Dan's Data letters #152
(page 2)Publication date: 11-Oct-2005.
Last modified 03-Dec-2011.
There's been some buzz about the first kilowatt PC power supply from Enermax called "Galaxy". Is it true that a 1000-watt power supply will draw more power than, say, a 450-watt power support in the same system? I was under the impression that the rating was what the power supply COULD do, not what it would draw all the time.
You're right. Different power supplies have different efficiency figures, so the amount of power one PSU will draw from the wall when running a particular computer won't necessarily be quite the same as the amount a different PSU will draw. These differences aren't large, though (they're more important when it comes to keeping the PSU cool than when you're calculating your electricity consumption), and PSUs with a high maximum rating do not, assuming their efficiency's the same, draw any more power for a given system load than PSUs with a lower rating.
A thousand-watt PSU certainly doesn't run at full power all the time. It'd actually be quite difficult to build a system that'd load it up that much.
You seem fairly erudite with cameras and imaging in general, hopefully you can answer this question for me. I recently acquired a "camera phone" - amazingly enough it was the cheapest one, with all the rebates - and while messing around with its various options I noticed it picks up part of the infrared spectrum. Specifically, it picks up IR LEDs in remote controls and the like, showing them as a bright white flash. Other sources of IR - heat lamps and the like - do not do this, and there is no noticeable distortion of the image if an especially hot item is in the frame. Why is the camera picking up this specific wavelength of IR and not others?
The camera's not picking up a specific wavelength - it's picking up near-infrared in general. Infrared LEDs emit light in one pretty tight band of IR, but CMOS and CCD sensors of all kinds are sensitive to near-infrared light in general. Near-infrared light goes from 700nm wavelength (the reddest light you can see without extraordinary measures) to about 1400nm.
LEDs exist that emit IR up to 1300nm wavelength, but thermal infrared is much longer wavelength than that - something like 8000 to 14000nm. Cameras that can see it exist, but they're not cheap. You can take a nice picture of, say, a hot electric stove element, with a near-infrared camera, but you're not photographing heat when you do it; if the element's hot enough to glow red in the visible spectrum, it'll be emitting plenty of near-infrared (and tons of thermal infrared, which your inexpensive camera can't see).
Ordinary photographic film, by the way, is sensitive to ultraviolet light, but not infrared. That's why people use "skylight" UV filters on film cameras; they stop the photo being "fogged" by UV light when shooting in daylight. Many people put UV filters on digital cameras as well; this is either (a) because they want something to protect the lens from scratches, and UV filters are the cheapest kind, or (b) because they're silly.
You can turn various cheap digital still and video cameras into true IR cams (which see near-IR and nothing else) by just putting an IR-pass filter over the lens. Fancier digital cameras and some cheap ones, however, have a "hot mirror" IR-intercepting filter built in to their sensor assembly, to protect them from image fogging due to near-IR. This is usually a good thing, but not if you want to experiment with IR photography. It's possible to remove the filter from many cameras, though it's a fairly hair-raising procedure that may be best left to professionals.
I think you're being a bit hard on the potential of bi-wiring.
The current system where you have a crossover in your speakers seems like a really bad idea - you can only include the simplest filter circuits, and these circuits will have to deal with many amps of current.
You really would be better off separating your two drivers, and splitting your signal between your different drivers at the line-level stage, *before* you have to start dealing with high-power electronics.
You could even add circuits to match the phase between different drivers, which could avoid strange effects where out-of-phase audio from both drivers interfere, resulting in strange frequency responses and directional effects.
You might also be able to design your speakers a lot more simply, as you could compensate for phase and amplitude problems in your speakers just by doing a little more filtering of your line-level signals going to your independent amplifiers.
With good-quality amplification chips being very cheap, such a system need not even be very expensive. I use 2xLM3876 chips in my home amplifier, and I hope that you're not snobby enough to suggest that these will be in any way noticeable way inferior to more expensive systems: with the huge amount of distortion introduced by the drivers, I doubt that many would be able to notice the difference between any competent power amplifiers.
Yes, this is indeed a good idea; it is indeed much more elegant to do the crossing-over before the amplification stage. It's not even very expensive; it's easy to make an "active crossover" (as such things are called), and the fact that you don't need high-current components does indeed also mean you can make the circuit smaller-tolerance and longer-lasting (and it doesn't have to be inside the speaker, either, completely killing the perennial problem of large electrolytic caps slowly drying out inside sealed wooden boxes).
Sources go to pre-amp as normal, then to the active crossover (commonly only one box for all channels), then all you need is a little amp for the tweeters, an optional somewhat-bigger amp for the midranges, and a big amp for the woofers - which also isn't necessarily expensive these days - and you're done.
Problem is, this isn't bi-wiring (or tri-wiring). It's known as bi- (or tri-) amping. Bi-wiring, in hi-fi jargon, is what you do when your speakers have separate terminals for treble and bass, and you run one wire to each terminal - from the output of a single amplifier. Many speakers have these kinds of terminals these days; they're by default connected together with metal strips so you only need to hook up to one terminal pair, but you can remove the strips if you want to buy more speaker wire for no benefit whatsoever.
Here's more on this, with a picture of the kind of terminals I'm talking about.
I have been looking at bicycles recently, as I need to buy a new one, and a thought of the kind about which I often read in your letters section occurred to me relating to bicycle lights: With modern LEDs and rechargeable battery technology, would it not be possible to make a bicycle dynamo that only engaged with the wheels when the brakes were applied, such that enough electricity was thereby generated to recharge the batteries for the bicycle's LED powered lights such that they need no external recharging, and such that it does not provide any impediment to speed (apart from the negligible extra weight) when cycling?
If the system was designed correctly, the dynamos could be used as a first phase of braking all by themselves (if it is set so that pulling the brake leavers lightly engages the dynamo, and pulling them a bit further engages the ordinary breaking), so that, when only light braking is required, only the dynamos need be used. What other useful bicycle-board electronic componentry do you think could thus be powered? Has anybody thought of this already? If so, why is the equipment for it not in my local bicycle shop?
There are two problems here, but they're both solvable.
One: You can't charge batteries fast enough. This is the problem that all regenerative braking systems face, and it's why people who make their own electric cars seldom bother with regenerative braking at all, since it only gives them a few per cent more range.
For a bike that's only using the regenerative power for lights (rather than propulsion), though, you could easily install a pretty huge capacitive buffer (probably using an "ultracap" or two, as mentioned here).
The second problem is that any system that physically connects the dynamo to the wheel every time you hit the brakes would either be nastily inefficient (like the old roller-on-the-tyre dynamos), or prone to failure.
Fortunately, you could probably get around that by just using a switch. Dynamos that aren't connected to anything spin quite freely, as you can demonstrate for yourself with any small brushed electric motor; spin the shaft with your fingers with the motor just sitting there in your hand and it'll keep spinning for a little while by itself. Short the terminals with a paper clip, though, and it'll suddenly be much harder to spin.
So - get yourself one of those cool hub dynamos that the kids like so much these days, attach a microswitch to the front brake lever so it clicks on as soon as you dab the brakes, and use that to connect the dynamo to the cap/battery charging system. Bingo. There'll be some drag from the freewheeling dynamo, but I (speaking as a person who has no intention of ever trying it) doubt it'd be a big deal.
You could run pretty much anything that normally runs from alkalines from a system like this. Not a 50 watt halogen headlamp, but radio/MP3 player, GPS, speedometer/odometer, visibility LED lights, all that sort of stuff. Run a 12V regulated system and you could use simple little DC-to-DC converters, ripped out of those cheap cigarette lighter power adapter things, to provide 9V or 6V or whatever else is needed.
Such a product may already exist, of course, but I can't find it either.
Hi Dan, I enjoy reading your letters very much, but lately an annoying popup [URL enclosed] comes up every time I visit your site, and the inbuilt Windows XP popup blocker cannot manage to suppress it.
What is most annoying is that the window (I use IE as browser) is rather small, and if I want to open new IE windows after I close the popup, they are still small. The consequence of this is that I must resize and close the full screen window for IE to come up the right way the next time. This is a hassle, and I will not visit your site checking for updates so often if this continues.
I looked at the URL Egil mentioned, and found... a Jamster advertisement. One of those, annoying? Surely you can't be serious!
Anyway, that ad looked likely to make me about enough money to pay for a 30-second phone call to my mother, so I just turned it off.
This, of course, is not a long-term solution to the problem of annoying ads on my, and other, sites. As I've mentioned before (end of this column, start of this column...), I strongly encourage people who're peed off by them to block them. I certainly do.
Also, if you're not locked into using IE by your place of employment or something, I strongly recommend you switch to Firefox or Opera (which is now free as in beer). Both are mature, functional, compatible browsers which are considerably less prone to annoying and/or dangerous problems than Internet Explorer.
Why do video projector globes cost somewhere from $200 to $10,000? Are they priced appropriately for what actually goes INTO them, or is this a case of the projector manufacturers making a killing on replacement components (like we haven't seen that done before...)? Is there a cheap alternative to give new life (at equivalent quality) to one's home entertainment device of choice, or are we stuck with paying whatever the manufacturer asks for their particular family of bulb?
What about transplanting globes between projectors? Ok, I'm sure they have slightly different input voltages and tolerances, just to make things tricky, but globes for some projectors might cost around $300, while similarly spec-ed globes for another manufacturer might cost $900. I'm sure this would have to be taken on a case-by-case basis to match expected input voltages, current and similar brightness requirements, (and possibly different form factors) but is there anything else you can think of that should mean such a project WON'T work? Like, for example, do the globes themselves include any electronics requiring the use of proprietary parts?
I'm sure there's a certain amount of price-gouging involved, but there is some justification for the prices.
First up: These aren't your plain nitrogen-filled incandescents, or even halogen lamps. Modern projectors all use short-arc High Intensity Discharge (HID) lamps, which produce nice white light (good for colour purity) at pleasingly high efficiency (good for power consumption and cooling), but which are temperamental little things that need smaller manufacturing tolerances and more expensive materials than common-or-garden lamps.
And you're not just buying the lamp; you're usually (always?) buying a lamp assembly, with the lamp and reflector built into some kind of cage that mounts in the projector chassis, to line the thing up right.
Automotive HID headlamps are (usually) considerably cheaper than projector lamps, but they're lower-powered and not nearly as optically touchy.
Amortised over the expected lifespan of the lamp - two or three thousand hours, these days - a $300 lamp price isn't all that bad for anybody who can afford a projector in the first place - especially if you've also got an ordinary television to watch the news and such on. Lamp prices can be a nasty landmine for someone who gets a "bargain" on an old projector with only 25 hours left on the lamp timer, though.
As far as I know, there's no cheaper alternative. The generic-printer-cartridge industry hasn't quite noticed projector lamps yet.
If you've got a modern projector that won't let you use it when the lamp timer's expired, you can generally yank the lamp and stick it back in again, which will reset the timer. Which is all very well, provided you don't mind a darker picture, or the fact that tired HID lamps occasionally explode.
It's technically possible to exchange HID lamps for each other, but it's very impractical. You'd have to match arc voltages and wattages (HID lamps need a ballast that's matched to them), and you'd also have to install the new lamp in the focal point of the other reflector - and the lamps probably won't even be socketed in the assembly, so you'd be cutting and soldering, then cleaning everything up fastidiously so acid in your fingerprints doesn't cause the lamp's outer envelope to fail.
Personally, I'd suck it up and buy new brand name lamps.
Here's some more on these touchy little buggers.
I was wondering if one day you'd like to review these babies. I saw them on a TV show today, and they appeared to work, as they make the water react the same way it does with detergent - interesting!
From the first page of Google results for "laundry balls":
Are you seeing a pattern?
(Of course, these searches also spawn a load of Google ads for people still selling the damn things, but that doesn't make them any less of a scam.)
And an especially entertaining pile of dirt here.
Had enough? If not, click here to go to page 3 of this letters column!