Friday, June 18, 2010

Guide to Managing Your EVO Battery: The Approach and Control Tests

In order to truly understand my device and my unique situation, I needed to control the variables. By systematically changing the variables, I can determine the impact of each on battery life. When I first received my EVO, I setup my gmail, added all the widgets I could, personalized my screens, and started downloading apps (63 to be exact).  And my battery life was horrible.  The problem was, I introduced too many variables and after three days of stuggling and trying every tip and trick I could find on the web, I was no closer to solving my battery life issues then when I started.  It was time to start with a clean slate so I wiped my EVO.

Baseline the Battery

To baseline the battery, I started with a simple configuration and a somewhat controlled environment to limit the variables.  I assumed Google designed Android with a couple of basic functions in mind, using your phone as a phone (go figure, right?) and integration with your email/gmail account.  My first configuration consisted of setting up my gmail account, turning off all radios except my 3G radio and my cellular radio, downloading an app called "Battery Indicator" to better monitor actual battery usage, and configuring a few basic settings.  Specifically, I disabled syncing to news and stocks, limited weather syncing to once every three hours, disabled auto sign-on of Google Talk (EVO automatically configures Google Talk when you setup your gmail account), set the screen to auto brightness, and removed all widgets except the HTC time and weather widget.

Test 1 - Minimal Usage, Minimal Syncing, Cellular Standby

I ran a 4 hour control test, only turning the screen on every 30 minutes to check the battery status.  I used the power button to lock the phone when I was done using it.  A few emails came in, weather updated, and not much else.  Over that time period, my phone battery decreased an average of 3.5% per hour.  That translates into about 30 hours of battery life.  Not horrible, not great. 

Given my configuration and usage pattern, I knew there were two primary variables that were impacting battery life, network usage for gmail and weather syncing and the 3G and cellular radios.  Given the number of emails I received, I didn't think network usage was a major battery drain.  That left the radios.  I only had 1 or 2 bars of signal so I did some investigating.  Using the "Battery Use" screen, I was able to see that my "time without service" (TWS) was about 80%.  That meant my cellular radio was working overtime, so 3.5% per hour didn't seem as bad.  Actually, I was feeling pretty good at this point since just that morning I lost 15% of my battery in the first hour.

Test 2 - Wireless

It was time to introduce a new variable to see if it had a significant impact on battery life.  For this test, I enabled the wirless radio, connected to my local network, and ran a two hour control test using the same usage pattern as the first test.  To my surprise, I actually saw better results.  My battery decreased only 2% per hour.  I suspect this was the result of a stronger wireless signal and/or more efficient transfer of data over the wireless radio than the 3G radio.  At 2% per hour, this translates into 50 hours of battery life.  Needless to say, I was pleased and maybe even a little excited with these results.

Test 3 - Bluetooth without Pairing

I wanted to test the impact of turning on the Bluetooth radio without connecting to a Bluetooth device.  When I get into my car I don't want to have to fiddle with my phone to get it connect to "hands free".  With my Blackberry Storm, I found that I could leave Bluetooth on even when I wasn't using it and it didn't have a perceivable impact on battery life.  During the next two hour control test, I confirmed these results.  Once again, my battery decreased approximately 2% per hour.

As a side note, I found the "Battery Use" screens to be inaccurate when determining Bluetooth power usage.  My test indicated no impact while the "Battery Use" screen indicated it had the second largest power consumption behind "Cell Standby".

Test 4 - GPS

Similar to the Bluetooth test, I wanted to test the impact of turning on the GPS radio when not in use.  Once  again I found minimal impact to battery life as the battery decreased approximately 2% per hour.

Test 5 - Call Time

Most of us still use our phones to place/receive calls.  I wanted to see if my EVO was going to get anywhere near the 6 hours of talk time the spec sheet said it would.  This test was straight forward.  I placed a 30 minute call and measured the decrease in battery.  Once again, I was pleasantly surprised.  The battery decreased 8% which translates into 6 hours of talk time.  Subsequent measurements have confirmed these results.

Test 6 - Simple Gaming

I like to play Soduko on my smartphone.  My next test was to install "Daily Soduko" and measure the resulting battery usage.  This introduced a couple of variables, CPU usage and screen brightness.  For these tests I used auto brightness.  I ran two short tests, approximately 15 minutes each.  The battery decreased 5% during each tests or 20% per hour.  Several subsequent tests have further confirmed these results.

At some point, I intend to rerun these tests with the screen set manually to 33%.

Test 7 - Live Wallpapers

This is the one you've all been waiting for, right?  I couldn't wait to get to this one because I absolutely love the Live Wallpaper feature and I didn't want to have to live without it.  However, if it was a major battery drain, I'd have to shut it down.  I ran two tests, another 2 hour control test with little to no usage to see if the Live Wallpaper would run in the background and drain battery.  This does not appear to be the case as my battery once again decreased 2% per hour. 

Next I ran additional gaming tests to see if Live Wallpapers impacted battery life while using other apps.  My results were the same as those in Test 6.

Conclusion, Live Wallpapers have minimal impact to battery life.  The only time they increase power consumption appears to be when you're on the launcher screens and the time I spend there is minimal.  If you're usage patterns are different, you may want to run some addittion tests to check impact.  As for me, I'm keeping my Live Wallpaper and I'm excited about it! 

By the way, these results do no apply to the Google Maps Live Wallpaper as this Wallpaper will access GPS and network resources to update the map and I suspect it will do this in the background even when you're not using your phone.

Test 8 - Minimal Usage, Strong Cell Reception

I have full bars at work, so I repeated Test 4 over a 4 hour period to verify my theory that weak cellular signal was the primary cause of my somewhat high battery usage in Test 1.  For this test, Bluetooth, GPS, and 3G were enabled, wireless was off, Live Wallpapers were on, gmail sync was on, and weather updates were set for every 3 hours.  The results were astonishing!  Less than 1% decrease in battery per hour!

Test 9 - Downloading Apps and Web Browsing

These were probably the most unscientific tests I performed.  It's difficult to measure exactly how much data is transferred and the impact of signal strength.  Suffice it to say, both activities burn battery fast!  I did find that a good wireless connection can reduce battery usage considerably compared to performing the same tasks over 3G.  However, I could easily burn 10-15% of the battery in as little as 10 minutes.  If these are activities you need to perform frequently, you might consider plugging in or purchasing an extended battery.

Next: Summary and Next Steps ...


  1. So is there a setting to change when your at an area with weak signal that will help battery life?

  2. I would check your bars. If you're not getting any bars, you might decide to go to airplane mode or disable your mobile network. I add widgets for both of these so I can quickly toggle them on or off. Since I don't want to miss calls or texts, I'll do this when I know I won't be able to receive calls or texts anyway.

  3. thank you for this site! Indeed you are a geek, but i appreciate all the testing you did, & i am trying it out on my phone now!