Your laptop is lying to you

Publication date: 28 July 2008
Originally published 2008 in Atomic: Maximum Power Computing
Last modified 03-Dec-2011.

Your mobile phone is lying to you. So is your laptop. And not just when they whisper to you in the night.

Mobile phones, and most modern laptops, have signal strength and battery life displays. One or both of these displays has probably been the focus of all of your attention at one time or another.

Neither display is actually telling you what you think it's telling you.

The signal strength bars on a mobile phone or laptop do, at least, say something about how strong the local signal is.

But they don't tell you the ratio between that signal and the inevitable, and often very considerable, noise that accompanies it.

A high-signal strength is like your friend shouting at you. A high noise level at the same time is like your friend shouting at you while you both stand in the front row at a rock concert.

(Actually, for phones and Wi-Fi, it's often more like your friend shouting at you while you're both attending a rock concert in an echo chamber.)

So if you've ever wondered why your phone can work fine with one rather unsteady bar of "signal", or be noisy and cut out when it's got five bars, that's why. The ratio of signal to noise can vary wildly from second to second, in many ordinary mobile-phone situations.

It's within the bounds of possibility that there's some phone out there whose "signal bars" actually do give a realistic estimate of the actual signal quality (rather than mere quantity), but you shouldn't count on that display telling you anything really useful if you're not actually on a call at the time.

And the exact same applies to Wi-Fi signal strength displays, though it's easier to see signal-to-noise ratio (SNR) figures on a computer - just run a utility like NetStumbler.

Signal bars also don't tell you anything about how well a cellular phone, or wireless network, base station can hear your phone or laptop. They only tell you how well your device can hear the base station.

(Some phones may give more realistic signal figures when they're actually connected to a base station because you're making a call. Then, they of course actually know how good the connection is, because they're sending and receiving data. If you've actually successfully made a call, though, you (a) already know that there's an adequate amount of signal and (b) probably have the signal-strength display pressed firmly to your ear.)

This explains many of those times when you seem to have a "bar" or two of Wi-Fi signal, but can't actually connect. If your laptop has too poky an antenna, it can't make itself heard to the access point.

Battery meters are even more fun.

It's possible to monitor the charge in a battery - any kind of battery - with considerable accuracy. There's a certain amount of guessing involved, because the "fullness" of different battery chemistries doesn't necessarily map well to terminal voltage even if the battery's powering an unchanging load. But modern "smart" batteries with little chips on them that keep track of how long they've been lasting lately really do work pretty well.

You may have noticed, however, that your mobile phone seems to spend an awful lot of time with its battery gauge saying it's full, or at least almost full.

Then, once you get to the half-full mark, the battery seems to go flat surprisingly quickly.

This phenomenon isn't as obvious with today's low-power-consumption phones than it was back in the days of 20-hour standby times, but it's still pretty hard to find a phone that doesn't do it.

There are two reasons for this. They are both bad.

Reason one: A battery that stays (apparently) full for a long time makes a phone look good. Even if it doesn't actually deserve to.

Reason two: When your phone still (apparently) has lots of charge left, you're more likely to use it. People who think their phone's going flat will make fewer, and shorter, calls. And that makes phone companies sad.

That's right - this is yet another example of the Curse of the Marketing Department. Both phone makers and cellular service providers want you to think that your phone is still pretty much full of charge even if it's almost half empty. For this reason, many of them tweak the charge meters to overestimate the remaining charge.

This is a particular problem in the US market, where most mobile phones are not just usually locked to one provider, but also customised for that provider, with specific firmware that can very easily include a battery meter that has an even larger Lie Factor than usual.

(This isn't the most spectacular achievement of the US cellphone companies. They also somehow managed to talk people into buying SMS message plans that charge you not only when you send a message, but also when you receive one. But it could be worse; if you're a Russian with a cheapo prepaid phone, you'd better be ready to be switched to half-duplex "walkie-talkie" mode, or dropped altogether, if someone who's paying more per minute needs your network bandwidth.)

Most modern cellphones have processing power to spare, and could easily have very clever battery-monitoring. They could tell you remaining capacity as a percentage, maximum capacity in watt-hours, minutes of battery remaining at current drain level, and so on.

What do you usually actually get, though? A phone which, whenever you plug a charger in for a few minutes, will upon unplugging proudly display a completely fraudulent full set of battery-capacity bars.

(Digital cameras are often not much better. It might be possible to dig through menus on your consumer-market camera to find a fine-granularity no-bulldust battery display, but it's not at all likely. Note, however, the CHDK firmware for Canon point-and-shoots, which adds tons of features, of which an honest battery display is the very least.)

The custom-rolled firmware from the US cellphone companies is also, of course, likely to tell fibs about signal strength. The telephony companies quickly realised that it's much cheaper to sell phones that show lots of signal all the time than to actually bother putting up more of those expensive cell towers. And, once again, the signal strength is not the same as the call quality. Lots of signal plus lots of noise is not as good as a little bit of signal plus almost no noise at all.

There's one other commonplace situation in which a "fullness" display lies to you - car fuel gauges. This time it's a safeguard, not a scam; the determination of some drivers to fill up as seldom as possible has led to a cultural convention in which "E" on the fuel gauge almost never actually means the fuel tank is empty.

Once an industry has made this sort of decision - for cars or for mobile phones - it's hard to change. If Car Manufacturer X boldly decided to make E really mean E for their cars, people would still expect that it didn't. Result: A lot of unhappy Manufacturer X car owners trudging down the road with jerrycans. Not to mention diesel cars with air in the fuel lines.

And if Phone Manufacturer X locked down their firmware so that the battery and signal displays actually gave an accurate idea of the remaining power and expected call quality, many customers would think those phones went flat faster and couldn't get as good a signal.

If laptops had as much battery life as phones, I bet we'd see "bar inflation" happening to their battery gauges, too. Manufacturers can't really get away with this scam when total run time is only a few hours, though - and modern laptops all have "smart" batteries that let the operating system quite accurately estimate how long the battery will last at the present drain level. So this is, at least, one piece of nonsense that laptop users have not yet had to suffer.

It's still worth noting, though, that there's no standard for what "two bars of signal" is supposed to mean, for a mobile phone or for a laptop.

It varies from device to device, and from firmware version to firmware version. On a laptop, signal displays can even vary depending on the monitoring software you use.

And remember: Measuring signal without measuring noise is like only counting the goals your team scores.