Studying iPhone 4S Battery Life semi-systematically
I'm testing Verizon iPhone 4S battery life during a variety of activities. I also have a general formula to estimate battery drain for active and standby use, and it would be great if people would use the formula on their own iPhones to check its accuracy. (See Conclusions).
Background: my fiancé's AT&T iPhone 4 seems to get dramatically better battery life than my Verizon 4S. We both bluetooth our music and podcasts in our car, she Facebooks, Foursquares, and Twitters while I Instacast and Instapaper. But subjectively at the end of the day, she's at 65% while I'm at 30%. This has bugged me to no end due to the very optimistic sentiment given for iPhone battery life - especially since my model appears to be the one that lost out.
But is the battery life that different, or it is really my own personal usage problem? So at the expense of my studying this summer, I have been compulsively studying my iPhone 4S's battery discharge based on a variety of factors. I would like to share what I've found and would like to encourage others to do the same.
1. Apple states the iPhone 4S should get the following performance (via their tech spec pages):
Standby - 200h vs 300h for the iPhone 4.
3G Internet - 6h vs 6h (iPhone 4).
Wifi internet - 9h vs 10h (iPhone 4)
Audio - 40h on both.
Talk time - 8h vs 7h (iPhone 4)
Even Apple admits iPhone 4S has somewhat less battery life than the iPhone 4.
Now for anandtech's ridiculous numbers (they themselves admit their battery test needs a revamp and that iOS is a little too good at it):
9.9 (iPhone 4S) vs. 6.5 (iPhone 4 Review) vs 7.7 (Verizon iP4 Review) vs 9.7 (iPhone 4, in 4S Review**)
11 (4S) vs 10 (4) vs 11 (Verizon iP4)
8.5 (4S) vs 7.5 (4) using the Verizon/CDMA talk times.
So their numbers seem a little odd for the 4, but it seems to come out to 10 incredibly optimistic hours for iPhone 4S, whether 3G or wifi.
I haven't seen anything like that, and having seen the discrepancy in iPhone 4 numbers across their different reviews, I think this may come down to them needing a new battery test. *It does seem that both Apple and anandtech say that the battery life is nearly equivalent on each carrier.*
2. For the past month I have wiped my 4S repeatedly to test different combinations of stock options. Since I wanted to compare to the iPhone 4 (which I reason with far less powerful hardware but the same battery, should have optimistically better battery drain) I have aimed for a configuration that require similar use as the 4 as it originally shipped (that's back on iOS 4, so a few features come to mind). Here's my preparation methodology.
- Wipe to stock and start clean (no restores from prior backups).
- Activate over wifi (3G won't be available until activation is complete, OR use iTunes.)
- Enable iCloud and Stores. (iCloud and push services are likely standard usage).
- Enable GPS (likewise).
- Disable Siri. (remember, comparing to iPhone 4. see drain below, OR research its consumption yourself.)
Once your iPhone is activated, sync to iTunes whatever you like - but do not restore from a previous backup, and do not enable Wifi Sync (if you want to compare to me, anyway). Then make the following adjustments.
- Disable Fetch (I only use push GMail and iCloud)
- Disable location based:
*Ads & Diagnostics (no reason to use these)
*Reminders (I like these, but large drain - see below)
*Time Zone detection (I've heard reports that early iPhone 4Ss had problems with this. I don't think it's a battery problem now, but traveling between states in the past month I've actually noticed sometimes I still have time zone problems. So I turn this off and do it manually).
*I have installed a minimal number of apps*. No Skype, no Pandora. (In particular I'm using Instacast, Find my Friends, and Pastebot). Nothing that could stick around for long periods in the background. (I disabled background features in Pastebot). *I found Tweetbot seemed to increase overnight drain and thus did not use it* - however I haven't repeated this to find out, and don't know the reason (its push notifications maybe?)
I measure the battery drainage in percentage compared to the total usage time for the task. I know that battery percentage change is nonlinear at the extremes of the curve but I'm willing to trust it for averages (especially the standby test, which I've checked many times now every night.)
a. Standby -
I check the charge, silence my phone, turn it face down, and don't pick it up overnight (usually an 8 hour span). In the morning I measure the drain. Underlying assumption is that this should be negligible given the very high standby times Apple states. (People who say their iPhone loses 20-30% just sitting around overnight should check this test against my methods and settings above).
b. Video streaming + Airplay
I stream (not predownloaded) the latest episode Vergecast over wifi to my Apple TV and measure percentage battery drain over time.
c. Audio streaming + Airplay
I stream a pre-downloaded podcast over wifi to the Apple TV. (I would have done this over a stream, but I just always download my audio podcasts and forgot about it.)
d. Phone call
Made in a good signal area (consistent 3+ bars) downtown with no wifi in use on my Verizon 4S.
e. Phone call (low signal)
Made in a poor signal area (2 bars or less) particularly, my house. Battery drain is always excessive on my phone calls and I wondered if this was due to bad signal at my location.
f. Wifi downloads (full wifi strength)
My goal was to test downloading a lot of stuff and to see how much drain pure data use has on the iPhone. (You would expect web browsing time to be *better* than this since browser data use is not continuous.) I chose to use Instacast to batch-download a bunch of podcasts for about half an hour. I also kept the screen on this entire time using auto-lock = never in settings. *I expected this to be the higher extreme for active data use battery drain.*
g. Location based reminders (ambient drain)
Similar to the standby test, but to see how much additional drain that the location/GPS reminders have when geofence awareness is active.
3. Here are my results (note that I'm not a statistician and don't have real averages for you.)
Drainages each night -
100 > 98% (2%)
92 > 91% (1%)
50 > 49% (1%)
49 > 47% (2%)
Consistently my overnight drainage with the above settings is 1-3%. If I am fully charged, 1-2% (remember battery level is nonlinear at the extremes so this may be equivalent). During the day I measure a steady 1% loss over 3-4 hour increments. This puts my iPhone 4S in the same ~300 hour standby neighborhood, so I am satisfied with this.
B. Video streaming + Airplay (wifi)
84 > 76% (1h, 8%)
12 possible hours estimated
Using Instacast app to my Apple TV. Seems to use less battery life than Apple's total video time on the device (10 hours), but streaming would seem less strenuous than local display use. It is worth noting: when streaming an internet video over Airplay, I think (but can not confirm) that the AppleTV actually does the network retrieval and merely keeps the iPhone alive to sync playback time and status, so this should not be a strenuous iPhone task other than not being allowed to go into true idle standby (even if the screen is off, this is marked as active time in Usage).
C. Audio streaming + Airplay (wifi)
66 > 58% (45 minutes, 8%)
9 possible hours estimated
Using Instacast, but this time podcast is pre-downloaded. This is interesting because now the brunt of the work is on the iPhone (which must transfer the local podcast file to the Apple TV) instead of the Apple TV for the network transmission. *This seems to take more battery life than Video Airplay alone*, but that makes sense considering the iPhone wasn't actually doing much during Video airplay (due to the Apple TV handling the entire video streaming process). I am satisfied here, since I always expected Airplay transmission to be a battery drain.
D. Phone call (good signal, 3+ bars)
91 > 86%, 22 minutes (5%)
7 possible hours estimated
Not much to say here. I'd like a longer call to be sure but I didn't have assistants, so I actually had to make a real phone call to a friend. Being downtown my iPhone also wasn't running wifi, so it was doing the bare minimum required (call transmission). This seems in between the Apple suggested talk times for the iPhone 4 and 4S, so I'm happy with this.
E. Phone call (poor signal, 2 bars or less)
58 > 29%, 1 hr 27 minutes (29% battery loss)
5 possible hours estimated
58 > 51%, 23 minutes (7%)
51 > 45%, 22 minutes (6%)
45 > 37%, 23 minutes (8%)
I always feel disappointed with talk times in my home and this may be the reason why. There seems to be a definite difference, in comparison to the good signal phone call, in battery drain at my home. (Then again I forgot that I also have wifi active in my home - perhaps this extra active connection takes more toll on the iPhone than you'd think - and something I will have to remember to disable in the future).
F. Wifi downloads
75 > 66%, 25 minutes (9%)
4.5 possible hours estimated
I wanted to see what drain heavy data use would take on the iPhone over wifi. I set Instacast to sync and download 8 new episodes for a variety of podcasts (Instacast also likes to download 2 at a time). Although extrapolating this one is uncertain I think it's a good ballpark figure for heavy data use and the resulting battery life to expect.
G. Location based reminders
Overnight drain 50 > 36% (14%) over 8 hours
Having activated this solitary option on a nearly stock iPhone after several nights of 1-3% standby drain, this proved to me that Location Reminders probably take too much idle drain to be worth it. If you're working the late shift overnight, you're essentially losing 15-20% of your iPhone doing absolutely nothing.
H. Siri - effect on ambient drain
Overnight drain 17 > 9% (8%) over 8 hours
*I have not repeated this test.* But merely enabling Siri without using it seems to increase the idle drain of the 4S. I decided this wasn't worth cutting my standby time to 100 hours or less although for day to day use, Siri drain likely depends on how often you use it. (I abused it when I first got this phone, but I'm thinking that may have been one of the reasons I was not impressed with the battery life). It's a great feature, but maybe it will be even better in iOS 6 or the new iPhone.
All in all, I think I'm still impressed with the battery life of my 4S. I think the difference between 4 and 4S battery life seems to be mainly down to usage (4S owners are more likely to use a dual-core processor and make tons of GPS-aware requests through a voice-recognition interface, for example). I found some interesting things.
- Signal strength seems to have a bigger effect on talk time life than I expected - Apple's numbers may account for optimistically good signal rather than average/subpar.
- Audio airplay takes more life than Video *streaming* airplay, which seems to delegate the actual streaming to the Apple TV. (I'd be curious to see what effect pre-downloaded Video Airplay, with the iPhone transferring the video over the network, would have).
- Location/geofence functionality seems to have a high standby drain that I think probably isn't justified, even if GPS reminders are really handy (and I've set enough up to like them).
- I was genuinely surprised that I could get decent standby overnight time on my 4S. I think this really depends on settings (using *Fetch every 15 minutes* with the iPhone's minimum wake time of 5 minutes each fetch would result in nearly 2 hours of active use just over night.
Based on Apple's numbers and confirmed by my iPhone, I think it's reasonably accurate to estimate the total drain using the following formulas.
% expected standby drain = standbyTime (hours) / 3
% expected active drain = activeTime (minutes) / 5
This estimates about 1% drain for every 3 hours of standby, and 1% drain for every 5 minutes of active use (normal smartphone tasks like email, browsing, app use); and therefore about 8 hours of total usage for the iPhone 4S. (Please note that phone calls and media use have different drain rates as seen above - that formula is only what I've seen for non-media/call use).
I find it handy to compare my actual drain to these expected formulas and see how accurate they are. I'll rework them if I find better numbers.
How do your iPhones match up? I don't expect anyone to collect tones of data, but I'd love if you would compare your usage and active drains in General/Usage with my expected drain formula.
** I don't understand why anandtech has 2 such radically different times for iPhone 4 3G use.