You are currently browsing the daily archive for 17 January 2011.

Charles tries to fix his laptop battery, and partly succeeds.

The reason, of course, is cost. Batteries like that are ‘way cheaper than they were the last time I was on the market — one to fit my old IBM is over $300, still — but $60 to $100 for something to make a computer that old work again is damned high.

Blame modern manufacturing.

Batteries are always going to be fairly expensive. Pick one up. Solid little sucker, isn’t it? And a goodly portion of that mass is stuff that’s hard (and therefore expensive) to get, process, fabricate, make somewhat safe to have around, and/or some other requirement. (Cue “electric car” discussion.)

Laptop batteries are something of a special case. Every laptop model, even from the same manufacturer, seems to use a different battery — Toshiba are less bad about it than most makers, but you still can’t pop the battery out of one model and have it work in another, at least in most cases. Sometimes the differences are “badgineering”, changing the appearance or feature list to provide the Latest Thing without fundamental changes, and in those cases the batteries might swap. It still isn’t guaranteed.

One of the reasons manufactured goods get cheap is economies of scale. If you’re making umpteen gillion copies of exactly the same thing, the main cost is in building the machines to do it; after that, one more turn of the wheel costs nearly nothing. If the machines cost a million dollars to build and 10¢ a cycle to run, the marginal cost per item is a dime. You can make one item at a cost of $1,000,000.10, or you can make ten million of them, in which case they cost 20¢ each. When computers came in some optimists expected “flexible manufacturing”, meaning that the machines could be reconfigured on-the-fly to make different items on each cycle, with little purple flags on every prime-numbered one, at minimal marginal cost. That turned out to be true — but “minimal” is not “zero”; you can still make them a little cheaper if you crank the factory to turn out scads of identical ones.

From the other direction we have pushes for differentiation and optimization, which are different vectors with a small dot product. The heart and soul of engineering is to have exactly the minimum part, at minimum cost to get and install, that will work, and engineers everywhere work hard to get as close to that ideal as possible. At the same time, the Suits want something that’s different — ideally better, but “different” is close enough; the advertising people can make up the difference — so that customers will seek out their product instead of somebody else’s. That’s why the batteries are different. Management wanted something different to attract customers; the engineers worked to make the new product any or all of lighter, faster, smaller, less costly, more powerful. The old battery didn’t fit in the new design, so they designed a new one. If the old-style battery had worked the battery plant could have made another million of them, thereby pushing the cost of each one closer to the marginal cost; instead they started over, so the cost of the battery stays high. You can bet it’s the cheapest possible thing that would work in that model, though — neither the Suits nor the engineers would have it any other way.

The same is true of everything — or, at least, for all manufactured products. You can’t replace the switch in your coffeemaker, because the engineers who designed it designed a switch for that specific model; it’s the best and least-cost alternative for that design, but it’s also unique — and when a new design comes along, the factory will throw away the tooling for the old one and start building the new style with still a different switch. It’s “flexible manufacturing” applied to whole production runs instead of individual items. They made just enough of the older style switch to equip all the coffeemakers it was suited for, then quit making those and started making the new type. You can’t buy a replacement for the older one — they never made any spares! And anyway there are thousands, maybe millions, of different styles of coffeemakers, each with its own unique switch. It would take a gigantic warehouse to store one of each, let alone enough to replace all that might fail, and an army of clerks to keep track of which was which and ship the right one when needed. That’s expen$ive, and we haven’t even got to the heater coil, the filter basket, the LED readout, or any others of the unique parts in each model. If something breaks, it’s ‘way cheaper for consumer and manufacturer alike to just toss the whole thing and get another one.

Greenies criticize that for waste, but it isn’t, really. The engineers have been busily beavering away at optimization,  doing the job with less materials, energy, and labor while making the device last as long as possible or reasonable. Companies spend tens of thousands of dollars per “seat” for software that will enable the engineers to save a pound of plastic in every ten thousand widgets, and engineers will squeeze the program to the limits to do exactly that, resulting in every bit being unique. Stocking, warehousing, identifying, and shipping billions of unique parts would be ‘way more expensive (in CO2 production, among other things) than simply tossing the broken gadget in the trash and replacing it. Fix or fling? Fling, most assuredly.

However —

You may not have noticed, but things don’t break as often as they used to. I certainly notice — I’m old enough to remember when the value of a used car went down precipitously at around 30,000 miles. Nowadays 100K is about what you expect to see on the lot, and the car will still be sound. That’s because of optimization. If every part of the car, or the coffeemaker, has exactly the right amount of the right material in exactly the right places to perform its function, there’s no reason for it to break unless you hit it with a hammer — and that means it lasts for a long, long time. The manufacturer might prefer that it break so you have to buy a new one, and people accuse them of that motive all the time, but think: it costs the same to make a defective part as it does to make a good one, and then you have to spend time (=money) sorting the bad ones out. It’s cheaper to let the engineers work as hard as necessary to make all good parts, and that’s what they do.

Things still wear, of course, but even there, if the part is really optimum for the application it won’t wear very fast, because wear is very much random. If it wears out quickly, consumers are unhappy and the profit margin suffers. Better to let the engineers go ahead and make it last as long as possible; it’s cheaper in the long run. Tornadoes, earthquakes, and burglars with crowbars can happen to good parts with about the same probability as bad ones, and even the best engineer can’t compensate for random chance.

Engineers in their cups have begun to mutter a phrase that isn’t for public consumption: “never break”. As they learn more and more about how to make exactly the right part for the application, the failure rate declines. The result of that has to eventually be that parts never fail at all. You and I won’t see it, nor are our kids likely to, but someday in the future the choice of “fix or fling” won’t exist — because it never needs to be fixed!

“Delightfully anthropocentric,” Prof. Snik’wah says of today’s xkcd. (When last heard from, the good Professor was apologizing for dead blackbirds in Arkansas. The problem, of course, is funding; if zi’i could afford to hire better transport contractors, such things wouldn’t happen. There really are universal principles of social interaction.)

String theory has a curious place in modern physics. On the one hand you have the string theorists themselves, who are ready, willing, and able to explain everything from the Big Bang to Jerard Loughner by reference to their work; on the other you find the majority of physicists and their allies, who point out that not only is there no way to set up an experiment to test it, the theory itself says there’s no such experiment. The cartoonist at xkcd is clearly allied with the latter group, which ridicules (or worse) string theory on the ground that if there is no experiment to prove it there are no physical effects of it, so it doesn’t matter if it’s true or not.

This is a somewhat shortsighted attitude. String theory produces some mind-blowingly abstruse mathematics, and mathematics need not have use — mathematicians themselves will tell you that the amazing thing about math is that it sometimes appears to reflect things that are going on in the observable Universe and is therefore useful. There is no apparent reason why this would be so, and really good mathematicians do it because it occupies their minds with enjoyable puzzles. It’s as if you could build space probes based on the principles of solitaire (which is partly true, come to think). It’s not at all uncommon for a branch of mathematics to be fully developed, complete with professional journals and meetings at conferences, years or decades before anybody discovers a useful application. String theorists find that immensely hopeful. Someday we will be vindicated, and all will bow before us!

Meanwhile mainstream physics occupies itself with a different puzzle, involving black holes. It is a principle of physics that entropy is conserved and is always increasing — never mind what entropy is; it’s a quality of the Universe and the things within it. It is also, perhaps somewhat less forcefully, a principle of physics that nothing can pass the event horizon of a black hole. It follows that all the entropy of everything which has fallen into the interior of the black hole must be visible on the surface, and since the things that fall into the black hole are three-dimensional and the surface of the black hole is only two-dimensional, a fundamental quality of the Universe needs only two dimensions to exist. This leads to the unsettling possibility that there really are only two dimensions, and the third is an illusion. Of course, since there are no black holes nearby for physicists to play with (which is, on the whole, a good thing) no experiment can be set up to test the notion, but unlike with string theory it is possible in principle to imagine one.

One of the major features of string theory is the existence of multiple dimensions. The number varies according to the whim and/or mathematical formulations of the specific theorist, but is always greater than the familiar three. We don’t experience the others — volume can be calculated in proportion to the third power of the “size” of the object, and it comes out right — because the additional dimensions are tightly curled, with all their energy (and entropy) confined to what we see as less than the Planck dimension, which is the smallest possible measurement that can exist in the Universe.

Professor Snik’wah chortles. Zi’i is not a physicist, of course; “he” is an allopodologist with a minor in exobiology, but he does have an intelligent layman’s knowledge of other scientific disciplines, and is more familiar with physics than most scientists in his discipline simply from riding around in starships and having contact with their crews and engineers. According to his understanding, there are an infinite number of dimensions — apparently this is contested — any two of which, if they are spacelike, form a Universe and create the necessary third, allowing particles (and us) to exist. Each such Universe has, theoretically, zero extent in any of the other dimensions. Imagining that the other dimensions must exist within a Universe such as the one we experience is hubristic self-aggrandizement, and leads to errors in thinking and mathematical formulations.

I dunno. They lost me at differential equations, so seeing humility as a mathematical quantity is ‘way over my pay grade.

Tip Jar

Donations (via PayPal)

Hit it, folks.
:fx:Calvin eyes:Puuleeeez?
You don't know many people who need it more.

When I Posted

January 2011
« Dec   Feb »