quote: WARNING: TL;DR INCOMING This is intended to put an explanation and "face" to the problems iPhone 5 users are facing. It also serves to compare the iPhone 4, 4S, and 5 from a comparative power consumption standpoint to explain the noticeable differences. The numbers and figures are not guaranteed to be 100% accurate, and I'm sure you'll find an error or a miscalculated capacitive load. Just trying to put minds at ease and give you a face to your battery life's enemy. Feel free to ignore all of this and enjoy your life. - Edsger Wybe PirateZac Dijkstra
Battery life seems very good
The battery is actually much better, but the draw of LTE kills battery life much quicker. Also, the change in processor power (single to dual core)changes things when comparing the 4 to the 4S/5, as well, which is why we've seen outrage in both releases. I actually figured these out the other day to get an accurate measurement. I found the capacitive load (CL) myself, but it may be off a little Also, this only takes into account serialized instructions, and it does not account for multiple threads or parallelized instructions. NOTE: nF(nanoFarads) = 1.0e-9 F iPhone 5
: 3.8V, 1440 mAh, 1.2 GHz(Dual)
CL: approx. .31452 nF (3.1452e-10F) iPhone 4S
: 3.7V, 1430 mAh, 1.0 GHz(Dual)
CL: approx. .3835 nF (3.835e-10F) iPhone 4
: 3.7V, 1420 mAh, 1.0 GHz(Single)
CL: approx. .38714 nF (3.8714e-10)
So, as we can see, the capacitive load is a bit less, and the voltage and the mAh are higher in the iPhone 5. The iPhone 4S actually looks to be be only marginally better than the iPhone 4, but what sticks out is a jump in battery life while making the switch to a dual core processor. This is one of the reasons their overall capactive load is so similar. When taking this into account, the clock speed is treated as handling serialized instructions at the max speed and voltage that they can handle. Because the 4S has a similar voltage and marginally higher mAh, it has a similar battery performance because the two cores are not taken into account so of course in theory, they should have had a similar battery life, right?
The problem is that we DO have two cores, and we have to acknowledge the trade offs of that. On one hand, if a portion of the instructions are parallelized, we can process two streams simultaneously to cut down on time spent processing those portions. That seems great, but we have two issues. The first is that not all instructions will always be parallelized so some of them must be run serialized, only utilizing one core. The second is that with the usage of a second core, we up the usage is power during that period. While dual-core may be more efficient in some cases, there are a ton where it consumes more power. So even though we spend less time processing instructions, we still have to deal with serialized instructions and also the case in which the dual-core sucks up more juice. There are also other factors that play into it, but that plays a factor in the 4S having a seemingly much worse battery life than the 4, even though they have similar battery specs.
Now, we move up to the iPhone 5. It's got much less of a capactive load, which is largely due to the use of the new Tri-Gate transistor, also called 3D transistors, technology that allows vertical transistors. This allows them to pack more transistors than they could traditional transistors into a similarly sized area, and they operate at lower voltages with less leakage. This should mean the battery should last much longer than its predecessors, right? Not quite. The problem is that the iPhone 5 has a much better power specs because it needs them for what they've added.
First, we have the A6 chip. It's got two-cores with a clock speed, from what I've seen, of around 1.2GHz. Obviously, this requires more power to operate, and with an mAh rating not much higher, that'll be an issue. The second major factor is the inclusion of LTE. In an AT&T test, the energy drain is proven to be higher for LTE. With 3G, the power drain is static. It's always going to be the same. With LTE, the drain is heavily dependent on the size of what's being transmitted. LTE has two DRX states for data transmission. When data is received, it goes into the short DRX state which consumes high power while receiving it, which is why we get mobile porn much faster on LTE.
The issue isn't with that, but instead, the problem lies in the with behavior of the long DRX state. The radio switches to the long DRX state, its tail state, after data is received and waits for more while remaining in high power mode. If it gets more, it loops into the cycle. Otherwise, it waits for a certain period before going idle again. This is where the big hit happens. You're no longer receiving anything or doing any tasks, but LTE is sitting in the long DRX state waiting for something else before switching off. Because 3G's consumption is static and its tail state is always in half power mode, it consumes less power.
The final numbers in the test had 3G consuming 34.67 J while LTE consumed 45.16 J. That means it consumed 30.2567% more energy than 3G for the test. That's a sizable difference, regardless of the fact that it could use less power at some points. Another factor that ties into that is that most of the people with iPhone 5's right now are using the shite out of them. It's a new, fast, fancy phone, and, of course, you're going to want to flex its muscles. This means you're probable also sending and receiving a lot of data, which in turn means you're using a frickton of LTE transmission. This would explain why the battery seems to be crying itself to sleep at night. I wouldn't worry about it too much, though.
Eventually, when your phone usage returns to normal "I've already played Doodle Jump a million times, and I just want to use my phone to text, call people, and occasionally, Google something to prove to someone that you're right and they're wrong" mode, you'll be alright. Well, normal people will that is. If you're a habitual TD poster, LTE may be bending your phone over and having its way with your battery daily.
This is long, I know, but it semi-explains the issue you may see and what could be causing it. It's much better to know why it's happening than to just be pissed off and clueless. Glossary: Capacitive Load
- The amount in farads that a transistor must have while “ON” when being used as a switch. This means no more power can travel across the transistor while in this state. Serialized Instructions
- Only utilizes one core while processing program instructions. Parallelized Instructions
- Instructions may be streamed across multiple cores for faster processing. Milliamp hour (maH)
- The total amount of milliamps that may be drawn from the battery within an hour to fully deplete is. For example, a 1000 mAh rated battery may last 1 hour drawing 1000 mA (1 Amp), 5 hours drawing 200 mA (.2 Amps), or 10 hours drawing 100 mA (.1 Amps). 1000 mAh = 1 Ah. Joules (J)
- The energy expended (or work done) required to produce one watt of power for one second, or one watt-second (W·s).
This post was edited on 9/30 at 10:01 pm