ICE Cream Sandwich battery screen

By now you've heard a lot of scuttlebutt about the Ice Cream Sandwich battery bug, and how it's affected the Nexus S and the Galaxy Nexus.  Or not.  Either way, we've heard it, too.  Since we happen to have several of each phones possibly affected, I decided to do a little independent research of my own.  Tested were:

  • 1 Samsung Nexus S (T-Mobile) with the official 4.0.3 update
  • 1 Samsung Nexus S (T-Mobile) with Android 4.0.3_r1 built from source
  • 1 Samsung Nexus S 4G with the official GSM update ported over
  • 1 GSM Galaxy Nexus
  • 2 Verizon LTE Galaxy Nexii

Hit the break for an explanation of what the bug is supposed to be, and the results of our testing.

The Bug

Android OS

Many people think that some issue in the kernel or the code is keeping the phones awake when it should be asleep.  When your phone's screen goes dim, when it's not actively doing something like syncing, it should idle and use very little power.  Based on a number found in the battery settings (open the battery stats and choose Android OS from the list, and check the "Keep awake time") folks think something's amiss, and the phone just won't sleep.  As a first-generation HTC Hero user, I know all about phones that won't sleep.

There's something to be learned here, even if you ignore any possible bug -- the percentage you see next to what's using your battery doesn't mean what most people think it means.  Think of it as a pie chart, where the entries have to add up to 100 percent.  Even if you've only used 5 percent of your battery, something has used 100 percent of that 5 percent.  Confused yet?  A perfect situation, where a phone is in airplane mode in your sock drawer will have phone idle using 100 percent of the battery.  That's not going to happen, but it helps understand how the battery statistics work.  Just remember, no matter how much or how little your battery has been used, something(s) used all of that juice.  They are reported as having used it.

Our results

It's not that we don't trust second- and third-hand information from the Internet, but we don't trust second- or third-hand information from the Internet if we can verify it ourselves.  The results were interesting. 

Android Central

None of the Nexus S devices tested had any abnormal battery drain, nor did they have a large awake time from the Android OS.  It made no difference if it was the stock OTA, an AOSP build, or even a hacked GSM build on a CDMA/Wimax version.  An interesting aside: An AOSP build seems to have better battery life than the stock OTA update, but I digress.

Android Central

The GSM Galaxy Nexus also showed no signs of this bug.  With 17 hours of uptime, it was showing that the Android OS had used 22 percent of the total battery used, keeping the phone awake a little over three hours total.  That means the phone slept about 14 hours, and all is well.

Android Central

Of the two LTE Galaxy Nexii tested, one seemed fine.  About 12-14 hours of idle uptime, and 39 minutes awake time from "Android OS".  The other, however, had 7 hours of uptime, and 5 hours of awake time from "Android OS" listed in the battery statistics. 

Houston, we have a problem.

At first we guessed that in the case of the Verizon Galaxy Nexus, it may have something to do with freezing or removing Verizon's apps from the system.  They certainly affect the way the Android Market sees your phone, so it was possible that they did something else we didn't know or understand.  The Verizon Nexus without problems had the My Verizon app disabled, the Nexus with the problem had it running.  Don't go jumping the gun just yet, that's been ruled out after freezing it on the phone with the bug -- no difference.  Honestly, it seems like nothing we did made any difference.  Two identical phones, both running bone-stock Android 4.0.2, one has the issue and one doesn't. 

Since it's obvious there is a bug, and it's of the worst kind -- not reproducible every time -- we assume Google is doing something about it, as Paul Wilcox, Google Community Manager has this to say:

If you received an update notification a little while ago but the update isn't currently available for your phone, this is likely the result of Google pausing the update in your area while we monitor feedback. The Android 4.0 update is continuing to roll out around the world so your phone will receive another update notification when it's available again in your region.

All we know for sure is that the people seeing the issue probably really do see an issue, and have brought it to Google's attention.  Wilcox says pretty clearly there that "The Android 4.0 update is continuing to roll out around the world ..." Passed over for some, rolling out for others. But halted?

Regardless, independent developers are hard at work building their own kernel fixes, and you'll find those at the usual places, but we can't say for sure that this is the solution.  We just have to hang tight and wait for the fix, which we hope comes as quick as the EDGE-volume bug did.

4G signal wonkiness, or not

And about those LTE network signal bugs?  There are none.  Read this: Anandtech.  Those guys are real nerds and broke it down.  We can't say anything they haven't, and we're smart enough not to try.

More: Google support

 

Reader comments

Supposed Ice Cream Sandwich battery bug seems hit or miss, but something is up

56 Comments

I wonder if it's the same random bug that's been around all through gingerbread, where the backgrounder gets all over stimulated... Android OS usage > Display usage.. much discussion of it on XDA.

First off, is some background. The Verizon CDMA/LTE Galaxy Nexus (codename “mysid”) uses a combination of Samsung CMC221 and Via Telecom CBP 7.1 for LTE and CDMA 1x/EVDO connectivity, respectively. This is virtually identical (unsurprisingly) to the Droid Charge......

This was taken from Anandtech's Site...

Ouch.

Blazing.. @ 1,300 mhz on my Droid/GN Charge.. Awaiting KAL-EL...

I am baffled by people saying there is no LTE issue. I get drops to 3G and to no service at all continuously. Real-time download speeds using USENET show a clear discrepancy between my GNex, and a VZ Novatel MiFi - please explain that.

I can also show you a comparison using speedtest.net

???

Well maybe its just your area.

Do you get dropped to no service if you lock it to 3G and not use LTE?

I did a side-by-side comparison of LTE speeds on my GN and my Thunderbolt after reading that some people had a bug, and the speeds were virtually identical all the time, and both were a very satisfactory 20-25mbps down // 10-12mbps up, depending on coverage.

It is possible that this bug is similar to the battery bug in this article, and that it makes no fucking sense whatsoever.

I had this problem with ICS on my Nexus S (i9020a), running kwiboo's v3.

Using root explorer, I removed google plus entirely.

Worked like a charm, battery life is close enough to where I was on gingerbread that it's more or less negligible.

Another option (for those of you with Titanium Backup) would be to simply freeze google plus and see how that works.

Try at your own risk, ymmv.

Edit: forgot a parentheses :P

That's a reasonable suggestion to me. I've had some wild battery drain with the Galaxy Nexus, and the number one culprit the first day, but just the first day, was Google+. Since then it's been AndroidOS and huge amounts of wakelock.

One issue with Google+ that might be worth testing if you don't want to freeze/uninstall may be the frequency of updates or having instant upload turned on. I've got instant upload on all my other androids and don't see this kind of drain, though, and frankly, instant upload is the only thing I like about Google+.

I'm gonna try freezing it just to see if it helps battery life. Thanks for reminding me.

Anandtech's article is flawed.

1. They only tested one Nexus, from what I can tell.
2. They only tested 4G, not 3G.

The problem could be a subset of Nexus devices, and it could also be a problem with 3G, not 4G.

Personally, I REALLY hope that there is in fact a problem, because my Nexus drops calls and fails at sending texts multiple times a day. When I do manage to make a call, the quality is horrible.

I have also put my phone side by side with a Droid Charge at work (which is also a 3G/4G Verizon tower), put both phones to 3G only and then compared dBm. The Nexus is WAAAAYYYY lower, every time I have done it. The Charge would be -50 to -60dBm (good), and my Nexus would be -70 to -90dBm (very bad considering I'm basically next to the tower.

Something is definitely wrong. The Nexus and Charge have virtually the same internal radios.

That's what bugs me when people refer to AnandTech. They get a lot of stuff wrong (such as how mobile browsers report themselves). Their attitude seemed dismissive from the start of that article as if others haven't tried every little thing already to try to narrow down what the deal is.

This needs to be taken seriously. I know lots of people can bitch and moan, but in this case it happens to be some fairly knowledgeable people and in droves. Something's going on.

Did you guys read the article? He flat out says that 3G may have an issue and that he hasn't finished testing it yet. The article is solely about the idea that the Nexus gets a weaker LTE signal than other phones, not about whether it has handoff problems or weaker 3G.

First off, is some background. The Verizon CDMA/LTE Galaxy Nexus (codename “mysid”) uses a combination of Samsung CMC221 and Via Telecom CBP 7.1 for LTE and CDMA 1x/EVDO connectivity, respectively. This is virtually identical (unsurprisingly) to the Droid Charge

This is copied straight from the Anandtech article....

Ouch.

Blazing on my 1,300 mhz Nexus-Charge.. Awaiting KAL-EL.

If they have the same internal radio as in the charge than you should worry. The charge has many radio problems, static on phone calls, dropping the phone call in a weird way like the other person can't hear you but you can hear them. Which just happened to me once on my nexus. So I hope Samsung used a different one than the charge.

My Verizon Nexus right now.
54% left
8 hours on battery
42% Screen

24% Android OS
CPU total 40m 41s
Keep awake 2h 2m 14s

The rest is phone idle/cell standby/other apps

Something seems to be wrong. I just got my Galaxy Nexus, disabled 4g radio, checked battery stats and after unplugging charger, been up 1 hour and 47 min and top user of battery listed is Android OS at 57%...doesn't sound like it's going into sleep mode at all. Thinking about returning phone. Not sure what to do, maybe Razr is better. Any thoughts?

"top user of battery listed is Android OS at 57%"

That number, by itself, is useless because we don't know what it's 57% of. How much battery is left? What is the awake time of Android OS?

You didn't read Jerry's intro very carefully.

Top user of the battery being Android OS means nothing by itself.
Android OS didn't use 57% of your battery. It accounted for 57% of what was used so far, which is far less than 57% , I'm guessing closer to 1-2% total usage for the hour its been off the charger.

57% of 2% is nothing to be alarmed about.

Say you take the phone off the charger and put it down on the desk
for X minutes and find your battery drained by 1%.

Look at the usage statistics, and Andorid OS will be very high. It might even be as high as 80%, but its STILL only 80% of that ONE percent.

After you surf the web, and check a few emails, take some phone calls, that 57% you saw will shrink.

So my advice is don't return it yet, and stop fretting about WHAT is using the battery and just see HOW LONG its lasts. (And as usual, run it to exhaustion ONCE, and charge it FULLY afterward, and then see how long it lasts. That calibrated the charging chip and your results will be much more accurate.)

sounds to me like verizon f*cked up their kernel...I'm sorry the customers have to pay for their mistake, but from the company's standpoint....good!! I hope they go bankrupt because of this...making the whole country wait on the update because they are greedy and don't want google wallet. F*CK THEM!!!

again, to the verizon customers: I'm sorry you are suffering for this.

To verizon, the company: ROT IN HELL!!!

lol. I'm having a chuckle at your expense because you have such a strong opinion without knowing any of the facts.

Un-rooted GSM Nexus S here (UK) and my ICS battery life is a bit less than what I got on Gingerbread... Currently at 63% after 9 hours 50 mins and Android OS is certainly the most (26%).

When I first got my Nexus S, it used to be that Android OS used the most battery, then after a while (possibly after updates) it was always the screen that used the most.

Have you been using the Screen much during these 9 hours, 50 minutes?

If not, that still might not be out of line. If you've been making calls, and surfing the web, it might be a bit high.

What did screen take? What did Voice calls take?

My SGS2 is doing this as well today and reboots arent solving it. I suspect the latest Google+ update that added hangout video because i'm now seeing the messenger and google search processes continuously coming back though I kill them.. which is new. It's also one of the few updates I've received in the last day or so. My android OS process remains at 33% , and top user, even while I use other things. I'm about to uninstall it for testing.

I've been telling this to the one friend who finally got his Galaxy Nexus. The 3 other V customers are pissed at him BTW.
Hey you guys wanted ICS now. NOW NOW NOW. You can deal with some bugs.

You can have any two of the following three choices. Good, cheap, or fast.

It can be Good, and Cheap....but it won't be done fast.
It can be Cheap, and Fast....but it won't be very good.
It can be Fast, and Good....but it won't be cheap.

Pick your poison. Consumer need to learn to settle the hell down and just wait until its fully baked. The above folks I mentioned are worse then kids with hard core ADD. Alternatively carriers and handset manufacturers need to stop rushing crap out the door until it has had a solid period of QC. While Apple is the most notorious for this crap. Its pretty damn obvious HTC, moto, Sammy are all guilty of this.

Nothing is perfect out of the door. NOTHING. No matter how much QC you do there will ALWAYS be bugs and issues with new tech. The Nexus is not only a new hardware device, it also runs a new software platform. If you bought it thinking everything was going to be smooth sailing, you made an unwise purchasing decision.

The growth of technology and certain product release cycles aren't going to slow down because of bugs like this. You release a product fully knowing you will have to continually monitor and update. This is how tech works. You sit on a product for too long because you want to make it perfect, you lose. Those who do so will be left behind. See Duke Nukem Forever for an example.

I disagree with you.
This was a *FLAGSHIP PHONE* and cost $800-900.00 off contract with Tax & Accessories..
The so called *Super Phone* ICS,NFC.. Blah Blah Blah..

It's a FAILURE! Period.

$99.00 phones have all these issues ( Read the complaints about those that purchased the Galaxy Nexus on Amazon..)

Battery issues, 4G issues, NFC down, Late 2010 early 2011 hardware running the phone at this price point? Really?

It's half-baked.. with old parts.. and.. They knew it..

And the sad thing is, They knew how bad so many of you could not wait to drink the Kool-Aid

I have said it before and will CONTINUE to say it:

Blazing on my 1,300 mhz Droid Charge.. Awaiting REAL 2012 Technology whether it comes April-May or June.. KAL-EL will be worth the wait... The Galaxy Nexus has my 2010 Droid Charge Hardware in it!
(((First off, is some background. The Verizon CDMA/LTE Galaxy Nexus (codename “mysid”) uses a combination of Samsung CMC221 and Via Telecom CBP 7.1 for LTE and CDMA 1x/EVDO connectivity, respectively. This is virtually identical (unsurprisingly) to the Droid Charge))) This was taken from Anandtech's artcle on the Nexus.. Really? This is just cracking me up!

Personally, I'm happy I can get 10 hours of use even in spite of this bug. And it's definitely affecting me. Solid bar under "awake time". Android OS and Android System showing far more awake time than CPU total. Hopefully, they fix this bug so that my phone can SOAR!

Nice way to slip the signal issues in at the end and then say there are no issues. There is a lot of denial going on out there by those who refuse to believe their "Precious" could have a serious signal problem! Hopefully it can be fixed via software, but my spidey senses are tiggling that it's more a hardware issue (crappy radios)!

I find it funny people are complaining about signal issues. It's Samsung, their radios are well known to have issues. Is anyone surprised.

I am pretty certain that my battery bug is due to the phone always searching for signal. Sitting beside a Droid2on 3g, my nexus signal -141dbm (no signal) droid2 signal -79dbm. People can say what they want, this isnt my first smart phone, the reception sucks! I have been with verizon for years an had smart phones as long as they have been out, i have never had phone as bad as this. Dropped calls, text that wont send, no 4g or 3g in places where i should have great service. I have had the phone 4 days and it has dropped more calls than every phone i have ever had , added together! (going all the way back to before the origonal motorola flip phone!)

I don't look at signal percentages.
I look at the fact that my Blackberry got data in the Chicago subway every day and the Nexus gets none. I'm not mad, cause I know they'll fix it, but the issue is real.

"And about those LTE network signal bugs? There are none."

This made me laugh. Not because of some specific bug that may or may not exist, but because I'm a software developer and there are *always* bugs. I don't care who wrote it. There are bugs.

I'm a bit confused. I'm seeing 5 hours off the plug with 2 hours of wake time from android. Is this the buggy behavior?

My battery life has sucked since installing ICS on my Nexus S.

I went to bed last night with 100% and woke up at 29% with nothing running except an alarm.

My Nexus S works like charm using ICS 4.0.3
Battery life is good, can go more than 15 hours with average usage and Wi-Fi on all the time. Android OS never went more than 18%. Already been using this for two days already.

This is my hypothesis: there must be some apps that you guys install that not compatible with ICS 4.0.3 yet. looping/hanging/error somewhere.

The only battery issue I've had with some of the ICS ports for my Nexus S 4G are that I'd wake up to a dead battery. When I checked the battery usage, it was Google Services.. I notice the Sync Browser option always has an error.

For what its worth, I took my phone off the charger at 7am on 12/20. Went home made about a 20 min phone call and went to bed till about 6pm, phone was on wifi the whole time. Woke up played a couple levels of great little war game (about an hour)and browsed a couple websites and looked at some videos. Its now 1am in the morning and my battery is still at 40%. I do however have the extended battery but I'd say that the battery life isn't too bad. All on 3g no 4g in Yuma, AZ :(. SCREEN 30%, calls 13%, android os 11%, GLWG 11%, idle 10%, standby 8%, handcent sms 7%, wifi 2%

just around 30 min for the entire day. My battery didn't die until 2 am this morning and i was playing games on it from about 12 to 2

Ok , just oitside of an LTE area and just like a Bionic that at release i promptly returned , a new Galaxy Nexus and Razr both here on their first day are at times losing their 3G signal , only restart solves the issue and on the Razr that i bought for my wife , it will not regain its data connectivity !

Really frustrating. This leads me to wonder if there is an hprd issue in my area for the Bionic and now both of the GNex and Razr to have the same exact issue?

my work blackberry , orig Droid , wife's Eris have never once lost data connectivity .

Something must be wrong with the explanation. I am also suffering from a much higher battery drain. My phone runs no longer than 6 hours on battery. With 2.3.6. it was about 15 hours.

With 2.3.6. Android OS consumed less than 10%. Display took more than 50% all the time. Well, could be that ICS changed the way it calculates that, BUT a colleague has a HTC HD2 running an early version of ICS (cyanogen). He reports 80% display and less than 8% for Android OS.

These results are in contrast to your findings. Also when charging on USB the curve that goes up is even more flat than the curve that goes down, when disconnected.

To those that have a Nexus S with normal battery drain: How much % is consumed by Android OS on your devices?

After a factory reset the situation improved a lot.

Android OS is at 34% in contrast to 55% before the reset. However most significant change is related to the Exchange Services. Before the factory reset they took about 30%, now 2%.

It would be nice if the author specified what hey tried when saying nothing they tried worked. My os is also running far too much...and I'm within one week of my 30 days expiring. If this isn't solved with an update soon...I'm gonna swap phones at risk of getting the same issue. Or go to another phone altogether.

I had the same problem, within about 5 1/2 hours of very light use, the battery would be at about 10 percent from a full charge. I fixed it by by doing a Factory data reset. I don't know if it will work for everybody, but I know it worked for me.

My phone is a Verizon Galaxy Nexus, completely stock. It's been about 48 hours since the factory data reset, I couldn't be happier now.

After the factory reset what is the "keep awake" time for your android os? Did it decrease from what is was before the reset?

I can confirm that a factory reset also makes a great difference on a Nexus S. Yesterday mine was down to 54% after 2 hours. Today after the reset it's at 89%.

Keep awake is down to 17m 49s

Took a battery hit with ICS on my Nexus S (T-Mobile version). The past two days saw me get 7-8 hours of battery life. On Gingerbread, I used to get 15-18 hours on similar usage. Android OS is getting 18% usage (vs 20% from the display) and it kept my device awake for 1:58.13 during my last cycle.

There really seems to be something fishy going on here. I'll try to make some tests with the phone sitting on idle if I ever get the time to do so.

may also be a change in how something like android consumption is measured.

I got 25% Android OS consumption yesterday, up from much lower pre upgrade, but still had roughly the same amount of run-time, with about the same amount of battery left at the end of the day (about 40%, after 16 hours). Went to work, went home, did my usual calls, read & replied to mail, etc. Nothing scientific, but no obvious issues except a call that got dropped in an unusual place (that normally has service). That could simply be a tower having problems.

One other note - it's a GSM phone, not LTE.

There's a great little app in the market called BetterBatteryStats. It will show you if it is kernel wakelocks, app wakelocks, or erant apps keeping your phone from deep sleep. And if you just hate paying for stuff, you can get it on XDA free. I prefer to help support the dev because this really helped me get a handle on my problems...like Google+, Facebook, etc.