Solving the myth of laptop battery life

Cliff Forster 5 Apr, 2009 at 4:01pm ET Article published in Tech

This label represents 90% PSU energy efficiency.

A PSU bearing this label is certified for 90% energy efficiency.

We must work together on this one. Communities like Icrontic will need to come together and propose appropriate programs and usage situations to properly and realistically test battery life. Futuremark’s 3DMark06 has been suggested in extreme scenarios, but I envision the battery life for multiple usage profiles listed on the package. I imagine something similar to a memory card which lists capacity for both SD and HD video. Perhaps your battery could list the life for a “productivity” user and a “power” user. What programs would those users run? How would you design an appropriate label for this?

While I am no benchmarking guru, one thing I do understand is the implementation of standards. I work in the cooling tower industry and anyone looking to be a major player must be independently certified by the Cooling Technology Institute. The technology industry is also no stranger to independent certification for other product categories: Consider Ecos Consulting’s 80 PLUS ratings for power supplies; they have revolutionized the way we view a power supply’s quality. What makes batteries any different?

So Icrontic — tech community — let’s stop being victims of our own consumer apathy, and let’s start coming up with solutions to battle a big myth in the industry. With a few weeks of honest discussion we can move ahead and submit our ideas to the industry contacts that are listening to consumers like us. Wouldn’t it be nice to live in a world where you could trust what is written on the package? Lets make that happen together.

Battery icon image courtesy Deleket.

« Previous

Pages: 1 2

Comments

5 Apr 2009 ~ 5:09pm drasnor

Cliff_Forster wrote:

Consumer apathy is a disease. We often accept mediocrity from corporations, perhaps because the issues seem relatively insignificant, or perhaps because we have accepted that our input is often of little value to the people in charge.

Or, take the OEM battery life and multiply it by a 0.7 "marketing lies" factor. That number corresponds to a reasonable power save profile running desktop apps or watching DVDs. You didn't want to play games on battery anyway.

-drasnor
5 Apr 2009 ~ 7:26pm Mochan I doubt that anyone in the industry will listen. If they do whatever we propose and rate their batteries accurately, they will come up with smaller claims. Compared to a competitor who is doing the "shady" marketing approach, their battery life claims for the same product will be like half the competition's. It's marketplace suicide.

Anyone doing this is killing themselves against the competition. The good guy does finish last sometimes.
5 Apr 2009 ~ 8:13pm Cliff_Forster dranson / Mochan,

Icrontic has been successful at gaining the attention of some high ranking representatives with AMD through the magic of twitter.

AMD raised this argument on their blogs. They suggested something needed to be done, and it required discussion. I think as consumers, we should give them the benefit of the doubt and challenge them. Does not necessarily mean that things will change overnight, but if we as a group stick enough ideas onto the pegboard, then agree to a benchmarking protocol, I say we take the charge, submit it to the concerned parties at AMD for follow up, and ask them to run it up the industry ladder.

Who better to represent the consumer community than the community at Icrontic that is full of intelligent technology enthusiasts? We should post ideas in this forum and go after improvements together.

I believe that the company's we serve as consumers should be accountable to us. AMD came forward and said hey, here is a problem we all know exists, but we have just let it slide. So, here we are, they put out the challenge, lets challenge them back, come up with a battle plan, and ask them to execute on behalf of technology consumers.

I know this is going to sound really far fetched, but take this ride with me for a second.

Icrontic has hardware and benchmarking experts. Why shouldn't Icrontic be the industry watch dog on battery life claims? Icrontic is independent, has no interest or stake in one manufacture of laptops, or portable devices vs. another. Why not??

First, it starts with the proposal. So as a laptop user, what would you want as a real testing protocol? What makes sense? I saw Robert and Peter having some interesting discussion on twitter, and they had some great ideas, but it all starts with a bunch of smart people smacking something up on the board.

Consider for a second, someone may actually listen, care, and eventually enact our ideas? Call me a dreamer or a foolish optimist, but I believe its very possible.
5 Apr 2009 ~ 9:40pm j Let me shed some light on the battery life. I have studied battery life of different chemistry for quite some time. The battery in today laptops are what's called Lithium ion. And the most they will ever last is 5 years. This is due to the lithium in side self degrading. Even if you never use the battery it will only last max 5 years, on avg 3 years. So why does it seem like it starts to fail in a year or less? This has to do with how the battery is charged. Li-ion have to be charged up to 4.2 volts, but they nominally run at 3.3v. Every time you charge the battery you damage the lithium electrode. It's unavoidable for this technology. It's just how the chemistry works. All hope is not lost. The new technology is nano lithum ion. It charges at a much lower voltage, and since the structure of the battery is like little nano sponges, the life span is significantly increased.

Also, one thing to consider is discharge rate of the battery. Setting aside internal resistance, the faster you discharge a battery and then fully charge it, the lower amount of charge cycles you can have. That's why battery manufactures always rate the batteries in charge cycles as well as capacity.

There isn't some conspiracy of why batteries are not better then what they are, It's chemistry.

If it were my choice I would use hydrogen fuel cell technology. Neat stuff
5 Apr 2009 ~ 9:44pm Thrax Nobody's claiming conspiracy or complaining about degrading batteries. We're pissed about the fact that the "six hours" in the advertisement is really 3 hours. Maybe two hours.

Over-stating your product's ability by two or three times would be an outrage anywhere else. All the article asks for is a system of evaluating laptop batteries so the retail estimation of its usage is closer to what a user will actually get.
5 Apr 2009 ~ 9:52pm j as Obi One said. what they told was true...from a certain point of view.

I think 6 hours is for they I use my laptop for MS office user. Real test. Measure the current consumption. Then divide the capacity of the battery by the current. That will give you time. If the numbers don't add up, then maybe they over advertised. But if they do come out, then well you guys are just running too much crap at 1 time.

IT'S A TRAP!
5 Apr 2009 ~ 9:54pm Thrax That's not how they test it, and even if they did, it most certainly wouldn't reflect how anyone uses a PC.

In fact they turn off all the laptop's peripherals (wifi, bluetooth, etc) and let it sit there until the battery is dead. That's it. The laptop never gets used.
5 Apr 2009 ~ 9:58pm j ok well, do that and see if it last 6 hours. if it does, than good. if not complain.

Put it in the freezer too. If the temperature drop by 30 degrees F. Than the heat is being lost from the laptop. And you could have warm peas.
5 Apr 2009 ~ 10:01pm Thrax BAD TROLL IS BAD.
5 Apr 2009 ~ 10:04pm j

Thrax wrote:

BAD TROLL IS BAD.

agree?
5 Apr 2009 ~ 10:37pm QCH Three levels of testing...

Economy - Test how the battery life if every non-essential function disabled, brightness lowered. (Kind of like what they do now).

Normal - Test how an average home or office worker would use a laptop... email open, office document open, web browsing. This translates to brightness near full, memory running at 65%, CPU at 50%, wireless ON, Bluetooth ON, hard drive access every minute for 10 seconds.

Max - Testing as if watching a DVD or doing some serious calculation. This translates to brightness at full, DVD drive running, CPU at max, memory at max, wireless ON, Bluetooth ON.

There... just some numbers. What do you think?
5 Apr 2009 ~ 10:48pm Cliff_Forster QCH2002,

That is a great start,

I agree that a three level methodology makes allot of sense. Three profiles will offer enough detail without it becoming overwhelming for the customer to understand.

On the Economy, perhaps it could simply be a standby metric? How long to drain during total standby? Similar to cellphones.

At normal I agree, office tasks, web browsing over wi-fi is likely the most common used application. How do you create a good loop for that? Perhaps a program that simulates hitting a new web link every couple of minutes? Mixed in with some typing?

The max user profile makes allot of sense to me for what more folks will use a laptop for vs. say a hard gaming benchmark like 3D mark, though for gaming capable laptops, I could justify a fourth profile as a gamer profile, or perhaps replace the DVD, multimedia test with a gamer profile for laptops that are specifically designed to handle a reasonable 3D load.

What do you think?
5 Apr 2009 ~ 10:54pm Thrax I suggest PCMark Vantage with its GPU benchmarks turned off to simulate productivity. It already does browsing, page loading, document creation, etc.
5 Apr 2009 ~ 10:57pm Soundy

Thrax wrote:

Nobody's claiming conspiracy or complaining about degrading batteries. We're pissed about the fact that the "six hours" in the advertisement is really 3 hours. Maybe two hours.

Over-stating your product's ability by two or three times would be an outrage anywhere else. All the article asks for is a system of evaluating laptop batteries so the retail estimation of its usage is closer to what a user will actually get.

This is why almost every ad you'll see will include the phrase "UP TO" in their battery-life claims. It's not "six hours battery life", it's "UP TO six hours battery life", sometimes with a footnote disclaimer in small print, sometimes not...

Unfortunately most people just read right over the "UP TO" part and then get pissed off when the unit doesn't meet those expectations. And that's true of a lot of things, not just laptop batteries.

Case in point is the new TV ads I'm seeing for the Chevy Volt, claiming it will "go up to 60km before using any gas". Sure... 60km, assuming no extra load, properly inflated tires, a light foot on the accelerator, and flat terrain. Put a 350lb eco-geek with a lead foot in there, and that drops significantly... but of course, it would be marketing suicide to actually SAY that in the ads!
5 Apr 2009 ~ 11:01pm Thrax Sure, the "up to" claim is always on the battery life claim, but we've completely done away with it in a bunch of other fields. Cliff's 80 PLUS citation is the perfect example... Power supplies used to be rated as "up to x% efficient," but it was complete bullshit because those were the values at sub-zero temperatures with ridiculously underwhelming loads. Now if a PSU says it's 90% efficient and has the 80+ Gold cert, it's 90% efficient--no matter WHAT happens.

Obviously you cannot calculate a precise lifetime for a battery given all the usage trends, but SURELY we can do better than 200 or 300% overestimates? I mean, really, is that the best we can do?
5 Apr 2009 ~ 11:30pm QCH MPG estimates were notorious for being far from accurate in the real world. EPA made some changes and they are much closer but at no time were the MPG estimates 200% over estimated.

The battery life needs to be based on real life use. Yes, battery life will never be as good as the first month. Over the life of a laptop, the fans become dusty and less efficient, the dust bunnies reduce the air flow so the fans run longer and the CPU runs hotter. This means the laptop will draw more power.
5 Apr 2009 ~ 11:37pm Cliff_Forster Probably no reason why you could not add a X charges till half life stat either?? You could really go after some detailed statistics. I think this is good stuff. These are things that people should expect to know when purchasing anything with a rechargeable battery.

Also, about the temperatures, if temperature, humidity, and environmental factor makes a significant difference, you account for it and control it as part of testing. With cooling towers we have to be at a certain air ambient wet bulb temperature to get a valid test result, so if temperature makes that much of a difference in how a battery holds its charge they you have to agree on an average room temp rating and control the environment for that. Its should be totally transparent how testing is done. If an independent organization were to do it, they should work to make the consumer aware of exactly what they do to arrive at the results.
5 Apr 2009 ~ 11:46pm QCH Agreed... a controlled environment that is REALISTIC. None of this 40 degrees or sub zero.
6 Apr 2009 ~ 12:33am Soundy The same lack of controlled, defined test parameters has been an issue in car audio for ages as well, especially when it comes to amplifier power. Simple physics dictates that an amp will put out more power at 14.4V (which is around the usual operating voltage of a car's charging system) than at 12V (which is the battery's voltage without the engine running). So most mid-to-low grade amps list their power output at 14.4V or higher input voltage, although few actually list the test voltage supply.

Meantime, the higher-end amps use regulated supplies, so their output ratings are the same regardless of input voltage, regardless of whether the engine is running. Regulated amps tend to have better overall performance (in part thanks to the fact that they tend to be better construction overall), while unregulated models will have widely varying output depending on the charging system's voltage.

Line-powered home and professional amps, on the other hand, ARE generally subject to defined specs and test procedures, and typical power supply design makes them less susceptible to variations in supply voltage, so it's a lot easier to compare specs and output ratings realistically.

And of course, all this correlates to the complete lack of any industry standards when it comes to battery performance in particular, and just about anything to do with computer specs in general.
6 Apr 2009 ~ 2:09am Leonardo I think there would need to be some type of widely accepted benchmark application, explicitly designed for laptops. The different classes of laptops and their demands would make almost any other type of testing so completely full of variables that testing would be next to meaningless without a plug-and-play (download and run) standard test regime.

I think the most difficult issue with designing a benchmark app would be to find a common denominator. The problem is scaling. If you do have an app that tests an expensive laptop for "typical" use, how can that application derive meaningful results when run on a netbook? The netbook wouldn't even be able run half the activities that might constitute normal activity on a high end laptop.
6 Apr 2009 ~ 2:37am Thrax A Dx9 GPU test would adequately stress both netbook and notebook alike. Proportionally, both would be put under the same load: Maximum. Battery drain is a function of milliamp hour discharge, and it makes no difference if 100% load comes from DX6 or DX10.
6 Apr 2009 ~ 2:46am Leonardo
Let's draw up a casserole dish of functions that all or most laptops, notebooks, and netbooks should be able to do:
- Internet
- email
- 2D display (obviously)
- office suite applications -- word processor, database (at least to some extent), presentations (2D), mathematics (spread sheet with functions capability), calendar/contacts, et cetera
- movie playback, at least from the hard drive (not all netbooks have optical drives, do they?)
- photo viewing and low resolution photo editing
- Adobe reader (requires more resources than MS Office Word or Open Office Writer)
- MP3 playback, at least through headphones or 'buds'
What other activities would comprise a reasonable baseline? For instance, on a netbook, will anyone be trying video or high resolution photo editing, or web page design? I'm not debating the merits of any mobile platform form factor, just trying to find a reasonable package.
6 Apr 2009 ~ 9:00am Cliff_Forster Do we think that netbooks and fuller featured laptops are different enough animals to have seprate standards for testing? Perhaps that confuses the market more, or perhaps it makes sense because the user expectations on each are differnt?
6 Apr 2009 ~ 9:23am Thrax I think the user expectations of what a laptop should do are the same, regardless of its form factor. If this was not true, there would be no reason for products like the Ion.
6 Apr 2009 ~ 9:58am Cliff_Forster Robert, I think thats fair enough for the standard user profile where web browsing over wireless along with some office productivity is being tested. I also think its good for some kind of video profile, but I think the netbooks realistic limitation hits at anything remotely like a 3D experience. I know it can do some basic DX9 stuff, but will anyone acutaly expect that of their netbook?

Do you support the idea of posting a few different profiles on a package, or would you prefer to see a universaly accepted single profile test?
6 Apr 2009 ~ 10:12am Thrax I think a few different profiles are appropriate, but I think it's also important to give users a realistic expectation of their "worst possible scenario." The only way to do that is to give the GPU -- discrete or otherwise -- a load. And like I said, a 100% load on a GPU is a 100% load, the battery will drain proportionally to the GPU's horsepower.

Even if the Netbook isn't made for gaming, people should know that there are tasks that can greatly reduce battery life.
6 Apr 2009 ~ 1:56pm Soundy

Thrax wrote:

A Dx9 GPU test would adequately stress both netbook and notebook alike.

Only those running Windoze. Unless M$ has come up with a version of DX9 for Linux and OSX...
6 Apr 2009 ~ 1:59pm Thrax It could be an OpenGL benchmark. It could be Glide. It could be OpenCL. It could be any 3D engine.

It doesn't matter WHAT engine is used or WHAT graphics library it's calling on. Battery life is PURELY a function of hardware load. If someone could make a 3dfx Glide benchmark that pegged a GPU at 100%, it would be just as valuable, because all that matters is how much the hardware is pushed in respect to the battery's capacity.
6 Apr 2009 ~ 2:18pm Leonardo

Do we think that netbooks and fuller featured laptops are different enough animals to have seprate standards for testing?

No, if there were two tests, it would just be wiggle room for the marketing dingbats to confuse the consumer. And boy, would they ever take advantage of that! Also, as soon as you ascribe a hardware format to correspond to a niche test, then a manufacturer will redesign the platform and add a new hardware feature to the platform. It's got to be a single benchmark suite that crosses mobile platforms.
6 Apr 2009 ~ 2:42pm Soundy ^Good point. From there, it wouldn't be much farther to some standardized test for phone-type mobile platforms (WinMo, iPhone, etc.)
7 Apr 2009 ~ 3:54am Atterno Realistic benchmarking is no doubt, the need of the hour. Whether it be automobile, IT or Engg, you need standards that are based upon real life situations. As the author has rightly quoted, certain standards like the 80 Plus gold are meant to deliver. Over advertising is a fact and Major players easily do away with their claims by adding clauses like "Upto" and "Conditions apply". So whom can you trust? None but your own experience and expertise. Reliable and independent benchmarking organizations must form and the time will come when Mfrs will look upto these certifying bodies with respect. For the highly optimistic, you can expect atleast some Mfrs to really listen to your feedback and do what is in their lot to make claims transparent like never before.
7 Apr 2009 ~ 12:08pm Audiguyy Great post! I aplaud the community for calling this out. I understand and agree that its not an AMD vs. Intel thing, its a consumer vs. OEM thing!

How do we, the little people make an impact? Do we tweet to the Best buy CMO on twitter?
Do we email or tweet Nigel Dessau?

Next steps?
7 Apr 2009 ~ 2:22pm Soundy

Atterno wrote:

Realistic benchmarking is no doubt, the need of the hour. Whether it be automobile, IT or Engg, you need standards that are based upon real life situations. As the author has rightly quoted, certain standards like the 80 Plus gold are meant to deliver. Over advertising is a fact and Major players easily do away with their claims by adding clauses like "Upto" and "Conditions apply". So whom can you trust? None but your own experience and expertise.

Well, and the price... in many instances, the most accurate indicator of relative quality is simply relative cost - or to coin a phrase, you get what you pay for.

Obviously it's not ALWAYS the case - sometimes you're simply paying more for bling, bells-and-whistles, little-blinky-lights, or a brand name. But if you can eliminate those from the equation, the remainder is generally the overall quality of the device, and the more expensive unit is typically the better one.

(Of course, that doesn't apply so well to the battery-life discussion, it's more of a generality when it comes to technology).
7 Apr 2009 ~ 3:44pm BuddyJ Audiguyy, the next step is to figure out how to get a legit test devised. Discussion here is valuable. Nigel reads it.

If you can come up with some tests, give them a try and post the results. I know Thrax has some good ideas and will be trying them on his MSI Wind soon enough.
7 Apr 2009 ~ 3:45pm wpeltola As I stated in AMD's blog, I find this to be an interesting topic. My so-called four hour battery on my D420 laptop gets about one hour from full charged to totally dead. The laptop is a year old and I don't really use it nearly as much as I do my desktops.

It will be great to have an organization committed to standards and practices in this field. Thanks for reading and great article Cliff !
7 Apr 2009 ~ 11:56pm Cliff_Forster Just wanted to add that I am in the process of forwarding this to a number of OEM PR people to gauge their interest and to get feedback. I will follow up.

Keep the ideas coming.
9 Apr 2009 ~ 3:11pm RADA Not sure I agree with this "3 to 5 year life span" on Lithium Ion batteries, though I can only go by my own experiences.

My only computer right now is my 5 year old Dell Inspiron 8600 with an Ati 9600 Turbo video card. I bought it new on Feb 9th, 2004.

I'm still on the original battery - I routinely use it on and off AC power and still gets just over 4 hours of general use time out of the battery.

I just used it to watch Bourne Ultimatum last week. Got through the whole movie on one charge..
9 Apr 2009 ~ 4:36pm AlexDeGruven While I can see the detriment it would have in the normal consumer market, in the business and enthusiast market, I'd love to see a dual benchmark (similar to what cellphone manufacturers provide with their 'Standby' and 'Talk' numbers).

Give us an 'Idle' number to tell us how we need to set our power-profile, but also give us a 'Minimum' or 'Heavy Usage' number that tells us essentially the lowest amount of use we'll see with the machine. I would love to be able to look at a box and see even a * next to the battery life with a note that says "Battery life varies significantly according to usage profile. Under heavy load, this configuration has been tested to 1.25 hours", or something to that effect.

Real information, and gives the people that really care the details they want.
9 Apr 2009 ~ 5:14pm QCH

AlexDeGruven wrote:

While I can see the detriment it would have in the normal consumer market, in the business and enthusiast market, I'd love to see a dual benchmark (similar to what cellphone manufacturers provide with their 'Standby' and 'Talk' numbers).

I think we are all in agreement with multiple benchmarks based on the use. The issue is different tests for netbook verses laptops. We feel there should be one predetermined set of tests done on all portable computers regardless of their perceived market of laptop or netbook.