4 June 2016, 08:22
My MacBook Pro’s reached that time in its life when the battery is diminishing slightly in capacity, so I’ve been looking into how to get the most possible usage between charges. There’s a lot of received wisdom about battery life and, as with so much computing advice, much of it is bullhonkey. Indeed, my research has found some surprising factors can dramatically affect battery life.
I used as a testbed a mid-2015 15in MacBook Pro (2.8GHz i7 CPU, 16GB RAM, 512GB storage). To measure power consumption in wattage I had two options. The first was to use a mains-power plug-in wattage meter. This would’ve involved disconnecting the battery within my MacBook Pro in order to remove it from the power-usage equation – simply not an option.
Instead, I chose to use Hardware Monitor, an app that amongst many other things monitors the Mac’s internal sensors in order to report a wattage figure. Alas, we have no real way of knowing if this is accurate or representative (I contacted Hardware Monitor’s developer and even he doesn’t know), so the test results below are decidedly unscientific. I advise you to get the app and test yourself.
The MacBook Pro’s magsafe power lead was disconnected during testing so that the battery was discharging. The wattage figure fluctuated massively at all times because of OS X’s seemingly infinite number of background tasks, all of which drew power as they did things like access the storage or network, so I simply stared at the wattage readout for 20-30 seconds for each measurement and made a note of the lowest figure reported. Nearly always the reading would stabilize for several seconds at this figure, or within 0.1 watt.
What we’re looking for
The amount of electrical power used by a device is measured in watts. More watts used = battery life used-up. Everything your Mac does involves the use of power, of course. My MacBook Pro tended to have a “resting” figure of between 6-10 watts when nothing was happening on the system. In attempting to eke out the most battery life, I was looking to reduce this figure as much as possible. Even one or two watts more energy usage can burn through battery life more quickly.
Apple reduced the energy demands of portable displays by switching to LED backlighting a few years ago. However, the screen is still believed the biggest consumer of power while you’re out and about. My tests supported this statement, but it isn’t necessarily as straightforward as “less brightness = more battery life”.
See the graph below. To create the measurements I started with a completely dark screen, then tapped the screen brightness button to increase the on-screen display (OSD) by one notch for each measurement I made.
As you can see on the graph, the lower brightness levels according to the OSD don’t vary much in terms of energy usage. Up to about four notches, which is perhaps the acceptable lowest setting for a moderately-lit room, and there’s little difference. After that, however, the wattage figure ramps up quickly. Beyond the mid-way point of eight notches on the OSD and you’re essentially increasing power usage by around one watt for each notch.
The moral of the story: Keep brightness beneath 50% but don’t punish yourself by setting it to a single notch so that you can barely see anything. Four notches seems the best choice.
This was a surprise: Setting keyboard backlighting brightness to full increased power consumption by a mere 0.4-0.5 of a watt. Increasing keyboard brightness to 50% reduced this figure only by 0.1-0.2 watts.
In other words, keyboard brightness has much less impact on battery life than you might think, although if you’re in for a long spell between charges then keep it off. Turning it down makes little difference – you might as well have it on full, or off.
Turning off Bluetooth and/or Wi-Fi
The first big surprise of my tests was that turning off the MacBook Pro’s Bluetooth and/or Wi-Fi function made no difference to power usage (0.1 watt, which is negligible). This is perhaps because Wi-Fi and Bluetooth chipsets are insanely efficient nowadays. It’s important to note that I’m referring here to merely having these features turned on or off. If they’re actually in use sending or receiving data then there will likely be significant power usage,
but I had no way of discovering exactly what this was because I’m unable to measure the power consumption of Wi-Fi or Bluetooth traffic discretely. UPDATE: I found a way of measuring Wi-Fi power usage when receiving data. See bottom of this piece.
However, you should still turn off both Wi-Fi and Bluetooth if you don’t need them because it’ll prevent the myriad of Mac background processes from using them, thus saving power not only by avoiding Wi-Fi and Bluetooth traffic but perhaps sending the background processes temporarily to sleep (and don’t forget that all tasks and apps on your Mac eat power). Your Mac is virtually always syncing iCloud data, for example, or seeking nearby devices for handoff tasks.
However, having Wi-Fi or Bluetooth switched on, per se, isn’t an issue. Myth busted.
You can skip this bit if your Mac doesn’t have a discrete GPU. If it does, keep reading…
Some MacBooks have both integrated Intel graphics plus a discrete AMD or Nvidia GPU. There’s no questioning the received wisdom that discrete (GPU) graphics eat battery life far more than integrated graphics – they do – but you can use an app like gfxCardStatus to force the use of integrated graphics only. This is free of charge. Just click its menu bar icon once installed and make your choice.
You might have read that some apps can’t manage without a discrete GPU. Once again, this is a classic example of computing wisdom bullhonkey. For example, Photoshop will run just fine with integrated graphics. It’s just that a tiny minority of specialist tasks like applying complex filters will complete much more quickly if the discrete GPU is used. On the other hand, you simply won’t see any performance difference for everyday tasks. Let’s rephrase that just in case it didn’t go in – it’s extremely unlikely you “need” a discrete GPU.
gfxCardStatus also reports when OS X decides to switch to the discrete GPU, and thereby shows just how seemingly random OS X can be in switching. Some web pages appear to cause OS X to switch, for example, even though forcing a switch back to integrated graphics via gfxCardStatus makes absolutely no difference to the browser performance. Using the Photos app always switches to the discrete GPU, although I can’t think of a single reason why this is necessary. (Ergo, here’s a tip: Don’t use Photos if your MacBook Pro has a discrete GPU and you’re using battery power.)
Do you know what? I’m inclined to think the whole GPU thing is bullhonkey too and little more than a sales pitch. Sure, if you use graphics intensively for things like gaming or rendering HD video effects then you might feel the pinch but for all day to day tasks it’s a waste of silicon.
The next huge surprise in my tests was that scrolling webpages using the trackpad causes a massive jump in power usage. In fact, all graphically-heavy tasks do so, such as opening Mission Control or Launchpad.
The difference is that you might open Mission Control once or twice every five minutes, if that, although you’ll be scrolling virtually constantly.
The solution if you really want to maximize that battery life is to scroll web pages using the Up/Down cursor keys. This is jerky, of course, but simply doesn’t cause the huge wattage spike. Even keeping your finger on the Up/Down cursor key does NOT cause a wattage spike.
Other interesting things and advice
Attaching a USB stick appeared to increase power consumption by one watt, which I thought kinda interesting, but that was the end of my tests. If there’s anything you’d like me to look into then please let me know in the comments below.
One other thing: The Safari browser uses a lot less energy than browsers based on the Blink rendering engine, which includes Google Chrome, Opera, Vivaldi, and others. Therefore, the use of Safari is advised when you’re out and about, at least. (For what it’s worth Firefox is roughly between Safari and Blink-based browsers in terms of power usage, so isn’t advised either.)
Notably, most if not all of the built-in OS X apps are more energy efficient than third-party options, although it’s only with an app like a browser that’s used near-constantly that you’re likely to feel any effect on battery life. However, learn to quit any app that you’re not currently using when on battery power, and quit all unused browser tabs. Switching to the old days of having just one web page open in your browser is a battery-friendly choice.
UPDATE: I managed to test the power used in Wi-Fi data transfer using the curl command in the Terminal to download a large file, while closing all other apps. A “resting” figure of around 6.5 watts jumped up to a fairly consistent 15.2 watts when the download began, maxing out my 70Mbps DSL connection. I suspect the curl command’s own energy usage is negligible, so that’s around 8-9 watts required to download data (Wi-Fi connection data for the curious: 5GHz, WPA2, Noise: -93dBm, RSSI: -47 dBm, Tx Rate: 975 Mbps, so probably 802.11ac in use). So, yes, turning off Wi-Fi if you can is a good idea.
Leave a comment...