Gah. There's some incoherent excitement in the blogosphere about a new kind of light source, developed by a company called Ceravision from Milton Keynes. The idea is that you zap a capsule of appropriate gassy mix with a strong microwave signal: the gas ionises and glows.
That's nothing new: what Ceravision has done that is clever (and quite possibly useful outside the world of lighting) is a power coupler that means most of the microwave signal is efficiently piped into the lamp. That's normally very difficult - to get radio frequency power efficiently transferred from source to target, both have to have the same impedance. A bulb such as used by Ceravision has wildly varying impedance from cold to hot, which should mean that most of the power will get lost as heat or you'll blow up your microwave source. Not with the Ceravision coupler, which uses a block of aluminium oxide to form a robust resonator that is very difficult to detune and thus won't change in impedance. Plonk the bulb in the resonator, and all is well.
All this is good and worthy. What's not so good and worthy is the idea that the device is a breakthrough in efficiency, "around 50 percent efficient", say the bloggists.
Curiously, Ceravision itself doesn't make great play of this astonishing figure on its website, prefering to concentrate on the many other advantages. And I'm happy to believe that the ionised gas does indeed turn 50 percent of the microwave signal to light - it sounds a bit high, but achievable. But that's not the same as the efficiency of the whole thing.
So where does this unfeasible figure come from? Step forward the Economist, which has published a rather fulsome piece on the technology - and which Ceravision has duly reprinted on its website.
Let me quote a couple of lines:
"Indeed, while traditional lightbulbs emit just 5% of their power as light, and fluorescent tubes about 15%, the Ceravision lamp has an efficiency greater than 50%.
Because the lamp has no filament, the scientists who developed it think it will last for thousands of hours of use—in other words, decades"
Unfortunately for a magazine of that name, the Economist can't do its sums. While the light source itself may convert 50 percent of the microwave signal to light, generating the microwaves themselves isn't nearly as efficient. The configuration described in the article is fed by a solid-state microwave oscillator and amplifier - and these are normally between 25 and 30 percent efficient. That brings the total efficiency to between 12.5 and 15 percent. There are more efficient sources of microwaves - your oven has one - but they're not at all suitable for powering lightbulbs.
We'll pass over the "thousands of hours of use -- in other words, decades" claim, except to note that that's an order of ten out (at least) and it's hard finding light sources of any flavour that don't last for thousands of hours (albeit not decades).
So - come on, Economist. You're my favourite magazine (well, after the Eye). Don't turn into yet another nest of bloody classics graduates who think science and technology is just too spoddish to bother with, or that the only hook to hang a story on is the bleedin' environment.