number of other variables taken into account at this stage, too. For example, your cell phone may be able to "hear" the tower just fine, but the tower may still have trouble hearing you. Mobile handsets are limited by the FCC to a transmit power of one watt, so the level of signal transmitted back to the tower is minuscule, compared to the 20 watts coming at you (and being diminished greatly as it reaches you) from the tower. This mismatch in signal strength is mitigated slightly by expensive, extremely sensitive receiver antennas at the top of towers.
Depending on the lower threshold of call quality on a particular tower, a phone will calculate a score on a scale of 1 to 5 to show the user how likely he or she is to be able to make a phone call. Let's say, for simplicity's sake, that the tower's call quality variable is on a scale from 1 to 100, with 1 being the best and 100 being the worst. If the tower's lower threshold is 80, and your phone is reading 65, your phone might display two bars, because it's near the bottom but not quite there yet. However, if the carrier's engineers set the threshold on the tower to a very low 100—that is, they'll allow cell phones with extremely poor signal quality to connect and make a call—a cell phone could still display two bars if it's reading something like 90, because it's still near the bottom, but not quite there yet.
"On the carrier's side, you can just crank that number all over the place to make the bars go where you want," former RF engineer and current Ars code monkey Clint Ecker points out. This variable lets engineers make the tower more accurately display the quality of the signal, and it's clearly in everyone's best interest to make sure it's as accurate as possible. After all, people's jobs, raises, and bonuses are dependent upon being able to provide expected levels of service. Still, that variable is capable of being tweaked in ways that misrepresent reality as well.
Manufacturers and standards (or the lack thereof)
The fullness of your mobile's five-point bar scale is also dependent upon just how much money the device's manufacturer sank into its RF discriminator (the hardware that filters usable signal out of the noisy RF spectrum). Manufacturers who use a more sensitive (and expensive) RF discriminator will be able to use signals that other handsets could never identify. Some handsets are known for having higher-quality discriminators than others. For example, Nokia spends a bit more on the RF front-end to help phones sort out signal and noise, and so Nokia phones tend to have a reputation for high-quality receivers. Other phone makers may spend a little less on the discriminator hardware, and so may not perform as well.
The bars scale is also dependent upon firmware from the manufacturer, OS vendor, and radio firmware provider. So the sensitivity of the bar scale, what will force it up or down, and what the baseline is of this variable on your phone is determined entirely by a string of programmers, none of which are following any specific standard for what qualifies as one bar versus two, three, or four.
Why the disarray? It turns out that there is no mention of, or even an attempt to specify, anything about "bars" or how call quality should be displayed to the end user in the GSM specifications. While manufacturers tend to stay within the same general range for each bar of signal, like women's dress sizes, they don't often match up exactly. This means that the value can (and does) vary between manufacturers, phone models, and even different firmware versions on the same model.
For example, Apple's recently-released iPhone 2.1 software update offers "improved accuracy of the 3G signal strength display." Notice that this doesn't claim to actually improve signal strength or your ability to make calls on the iPhone. Translated into layman's terms, it means "we cranked up the bars for the same level of signal." And that (cranked up bars, anyway) is what users have been noticing, too.
It's an art
If you've read to this point in the article, you're pretty well aware that the bar scale is a simplified meter for signal that is dependent upon many different factors, and those factors can vary by carrier, cell towers within each carrier, handset manufacturers, phone models, and software. This is why it's next to futile to directly compare two phones by the number of bars being displayed, or to compare the same phone from two different carriers, or even the same phone from the same carrier with different versions of the firmware.
But what you should take away from this isn't that bars are completely meaningless. They are, after all, there for a reason, which is to give you a "good enough" estimate of your ability to make a call. Just take the bars with a (very large) grain of salt when the service you get doesn't seem to match up with what the bars tell you, because the relationship between bars and call quality is much more of an art than a science.
Clint Ecker contributed to this report.