January 07, 2005
Department of "Huh"?
The Economist writes that "consumers face stark choices" because of a "mobile-energy crisis" in battery technology:
Economist.com | Mobile devices: Matsushita's engineers have spent eight years working on their new battery, yet it lasts only 50% longer than an ordinary disposable battery. The technology behind the rechargeable lithium-ion and lithium-polymer batteries that power mobile phones and laptops is not evolving much more speedily. According to unpublished research by the Boston Consulting Group, the amount of energy that a battery can store (its energy density) is growing by 8% a year. Mobile-device power consumption, meanwhile, is growing at more than three times this rate, as backlit colour screens, high-speed wireless networks and more powerful microprocessors draw ever-larger amounts of power...
Let's see... 8% per year... that means battery energy density doubles every 9 years... increases 32-fold in 45 years... 1024-fold in 90 years.
Technological progress in battery technology is slow only by comparison with progress in the information-technology trinity: microprocessors, memory, and hard disks. The fact that the Economist can call an 8% per year growth rate a "crisis" is an index not of slow technological progress but of our extraordinary expectations for improvement.
Posted by DeLong at January 7, 2005 09:34 PM
TrackBack URL for this entry:
Listed below are links to weblogs that reference Department of "Huh"?:
» Physics and expectations from Descriptive Epistemology
Brad DeLong makes a good point, while commenting on an Economist article: Let's see... 8% per year... that means battery energy density doubles every 9 years... increases 32-fold in 45 years... 1024-fold in 90 years. Technological progress in battery t... [Read More]
Tracked on January 8, 2005 04:28 PM
It is also example how one kind of improvement cancels another kind. For example, improved microchip technology may decrease power consumption or increase other performace parameters. The same with software.
I guess one can offer a bare-bone notebook computer that would allow someone to read and write documents for days without recharging. And ultra-light to boot. Naay.
Posted by: piotr at January 7, 2005 09:49 PM
http://noeticcenter.tripod.com/ : Wave function collapse is universal, but complex mental phenomena are unique to humans.
Sorry, just could not resist.
Posted by: a at January 8, 2005 01:20 AM
"Let's see... 8% per year... that means battery energy density doubles every 9 years... increases 32-fold in 45 years... 1024-fold in 90 years."
This is not a valid critique. If power consumption increases by 24% per year density/consumption goes from 1 this year to 1.08/1.24 = 0.87 next year to 0.25 in ten years. Assuming density factors linearly into usage time (any experts on this?) the Economist's crisis talk makes sense. Where the Economist errs is the belief that engineers can only improve energy *supply* to close that gap. As soon as returns from developing energy saving technologies are higher than from developing better batteries R&D will switch to the former. I'm quite certain this mirrors the development of, say, digital watches and pocket calculators.
Posted by: ogmb at January 8, 2005 02:54 AM
Simple exponential extrapolation rapidly runs into limits to energy density of chemically powered batteries.
More likely the energy needs of the devices will drop as they grow more efficient, and the batteries will be supplemented by photovoltaic generation. Heck, there are plenty of light powered calculaters. Another alternative is wireless transmission of small amounts of energy: that would not be too much more difficult than wireless networking as a utility.
Posted by: Mike Huben at January 8, 2005 05:51 AM
Battery technology is a problem. I work in renewable energy technologies. Batteries (or other forms of storage) are a necessary part of exploiting an intermittent power source. Battery technology has always lagged behind for a couple of reasons. First, it isn't sexy and more importantly it is a very difficult physical problem. The more attention to the problem the better from my perspective. If advances are made because of the demands of relatively trivial applications, so be it.
Posted by: SW at January 8, 2005 08:02 AM
The battery problems will be solved very soon.
Posted by: northernLights at January 8, 2005 08:42 AM
After 90 years at 24% per year, the cell phone's power consumption would have gone up more than 250 million times, dwarfing the battery gains. But this just shows that the 24% trend cannot be extrapolated indefinitely. Anyone really think the cell phone of the future is going to use more power than today's Hummer--or today's hairdryer for that matter?
This not only not a crisis, it's a case of mass market incentives driving an important technology (compact energy storage) that hasn't advanced as fast as we'd like.
Posted by: Paul Callahan at January 8, 2005 08:58 AM
Or to put it another way, why not discuss the crisis of refrigerant technology? The cell phone of the future is going to be generating an awful lot of waste heat if the projection holds.
Posted by: Paul Callahan at January 8, 2005 09:34 AM
Economist meet Laws of Physics.
Laws of Physics meet Economist.
You guys work this out yourselves.
Posted by: Unstable Isotope at January 8, 2005 09:58 AM
The history of laptop engineering is a kind of punctuated evolution, where things level off every once in a while until a next gen technology is introduced that enables a jump to new levels of functionality. 4 years ago the combination of compute/graphics power, screen res/brightness and battery life available today seemed impossible, but it's shipping today nonetheless.
eg. 40%+ of the power of a laptop is consumed by the screen due to the inefficient nature of LCD technology (i.e. very bright hungry backlight that is filtered through coloured LCDs. Next gen OLED/SED/FED technology will reduce this by half or more.
Next gen cell phones will use more power, but small methanol fuel cells will address that problem. Today's cell phone has more compute power than a older gen 486SX desktop PC, in some models with higher resolution.
Net-net: They are making a mountain out of molehill. The actual story is how new exotic new technologies are developed and standardized in response to meet new consumer or technical demands.
Posted by: quietpc3400 at January 8, 2005 10:00 AM
Battery technology is hard (I've worried about batteries for use in small robot submarines since 1991). Originally we used Silver-Zinc cells which lasted maybe 30 charge-discharge cycles. Now we use Li-ion cells that give a few hundred cycles, but provide the same energy. This is considered a big win. (They are generally safer than Ag-Zn, as well).
The ideal battery packs a lot of energy into a small volume and mass - which is also the definition of a bomb! The difference is that you can extract the energy in a controlled manner from a battery (always harder than letting it go uncontrolled). Of course, what you *really* want is a rechargeable battery, so now you have to be able to take the energy out in a controlled & reversible manner. And, by the way, these are consumer devices so they must be incredibly robust to mishandling.
I think an 8% per year improvement in energy density (or in specific energy) is great progress! And, the fact that battery limitations forces other technologies (e.g., displays, hard drives, chips) to be more energy efficient is fine by me - I *want* our technology to be more energy efficient, given the true costs of our oil habit.
As to methanol fuel cells, etc., I'm not holding my breath for them to take off (although they seem to be set for certain niche applications). The infra-structure isn't really there yet, and I'm not convinced that the improvements in performance can drive the creation of the infrastructure. (If it ran on butane, and we could simply click in a disposable cigarette lighter instead, then it could take off).
BTW, a big driver for this is the military. I recall a meeting in the late 1990's with the manager of a DARPA energy program who said, when asked about methanol fuel cells, "Don't talk to me about methanol. I can't introduce a new fuel into the logistics stream. Gasoline, diesel, and jet fuel, those are what I have to work with for fuel cells."
Posted by: Jim at January 8, 2005 10:42 AM
_The Economist_, a little surprisingly, seems to hire the same sorts of idiot tech journalists that the US mainstream media hires, people who are easily impressed by bright shiny things and who will happily reprint anyone's press release without bothering to add any analysis. Their quarterly Tech supplement reads like the worst sort of crud generated by Ziff Davis. Anyone who can produce four pages on the crisis of why software is so complicated while not once acknowledging the fact that this complexity is far far more a sociological/political problem than a technological problem really has no useful insights into tech.
Posted by: Maynard Handley at January 8, 2005 01:48 PM
I'm calling Bullshit! on the claim that "mobile-device power consumption" is growing at more than 24% per year.
Maybe a few particular devices over a small number of years, but that's hardly representative.
Designers of portable electronics will forever want more power from less weight and volume than batteries can provide. Where's the news?
Posted by: Ottnott at January 8, 2005 04:01 PM
The amount of energy you can extract from a battery is limited by fundamental chemical constraints. I suspect the ultimate limit is only a few times the best achieved today. Jim's comments are quite apt.
A little-known fact is that IC designers now face technical challenges of a different kind and order than in the past. Formerly, shrinking the transistors and wiring was a win-win-win proposition: more transistors per chip, switching faster, with lower power per transistor. But fundmental physical constraints are coming into play, so that's no longer so, and power consumption problems in particular are making circuit design enormously more complicated.
Of course, for portable devices, regardless of battery technology, there is the fundamental constraint that if a handheld device consumes very much power, it'll be too hot to handle. A 40-watt cellphone would get just as hot as a 40-watt light bulb.
Posted by: jm at January 8, 2005 04:03 PM
jm: I have been reading in the computer press, and observing in one case, how the power consumption of current laptops is leading to severe fan noise problems when running the CPU at the full clock rate (and perhaps other components in full-speed mode). Only with a reduced-clock regime you get reasonable noise levels and battery lifetime.
Interestingly enough, the reduced clock frequency of today's 1.6 GHz Intel CPU based machines appears to be around 600 MHz, which is pretty much the full speed of my 5 years old laptop (that I don't intend replacing until it breaks).
Posted by: cm at January 8, 2005 08:22 PM
I have a feeling that _The Economist_ is somewhat biased.
Posted by: Aaron at January 8, 2005 10:53 PM
cm: The thermal problems of current laptops have also led some in the medical community to advise that male users for whom, ahem, fertility is a concern might not want to work with their laptops on their laps.
One of the desperate measures chip designers are having to resort to is adding more circuitry so as to be able to selectively supply the clock signal only to those parts of the chip that really need it, to supply slower clocks to selected areas, and to monitor the chip temperature and back off the clock speed before thermal runaway can start.
Posted by: jm at January 9, 2005 12:16 AM
jm: As long as advertisement and selling is done on one-dimensional "performance" indicators (clock rates, RAM size, engine horsepowers for cars), we will be stuck with those problems. I'm afraid that's just how it works. But battery lifetime (and the underlying issue, energy consumption) is increasingly becoming a concern. Likewise, though maybe to a lesser extent, gasoline consumption and pollution in cars. So all hope is not lost.
Posted by: cm at January 9, 2005 11:47 AM
>One of the desperate measures chip designers are having to resort to is adding more circuitry so as to be able to selectively supply the clock signal only to those parts of the chip that really need it, to supply slower clocks to selected areas, and to monitor the chip temperature and back off the clock speed before thermal runaway can start.
Asynchronous designs are actually a very cool thing, and not because of overcoming thermal problems.
Posted by: Felix Deutsch at January 9, 2005 01:28 PM