Useless "power savers": The saga continues

16 June 2009 — dan

From: "tan" <tan@epowersaver.com>
Date: Mon, 15 Jun 2009 14:04:39 +0800
To: dan@dansdata.com
Subject: Removing Links - Website

Hi Dan,

We have read your write up and would like to suggest that you remove the
link from your site to our website.

You can still quote some phrase BUT please do not simply add a link without
any permission.

http://www.dansdata.com/gz088.htm

Your immediate action on this is much appreciated.

best regards,

Steven

Use Power With Le$$ Cost

http://www.epowersaver.com <http://www.epowersaver.com/>

If you do not want people to link to your Web site, Steven, then don't have a Web site.

(I know this post will be a bit repetitive for regular readers, but clearly the people selling these gadgets are not giving up, so I reckon the world could do with one more page warning about them.)

Steven's company doesn't even have a stupid linking policy. I'm very disappointed. They may yet cheer me up by getting a lawyer to send me a letter, though!

The Ground Zero column that's bothering Steven is the one in which I talk about the numerous worthless "power saver" gadgets on the market today.

Some "power savers" - including some of the things currently found in my above-linked eBay search - actually work. You can, for instance, buy powerboard gadgets that monitor one socket on the board for current draw, and only turn all of the others on when you turn on that one monitored device. A setup like that can, for instance, turn on all of your home-theatre gear only when you turn the TV on, so everything else won't be sitting there in standby mode drawing a watt here, a few watts there, of "vampire" power when the TV isn't on.

The most common kind of "power saver", though, is the kind that Steven and his buddies at ePowerSaver sell. (Oh no, that's another link - whatever will they do?)

The ePowerSaver device is alleged to save you money not by turning things off, but by correcting the power factor of electrical equipment in your house.

A single small plug-in device like the ePowerSaver cannot actually effectively correct the power factor of other stuff plugged into the same circuit. Apart from the fact that the ePowerSaver is not nearly big enough to contain the hefty high-power componentry it'd need, power-factor correctors have to be matched to the load. Too little correction - which is what you should expect, if your power factor is bad enough to need correcting in the first place and the corrector you purchase is one of these wall-wart-sized "power savers" - and they won't entirely compensate for poor power factor. Too much correction - which is actually possible, even with a cheap plug-in corrector like this, since the overall power factor of a modern household can actually be very good - and they can make a bad power factor worse.

But, and here's the punchline, it doesn't actually matter whether these devices correct power factor at all, because nobody but certain large commercial electricity users is billed by power factor. Normal domestic electricity meters can't even measure it.

The plug-in power-saver idea is so dumb that even TV news can figure it out...

...although it seems that that the difference between apparent power and real power, which is the core of this issue, was somewhat beyond KUTV's Bill Gephardt.

Sellers of power-saving gadgets count on this. Their target market is all of the people whose eyes glaze over when someone who actually knows something about electricity attempts to explain that extra current flow from mains-waveform phase distortion in the starboard Jeffries-tube boson inductors is not the same as actual extra power consumption.

(Just because one TV news show managed to figure out these things are a scam doesn't mean that other stations aren't perfectly happy to repackage VNRs from "power saver" companies and call 'em news, though. I particularly enjoyed this awesome piece from an Atlanta CBS affiliate, which cheerfully advertises the previously-mentioned Power-Save 1200. The voiceover proudly states that "the US Department of Energy endorses the device", which is a piece of information which appears to be news to the Department of Energy, and indeed the rest of the US government. Nothing seems to have changed in Power-Save 1200 Land since I wrote that piece in 2006; now, as then, the closest thing they seem to have to an actual DoE "endorsement" is a report - PDF, from a Power-Save site, here - that says that improving power factor is a good idea if you're a large commercial customer, blah blah blah.)

I suppose it's perfectly possible that the sellers of these power savers are like their customers and TV-news talking heads, and don't know the difference between apparent and real power either. I mean, just look at the ePowerSaver Product Testing page, where they proudly show a bank of fluorescent lights (with, I presume, lousy low-power-factor ballasts; modern fluoro ballasts should be much better than this) drawing 1.306 amps with no "power saver", and then only 0.642 amps when the "power saver" is connected in parallel. That's the sort of reading you might perhaps get if you plugged a large enough capacitive power factor corrector in parallel with a highly inductive load, but your electricity meter will notice no difference at all.

(Note that even if you do have ancient awful-power-factor fluorescent lights in your house, then plugging a PFC-modifying capacitor gadget like this into a socket elsewhere in your house is unlikely to achieve very much, even if by some miracle it's got enough capacitive reactance to cancel out the inductive reactance of those magnetic fluoro ballasts that you should have replaced years ago. The "power" circuit and the "lighting" circuit in normal residences diverge from the breaker panel separately, so whatever electrical ebbs and flows the capacitor in the PFC gadget produces will have to interact with the ebbs and flows from the fluoros many, many metres of wire away from them. And then there are houses like mine, in which the power and lighting circuits are completely separate, coming from different wires on the pole outside. Plugging a PFC doodad into a power socket here will have precisely zero effect on the lighting circuit's power factor.)

I think it's actually rather implausible that the people selling these things have never gone so far as to see if the little disk in the electricity meter starts spinning slower when their product is plugged in. But who knows - maybe they went straight to market after doing that clamp-meter current test. And I suppose it's possible that people who are too clueless to understand that sending complaints to people who link to their Web site is not a great idea may also be too clueless to figure out that their product does not in fact work.

If you don't believe me (and Bill from KUTV and his friends from the University of Utah), here are a few other references about these devices:

The Energy Star program run by the US Environmental Protection Agency and Department of Energy would like you to know that these things are scams. (The Department of Energy unaccountably fails to mention how they absolutely 100% you-betcha approved that thar Power-Save 1200.)

British Columbia's BC Hydro and Power Authority, like many other power companies, charges large commercial customers extra if they run gear with a lousy power factor. Here's their page about power factor correction, which talks about choosing motors to match loads, installing capacitors across motors, et cetera. Note that the power bill they quote in their power-factor surcharge example is about one thousand four hundred Canadian dollars per month. If you're not running motors that cost that much a month to operate, you're probably not being billed by power factor.

The Texas Attorney General has busted the makers of the XPower Saver that was tested in the KUTV video clip.

Electricitysaver.com.au used to sell plug-in power savers, until they realised they were a scam, apologised to their customers, and handed out refunds. Now they sell one of those things that just turns off "vampire" devices.

Michael Bluejay's Mr Electricity site.

Peter Parsons' home energy cost reduction site has a section about power-factor devices. (He's also got pages about some other "green" scams, like ineffective insulation, the unavoidable fake "fuel catalysts", overpriced electric heaters like the ones I've looked at, a fuel doodad for oil-fired furnaces, and fraudulent home-power-generation schemes.)

Silicon Chip magazine here in Australia has checked out the "Enersonic Power Saver" and the less-creatively-named "Electricity-Saving Box", and concluded that both are a complete waste of money. You'll need to pay a subscription fee to read the whole reviews, but the circuit diagrams are rather entertaining (note that the second one appears to be missing a "µ" in front of the capacitor-value "F"). If you ignore the components that only exist to light LEDs and protect a capacitor from damage, you're pretty much left with... a capacitor.

In a box.

Imagine my surprise when I saw what you get when you search for "just a capacitor" power saver.

NOTE: I cordially invite any power-saver manufacturers who're now itching to send me a nastygram about this post to, instead, send me their product to review. I will test its effect on a normal household electricity meter while running various household motor loads, like an electric fan, a vacuum cleaner, a washing machine and a refrigerator.

Should your device do what you claim, I will immediately and cheerfully retract all of the above.

Posted in Electricity, Scams, Science. 20 Comments »

Or you could just wait 'til there's smoke

29 May 2009 — dan

A reader writes:

Back when Dick Smith still had Dick Smith in the logo and sold more than plasma TVs and pre-paid mobile phones, you could find there a little electrical detector thing made by HPM. I remember seeing it as a kid, but can't now remember what it was supposed to detect. I came across one in a laboratory undergoing decommission early this year, and took it home and plugged it in. It cheerily lit up a red and green "OK"

Satisfied that something was "OK", I left it plugged in in the bathroom to entertain my housemates.

That probably isn't good.

Last week, though, it switched to red and flickering orange, which I suppose means "Not OK". It has stayed like this ever since, but returns to normal in other people's houses. Lacking the manual, I wonder if you knew what this could signify?

Jonathan

PS Manuka Coles supermarket is selling the power meter you reviewed last year, this time branded Arlec (nostalgia!), for $29.90. Power metering for the masses!

I remember those things! They're "receptacle testers" or "outlet testers", thousands of them no doubt still lurk in kitchen drawers, and more refined versions still exist today.

I don't remember the exact wiring inside the old kind, but their basic idea is to light "Not OK" combinations whenever they see something other than the correct socket wiring. They want to see volts on active, no volts on neutral, no volts on earth, neutral and earth not electrically connected, and earth actually doing some earthing. (I think that's all.) Anything else gives you Not OK lights.

(I think these testers only exist in places, like Australia, where standard electrical sockets always have an earth pin, and only work one way around, so even a two-pin plug cannot be plugged in backwards. The USA, in comparison, is full of non-polarised earth-less sockets, so you can't make a tester that can figure out if there's something wrong, beyond incorrect voltage. UPDATE: As commenters point out below, I was wrong here - the States may still have a lot of two-pin either-way plugs, but any slightly modern building should have three-pin polarised sockets.)

If the tester's working properly, the Not-OK combination does indeed indicate a serious electrical problem. It's entirely possible for a building's wiring to originally be OK and then go bad; that rules out the classic "amateur electrician connected the wires wrong in 1978" problem, and if your appliances still work then that narrows it down further, but there are other ways for wiring to go bad. You could have no earth at all any more, for instance; that can have zero effect while everything else is fine, but if a wire comes loose inside your toaster and touches the chassis, then instead of shorting to earth and popping a breaker, it'll just sit there waiting to shock you.

I suppose this could perhaps just be something wrong with the old socket tester, but since it works in other houses I rather doubt that this is the case. But all things are possible, in this best of all possible worlds.

You can look at the socket wiring yourself without greatly endangering your life, by setting a multimeter (a $10 cheapie meter will be fine, as it almost always is) to the appropriate AC-volts range, and sticking the probes into the appropriate pin-holes of the socket.

(Deem usual warnings about how it's not my fault if you do the above after putting the other ends of the test leads in your mouth to have been included here.)

Looking at the socket, here's what the holes should be:

active /   \ neutral

         | earth

Active-to-neutral should give you 240 volts AC (Australia is now nominally a 230VAC country, but I think pretty much everywhere still actually has about 240V). Neutral-to-earth should give you zero (a small voltage here does not indicate a serious fault; capacitive coupling between active and neutral wires can give neutral a few volts with a safe near-zero current capacity). Active-to-earth should give you 240V (or near offer) again.

It's possible, but unlikely, that doing the last test may trip a "safety switch" (ELCB or RCD), since the multimeter will pass a weeny bit of current. This doesn't by itself mean there's anything wrong; it's normal for perfectly safe appliances in the average house to also leak a very small amount of current to earth, which'll use up some of the trip capacity of the safety switch and leave it susceptible to tripping when something, like a multimeter, adds only a little more leakage.

The impedance of a cheap multimeter in AC-volts mode should be a couple of million ohms, so it should only pass a small fraction of one milliamp from 240V. That isn't likely to bother even a twitchy 5mA safety switch, let alone one of the more common 20-30mA ones. But if your $10 multimeter does manage to trip the RCD, just go to the breaker box and un-trip it again.

(If your safety switch trips when it sees a 20-milliamp difference between the current in active and the current in neutral, and stuff in your house is already leaking, say, 17 milliamps in total, then you can get irritating "nuisance tripping" at random moments. This may be curable by a process of elimination, finding the one appliance in the house that contributes most to the problem; if that doesn't work, the annoying safety switch may be trying to tell you about a real wiring problem.)

A safety switch may also instantly trip when you plug an outlet tester in, because the old-style three-lights tester comes from a time before safety switches, and inherently passes a small current between pins that should not normally be connected. Modern outlet testers are more sophisticated (and more expensive); they typically have an "RCD Test" button, and only pass current from active to earth when you press that button.

If I were you, the first thing I'd do would be to unplug everything unpluggable, and see if the outlet tester returns to happiness. If it does, plug things back in one at a time until the problem recurs, then unplug the offending appliance, cut its cord off and dispose of it. (Or get it fixed, if it's your hundred-inch plasma TV.)

If unplugging stuff doesn't help, and especially if the above quick multimeter test reveals a problem, it's Qualified Electrician Time. If you're lucky, the fault is in the breaker box or the Lovecraftian wiring that lurks behind it. If you're less lucky, the problem's in the walls somewhere. Either way, it's something you need to attend to, lest you get zapped, or awaken to a thrilling housefire.

UPDATE: Modern equivalents of the old socket-tester do exist. Here's one with an Aussie plug that only costs $AU21.95.

Posted in Electricity. 13 Comments »

Basic electronics to make your organ glow

2 May 2009 — dan

A reader writes:

Thanks for your Embarrassingly Easy Case Mod page.

Sorry to be such a techno-dummy, but: You said that because each color-changing LED was 3V, you could connect 4 of them in series to a 12V source. So the LEDs divide the voltage between them? If that's true, how can you connect multiple AC devices to an exension cord and have each of them receive 117v?

Anyway, I'm converting an electronic theater organ to MIDI, and would like to add 20 color-changing LEDs to the console. (Thought you'd appreciate the details, eye-candy-wise.) How do you suggest I do that? If I wire them in series, what sort of DC power would I need?

I know you're a busy guy, so thanks very much for giving me a clue about this. I promise I'll do my best not to blow myself up.

Andy (Vancouver, BC)

Yes, you can run a string of four RGB LEDs from a 12V power supply. They're odd little critters, though, and it's important to understand why this works as well as just the fact that it does. You can make electronic things that work by just blundering around with no understanding of what's really going on, but it really does pay to spend some time learning the basics of at least DC electronics before you start on any electronic project. Hence, this lecture.

In the four-RGB-LEDs-from-12V situation, the LEDs can be regarded like ordinary passive DC-circuit components, like resistors or batteries. But LEDs can't usually be treated that way. Two-leg 5mm RGB LEDs may look like the usual kind of LED, but they're actually three LED dies with a tiny controller circuit, all in a normal 5mm LED package.

If you make a string of simple resistors that all have the same value - let's say, five two-ohm resistors - and hook one end up to the positive terminal of your DC power supply and the other to the negative, a current will flow that's determined by the total resistance and the voltage, according to Ohm's Law: Voltage in volts equals current in amps multiplied by resistance in ohms, or V=IR.

(Ohm's Law is usually written with "I" as the symbol for current, rather than A-for-amps, because when Georg Ohm came up with the Law nobody really knew what current was, and it was referred to as "Intensity". Feel free to write it with an A if you like.)

If the power supply is outputting, let's say, 12 volts, a string of five two-ohm resistors in series will work out as follows:

12 = I*2*5
12 = I*10
12/10 = I
I = 1.2

So the current in this circuit would be 1.2 amps. Because the resistors all have the same resistance, each one "drops" the same voltage. If you measure the voltage "across" the central resistive element of one of the five resistors in this circuit, it'll be 2.4 (12/5) volts. Measure across two resistors and you'll see 4.8V, three will be 7.2V, et cetera. (If the resistors in the chain have different values, they'll drop different amounts of voltage, and dissipate different amounts of power, making you use a polynomial equation if you want to figure out which resistor's doing what.)

To visualise this, think of the current as a flow of water in a hose and each resistor as a narrowing, or kink, in the hose. The higher the resistance, the narrower the path for water flow, and the more pressure (voltage) you'd need to achieve a given flow rate (current). (I've got water analogies for capacitors and inductors, too!)

To really get a grip on all this, I highly recommend that you get a little "breadboard" that you can plug components into without soldering (this sort of thing), and a selection of jumper wires (like this, or you can of course make your own), alligator-clip leads, resistors, capacitors, inductors, LEDs, battery holders etc to play with. And destroy - blowing up resistors, caps and LEDs can be very educational. Wear eye protection, especially when playing with electrolytic capacitors:

A proper adjustable bench power supply would also be nice, but would cost way more than all of the rest of this stuff put together. A lantern battery or hacked-up plugpack or PC power supply would be an adequate substitute, for this elementary stuff.

(You'll also need a multimeter, of course. A $10 cheapie like this will be fine.)

The gold standard for basic electronic education is a "science kit" sort of setup, like the classic Gakken My Kit 150 and Electric Block EX-150. But, again, they're a bit expensive.

OK, back to LEDs. Ordinary LEDs do not behave like simple DC components; they don't just have a "resistance" where hooking them up to a given voltage will cause a given amount of current to pass. A blue or white LED might be specified "3.6V, 20mA", but if you connect it directly to a 3.6-volt power supply, it'll get warmer and warmer and pass more and more current - "thermal runaway" - until, if the power supply's internal resistance is low enough, the LED burns up. This will happen for series strings of LEDs as well; if you make a string of ten "1.8V 20mA" red LEDs and connect it to an 18V power supply, it will probably not last long.

(Power supplies that have high internal resistance are a special case; you can connect LEDs directly across such a power supply and they'll work fine. This is why Photon lights and "LED throwies" work; they connect an LED directly across a lithium coin-cell watch battery, but the battery's internal resistance keeps the LED safe.)

The simple solution to this, as I explain in my old piece about building a caselight, is to put a resistor in series with your LED or LEDs. It's easy to figure out what resistor values to use for a single LED or even an array, but again, doing this without understanding what you're doing is not a great idea.

Series and parallel are bedrock concepts, here, with direct application to a number of everyday situations. Take your question about the powerboard that delivers full mains voltage - in your case 117V, a nominal 230V where I live - to everything plugged into it. It does that because the powerboard's outlets, like the wall outlets in your house, are all in parallel. (There are some tricky things about household power wiring in some countries, but they need not detain us now.)

Now, consider the old-fashioned kind of Christmas lights, with a long string of little bulbs that all go out if one bulb blows, so you have to replace every bulb in turn with a fresh bulb until you find the one that's actually died. That sort of behaviour is a dead giveaway that you're dealing with devices wired in series. In the Christmas-lights case, they're a string of low-voltage bulbs whose total voltage adds up to mains voltage, and they "share" the voltage between them just like a string of resistors. If mains is 240V, twenty 12V bulbs in series will run from it happily.

(Mains power is, of course, alternating current, not direct current. The two are very different, but incandescent light-bulbs don't care.)

OK, now let's finally get to your specific application, adding trippiness to an electronic organ. If the organ is at all modern, it'll run from low-voltage DC inside, and contain a power supply that converts the AC mains to whatever voltages it requires, just like a computer PSU. This doesn't mean it's safe to go fiddling around in there while the organ is turned on, but it does mean that there's probably some supply rail you could easily use to power plenty of LEDs, since they don't draw much power.

You will probably have to fiddle with the organ's guts while it's powered up to find a suitable power rail (unless you've got a schematic or service manual, or the innards of the organ are unusually well-labelled), so all usual safety disclaimers go here, along with my traditional link to the Sci.Electronics.Repair FAQ. But I wouldn't be surprised if you could easily find a 12V-ish rail across which you could connect a string of RGB LEDs, or even multiple strings in parallel.

That last bit is a "series-parallel" array. If you've got 12V and want to run more than four 3V RGB LEDs, you make up multiple strings of four and connect them all in parallel. People often seem to find this concept a bit slippery, but it's another of the things that it's important to grasp if you're to know what you're doing.

Here's how I wired that LED caselight:

LED-array board layout

Those are 18 2-LED strings - and just one current-limiting resistor for the whole thing - all connected in parallel with each other. The little piece of "strip board" I used to make the caselight curls all of the copper traces around to make a rectangle and so is a bit confusing-looking, but electrically it's the same as two long wires, one positive and one negative, connected by 18 two-LED strings like the rungs of a ladder. (Rob Arnold's above-linked LED array wizard is very handy for figuring out LED array configurations, but remember that two-leg RGB LEDs aren't normal LEDs, so you really can just treat them as 3V DC components and not worry about resistors.)

If the organ doesn't turn out to have any tappable power rails, or if you just don't want to fool with them, the LEDs could less elegantly be run from a separate power supply, like a 12V DC plugpack. There's unexpected complexity waiting to ambush you here as well, though; if this page hasn't already turned you off electronics for life, try my essay on Humankind's Endless Quest for a Substitute Plugpack!

Posted in Electricity, Hacks. 23 Comments »

LED-Brite

10 April 2009 — dan

A reader writes:

I live in Taiwan, and I just came across a new LED device which seems very cool.

First, here is a link. It's all in Chinese, unfortunately, and I can't read it to translate for you, but there are at least some photos to give you an idea.

'Aurora' LED sign

[Here's a goofy machine-translation, which gives the thing the name "Aurora", which sounds good enough to me. The price, 1699 Hong Kong dollars, is as I write this about $US220.]

Basically, this works like a Lite-Brite, but with LEDs. There is a black PCB, entirely pierced through with holes. It has no wires, and there are no visible electronic components except for the DC input at one corner. You can plug in LEDs on either side, front or back, in any pattern you like. It's powered by either a wall-wart, a small battery pack, or a USB power connector.

A friend of mine here showed it to me tonight, and it was very impressive. Water-resistant, even - he poured a beer all over both sides of one with many lit LEDs, and there was nary a flicker.

Anyway, if you're interested, I could probably find out more about it.

Doug

I immediately, and completely wrongly, picked the Aurora as a cheaper clone of the Bandai Luminodot (dodgy translation), which was all the rage on the gadget blogs a few months ago. Some hipster has presumably bought himself a Luminodot for $US200 delivered on eBay by now, but I sure ain't.

Doug was quick to point out, though, that this thing is not a backlit-plastic-pegs device like the Lite-Brite or Luminodot, but a bunch of little powered breadboard-ish holes, into which you can plug as many or as few LEDs of whatever colour you like, and have 'em all Just Work with no fooling around with supply voltages or current-limiting resistors or fancy driver pucks.

(I think a cheaper version of the Bandai doodad might be makeable with a laptop CCFL backlight panel and little black shutters that open to let light out when you push a peg though them. Or you could do it the Lite-Brite way, and put a new sheet of black paper over the light for the pegs to puncture every time you want to make a new picture.)

Undaunted, I immediately developed total certainty regarding the Aurora's similarity to another light-array gadget.

Peggy 2.0

That gadget is the open-source "Peggy" invented by Evil Mad Scientist Labs, which is now up to version 2, and available as a kit.

The different Peggy versions are capable of various kinds of animation, and can even be used to display (very low-res) video.

The array Doug saw may, like the Peggy, only actually light one row or column of LEDs at a time, but cycle through them too fast for any flicker to be visible. (It may or may not do the same devious multiplexing as the Peggy, and is almost certainly a lot less "hackable".)

The Aurora is clearly being promoted as being useful for commercial signage, as an alternative to the custom-made, ultra-bright LED-array signs that I've seen sprouting around the place.

Doug was under the impression that the retail price "for a board about a foot square" was only around $US30, plus another $US10 for the power supply. That'd make it worth buying just for the amusement value, but doesn't line up with the $HK1699, $US220-ish price on the product page.

Never mind, though; when an odd toy starts being sold on any Web site ending in .tw, its price will probably be in free-fall soon.

Posted in Electricity, Hacks, Toys. 13 Comments »

Greener Gadgets: This time for sure!

2 March 2009 — dan

Last year, one of the award-winners in the Greener Gadgets design competition wasn't all that it might have been.

To summarise: It didn't exist, and was physically impossible.

This year, the winner is an actual object that actually works. It's the "Tweet-a-Watt" prototype, a system that gives wireless computer access to as many power-usage meters as you like, so you can have your Internet washing machine - or, at least, the power-monitoring feature thereof - without having to wait for an appliance company to make one.

Some of the other shortlisted entries seem quite good, too. You can't argue with something that's actually built and working, for a start. And I spent a while trying to think of something wrong with this little cardboard PC case, but couldn't (here's a bigger one).

This roll-up solar flashlight wouldn't be cheap to make with current tech, but it looks workable. And the BugPlug looks fun and useful, while these evil eyes for standby LEDs are hilarious - I don't care if the light-guide design doesn't seem quite right.

The Zeer evaporation cooler looks good, too. As do the WattBlocks, except for the slightly embarrassing detail that they seem to be turning off "vampire" standby devices by adding a bunch of little doodads that must themselves be in standby mode all day. The clockwork Vampire Plug looks like a better bet, but doesn't do the same job.

Many of the other candidates seem to be the usual "cool design project" things that look really neat at first but become less and less appealing the more you think about them.

Take the RITI printer, for instance, which uses "coffee or tea dregs" for ink.

What a great idea! Make the coffee-ring work for you!

Except that dregs are full of particulates, which I strongly suspect would instantly clog any printer nozzle capable of output resolution better than that of a nine-pin dot-matrix. I suppose they could put a filter in the thing, but that'd have to be a frequently-replaced consumable for your "eco-friendly" printer.

(And apparently the whole thing's run by the user moving the print head back and forth by hand. So it doesn't just need sub-millimetre head-tracking accuracy to know when to squirt out a dot - which would seem to rule out the usual rubber-belt system - but apparently when you're sending a document to it, you have to whip the head back and forth so the printer can hear you. And yet it still features an on/off button.)

This neighborhood intercom doodad might perhaps work, but there'd be non-trivial security issues, unless you made setting the system up no easier than just instant-messaging your neighbours. (Not that it wouldn't be great to just drive through the city with a spare "Eco-Neighbuzz" transmitter broadcasting "Big dog-fight this evening at number 29! No homos, negroes, fat chicks or Jews!")

Some of the entries are more art object than realistic product (this one's on the border line), which is fine by me. I've got no complaints about those, unless they win prizes.

But then, along with various solar devices that don't seem to have enough cells to perform as advertised, there's this solar battery charger, which does not appear to have any solar cells at all. Perhaps the green block in the middle is Kryptonite, or something.

The Urban Fan is entertaining, too. It's a ceiling fan that plugs into a light-socket. Commenters predict rather unpleasant failure modes when you try to hang it from a loosely-anchored socket, and I'd like to add that if it uses an Edison screw fitting, it may unscrew itself after a few on/off cycles.

Oh, and then there's the Enviro'clock Bandage, which commenters observe appears to be a sticker with mental telepathy.

And it wouldn't be an eco-gadget competition without yet another dodgy small wind turbine. This time, there's one that has one little turbine that's meant to work in both wind and water. And the "Wind-Helmet" seems to be trying to set a record for personal-wind-turbine smallness and uselessness.

(Some commenters on both of those entries are, again, unimpressed.)

This solar-powered air-cleaning fan is only mildly stupid. This home 12VDC socket idea could work, but seems to me to be almost completely unnecessary (it's remarkable how many of these proposals have glaring spelling errors). And then there's PpMm pre-perforated paper, which aims to end the endless nightmare that is... tearing up a piece of paper.

If you find any other howlers in the top 50 candidates, please point 'em out!

Posted in Electricity, Nerdery, Science. 33 Comments »

Half theremin, half Stylophone

21 February 2009 — dan

Gakken SX-150

I bought a Gakken SX-150. It's the first electronic musical instrument from their brilliant "Otona no Kagaku" line of "magazine kits", which all come in a funny box with a magazine attached to it that contains instructions for building whatever the thing is.

(Gakken also make the Cross Copter and Mechamo Centipede about which I have previously written.)

The instructions, like Gakken's Web sites, are always in Japanese, but this seldom poses much of a problem. Particularly not in the case of the SX-150, which is quite trivial to put together. As I write this, the Hobbylink Japan page for the SX-150 says "It requires both cement and painting to complete or use. 124 parts" is incorrect. You actually only have to screw the circuit board into the casing, screw down the contacts for the two ends of the ribbon controller and the stylus, and screw down the edges of the little speaker. And put four AAs in it. And cut out and attach the cardboard back panel, if you like.

(I found that Hobbylink Japan had it the cheapest, for Australian shoppers anyway, at about 4380 yen delivered, which is under $US50 as I write this. But it's also out of stock at the moment. The Make: store has it for a higher price, though, as do several other dealers.)

Herewith, a quick SX-150 FAQ.

Does it have to sound like a Stylophone?

No.

The SX-150 doesn't have what you'd call a huge palette of tonal variety - mainly pitch and resonance variations on a, yes, distinctly Stylophone-ish screech - but you can also coax a decent bass tone out of it, as well as various sweeps and bleeps of no use for melodies.

This discussion on monome.org mentions people not seeing the point of the SX-150 until they heard a "mid 90's acid track"; I concur.

Apparently, someone at Gakken said "Let's make a small device with which people will be able to approximately recreate the lead-synth line from Da Hool's "Meet Her At The Love Parade", and somebody else said "Well, it'll need at least a Resonance knob, then", and the SX-150 sort of grew from there.

(Their next product will presumably be the Europa-8, purpose-built to allow you to play the lead synth line from "Axel F".)

The tiny built-in speaker is of course not a bass-monster, but it's easy to plug the SX-150 into other speakers. Its "Output" socket is a fairly hot line-level, so can't drive full-sized headphones very loudly. (It'll probably be OK with little earbud headphones.) It should work fine with any guitar amplifier or effects pedal/unit, though, or with hi-fi gear and headphone amplifiers. I have already connected it to the stereo through an old cheesy digital reverb unit, with entertaining (for me, anyway) results.

Can you actually play a tune on it?

Yes. I was very pleasantly surprised by how musical this tinny little thing actually is.

The standard pitch control on the SX-150 is the prominent black resistance "ribbon" on the front, which you play with the little wired stylus. Left is bass, right is treble, and the total pitch range of the ribbon is a bit more than four octaves.

Some people have achieved tune-playing on an SX-150 by hacking an actual keyboard onto it, with keys connected to the stylus terminal that make contact with the stock ribbon controller at the appropriate points. But you don't need to do that. Even with the standard ribbon, someone with reasonable dexterity can play actual repeatable notes.

The ribbon makes the SX-150 a "fretless" instrument, like a violin or fretless bass. So you'll never actually hit exactly the same note twice. But the pitch-change-per-millimetre is constant - an octave is about 19mm, no matter where on the strip you're playing - and this makes the SX-150 much easier to play than many real fretless instruments. In all regular string instruments, the notes get closer and closer together as they get higher - you can see this effect in the spacing of the frets on fretted instruments.

So in this respect, the SX-150 is like the ondes Martenot or its younger, poorer cousin the Electro-Theremin, which can both make very theremin-y sounds (that's an Electro-Theremin in "Good Vibrations", for instance, not a proper theremin), but are operated by simply moving your hand a set distance for a set pitch change, no matter what pitch you're starting from.

(And then there are trombones, which I have yet to be persuaded do not produce entirely random tones.)

I don't know much about electronics. Can I still do interesting things with an SX-150 (besides just trying to play it)?

Yes. Adding actual new non-trivial features to the SX-150 isn't for beginners, but this thing is genuinely educational, in the very best way. It can teach you things about electronics, and about analogue synthesisers.

Some basic facts: The probe is negative, and the probe-to-strip voltage varies from about 1.6V at the high end of the strip to about 0.8V at the low end. The end-to-end resistance of the strip is about 50 kiloohms.

What this means is that when you connect the probe to the top end of the strip through a multimeter, as I did to get the above numbers, the SX-150 will play a very low note, as a tiny amount of current passes through the multimeter's voltage range.

Many similar tricks are possible. Hold the probe-end in one hand, for instance, lick a finger on the other hand and press it to the top of the strip, and you'll get a low-bass note. Sliding your finger down from there will get you lower and lower bass, far beyond the ability of the tiny speaker to reproduce.

Use a paper-clip as a second stylus, touching the lengthy bit of bare metal on the proper stylus to the paper-clip and then disconnecting it again, with the other end of both stylus and clip touching the ribbon, to create a yodelling effect!

Observe the small but noticeable change in pitch and noise when you hold the stylus close to the tip - so your skin touches the stylus metal - as opposed to holding only the plastic handle!

And the SX-150 is a very limited instrument, of course, of very little use for "real" music. But limitations focus you on what you can do, and this really is a bonsai analogue synthesiser to play with, not just a Stylophone.

Does the "EXT.SOURCE" socket actually do anything?

Yes, imaginary questioner, it does. How convenient that you just asked exactly the right question for me to be able to continue what I just wrote.

The EXT.SOURCE input is a simple example of what all the fuss is about with analogue synthesisers, and the modern software simulations thereof. If you plug a very "hot" signal into that input, it converts amplitude to pitch. Line-level isn't good enough (which is why many people seem to have concluded that it doesn't do anything at all), and most headphone sockets won't go loud enough either; Gakken made this input to interface with their little Theremin. If you've got a loud enough input, though - like a headphone amplifier, or a normal amp turned up only a little bit - there it is; the louder the input, the higher the tone from the SX-150.

This is not very useful, if you don't have the little Theremin. Actually, I think it's probably not terribly useful even if you do. But it helps you make the one great conceptual leap of the analogue synthesiser, especially the modular analogue synth that's a wall of separate "modules" connected together with a spaghetti of patch leads.

That conceptual leap is to realise that audio signals, when conveniently converted to electricity, can readily be transformed in this sort of way. If amplitude becomes pitch a "BOOM tish BOOM tish BOOM tish" drum line becomes "peep boop peep boop peep boop".

That's the whole point of the modular synth. It's all just voltages that different modules create or modify in different ways, and how and where those voltages become actual sounds is entirely up to you.

The SX-150 doesn't take you all the way back to Jean-Jacques and Delia, recording individual oscillator-noises on tape and then endlessly dubbing and splicing. But no mere human has the patience for that. It does, however, give you a real little insight into the dawn of the true synthesiser. So even if you have to pay $US75 for it, I reckon it's a pretty good deal.

UPDATE: Here's someone playing an SX-150.

(The reverb effect later in the clip is, of course, being created by outboard hardware.)

Here's one of many modified versions:

Posted in Electricity, Music, Nerdery, Toys. 5 Comments »

Science Sunday

15 February 2009 — dan

I just measured the electrical resistance of ice.

Ordinary water-ice is a conductor, but not in the usual way. The charge-carriers in most electrically conductive substances are electrons and/or ions, but in water-ice they're protons - mobile hydrogen nuclei. Technically those protons do still count as ions, since they're a hydrogen atom without its electron, but proton conductivity is a distinctly different phenomenon from the usual kinds.

(I am indebted to the inimitable Bill Beaty - previously - for this information.)

So I was sitting here, and I thought, "let's see how conductive ice actually is".

First experiment: Plug multimeter probes into bench power supply.

(Just yesterday, I discovered that you can plug "shrouded" banana-plug multimeter probes into the usual sort of knobs-with-banana-sockets outputs from a power supply; just unscrew the knobs, and the probe-shroud fits neatly around the bare socket!)

Get ice cubes from fridge.

Wind power-supply up to maximum voltage (31V, for this eBay-cheapie supply). Stab probes into ice.

Current-display reading: Zero.

The scale on the bench supply bottoms out at 0.01 amps, though. Perhaps the resistance is just too high for 31 volts to be able to push 10 milliamps through it.

OK, let's try again.

Grab $10 yellow multimeter which for some reason I use much more often than my much more expensive Protek meter.

Set yellow multimeter to its highest resistance scale, which tops out at two megaohms. Stab probes into ice.

Reading: Off the scale, just like when the probes weren't touching the ice.

Suddenly remember that there is a reason why the Protek meter cost more. Its resistance mode tops out at 40 megaohms.

Set it to resistance mode, stab probes into ice.

Eureka! A reading!

With the probes separated by about an inch and only sticking into the ice a millimetre or two, I got a reading of about ten megaohms.

(I took care to avoid letting liquid water bridge the gap between the probes. The ice cubes were made from ordinary tapwater, and clean tapwater is a very lousy conductor - but once you're talking megaohms, all sorts of unlikely things are conductive enough to mess up a test like this.)

No wonder I didn't get a reading on the bench supply. 31 volts across ten million ohms gives a current of only 3.1 microamps. Even with the bench-supply probes really close together I was a few orders of magnitude short of the 10 milliamps that's the least the bench supply can display.

Perhaps it's not surprising that there are all those magic-water quacks. Water may seem to be a straightforward enough substance, but look just a little closer and it becomes as strange as electromagnetism. Hydrogen bonding, proton conductivity, a multitude of different kinds of ice, weird high-temperature, high-pressure behaviour... it goes on and on.

But I think there'd actually be just as much water woo-woo if water didn't do a single unexpected thing, not even expand when it froze. Crackpottery spontaneously generates all over the place, and bothers with scientific evidence only when some portion of that evidence can be used to support it.

(On the subject of ice, by the way, I highly recommend this book. It's full of gorgeous pictures of snowflakes, but it's not just another glossy coffee-table picture-book; it also has a lot of information about how ice forms and why it looks the way it does.)

Posted in Electricity, Science. 14 Comments »

I run MY ThinkPad from a Leclanché cell

8 February 2009 — dan

Back in the day, you couldn't spit without hitting someone saying something completely wrong about memory effect. But today, really loopily idiotic writing about batteries is quite hard to find.

So I am indebted to the reader who just pointed me to one Dave Thompson's article in the Sydney Morning Herald, entitled "The big fat lie about battery life". (It also appeared in the Melbourne Age - the two papers are published by the same company and share a lot of material.)

My correspondent gave his heads-up e-mail the title "Worst. Battery. Technology. Article. EVER.", and I am delighted to say that I concur.

Mister Thompson is apparently under the impression that "average" laptops currently come with nickel-cadmium batteries. This hasn't actually been the case for more than ten years. Laptops with NiCd batteries were still easy to find as late as the mid-Nineties; the famous ThinkPad 701, for instance, apparently straddled the gap, with a NiCd battery for early-production 701s.

Nickel-metal-hydride batteries superseded NiCds, and then lithium-ion or lithium-polymer (generally a distinction without a difference; see this piece, from 2001, for more...) took over in the last few years of last century. The demise of NiMH in the laptop market was quite rapid, even though early lithium-ion batteries had a distressing tendency to drop dead after only a couple of years. But lithium batteries gave a lot more capacity per kilogram, and laptops were expensive enough items that manufacturers could put cutting-edge battery technology in them without greatly - proportionally speaking - increasing the price of the computer.

I don't think it's actually physically possible to buy even a NiMH-powered laptop any more, let alone a NiCd-powered one. Lithium-ion dominates the market, including the low end, and I don't just mean laptops. $20 Chinese Picoo-Z-knockoff helicopters, $12.50 tiger-shaped MP3 players, entry-level mobile phones, $US300 netbooks, you name it. I have a mobile phone that retails for a flat fifty bucks unlocked, and it has a lithium-ion battery. I cannot imagine how Dave Thompson has come by his view of the world.

All undaunted, though, Dave ploughs on with a number of fascinating details about the "NiCads" he alleges are still the standard power source for laptops.

Like, apparently they have a limited lifespan. Well, yeah, everything does, but NiCds are actually likely to work fine for many years if not abused. Few rechargeable lithium batteries are likely to be useful for more than five years.

"If not used properly they simply stop working". I thought he might have been thinking of memory effect or something, but no, he reckons they die if you don't use them, and need regular cycling.

You hear this all the time - it's not right out of left field like the bit about NiCds still being in common use - but it's not actually true. NiCds are actually known for their very long shelf life. If they've been on the shelf for a year then most of the charge will have leaked away, but even if they've been on the shelf for ten years you'll probably just be able to give 'em a charge and put 'em to work.

All you achieve by cycling most NiCd, NiMH or lithium batteries is wearing them out faster. There are certain situations where emptying and refilling a battery can be good - NiCds suffering from voltage depression, say, or LiI batteries whose monitoring hardware has lost track of how much capacity the battery actually has - and I think lithium-ion often hits its shelf-life limit before even someone like Dave can cycle it to death. There are special cases in the radio-control world, too, where absolute battery capacity may be less important than high current delivery and a shallow discharge curve, so your electric car or plane is almost as fast at the end of a four-minute race as it was at the beginning. But as a general rule, cycling your batteries is like "cycling" your car's fuel tank, by driving round the block until it's empty then filling up again.

Dave is, at least, correct that NiCds are a pollution risk if you throw 'em out. Cadmium is quite a lot more toxic than lead, and I think there's still no good way to recycle NiCd batteries, here in Australia at least.

So it's a bit of a shame that he's encouraging everybody to wear their NiCds out faster. Good thing laptops aren't actually powered by NiCds any more, ain't it?

The toxicity issue is one of the big reasons why the much-less-toxic NiMH batteries became popular; nickel pollution is a problem too, but nickel is rather less toxic than lead, and far less toxic than cadmium.

Fortunately, old dead NiCds aren't particularly dangerous just sitting there. So you might as well just toss any dead NiCds you have into a sealable container, put it under the house and forget about it, until someone comes up with a way to recycle them that doesn't involve sending them to China to poison people there.

Dave has noticed that, sometimes, someone who last used their laptop on battery power a long time ago discovers to their dismay that it now has "20 seconds" of battery life. He thinks this is because the battery hasn't been cycled. It's actually because modern laptops have lithium-ion batteries, and lithium-ion batteries have a relatively short lifespan (improving all the time, though - things aren't as dire as they were when I wrote this in 2004). If your laptop battery had 25% of its capacity left when you last disconnected the mains power, a year ago, then yes, it's very likely to be completely dead now, and there's nothing anybody could have done over the intervening months to avoid this. (It's possible that the battery is actually OK but the capacity-monitoring hardware has gone nuts, though; cycling might actually help, there. It's also possible that the laptop has a dumb charger circuit that's slowly barbecued the battery; cycling would in this case be a waste of time.)

Well into the article, Dave remembers that NiMH batteries exist - but then immediately refers to "NiCad's well-known memory effect", resetting the clue-meter to zero just when it looked as if he was making some progress. And then he signs off with "All batteries like to be used, so run them down every few weeks and charge them back up properly just to keep them in top shape", cementing his position in the I Hate The Environment, Die, Environment, Die, Coalition.

Maybe, I thought, Dave just had a small stroke while writing this article and is usually quite sensible. So I had a little look around for other examples of his work.

Apparently cameras, wireless peripherals and "pen-drives" can reasonably be expected to work only once, which is news to people who've been using the same wireless Logitech mouse-and-keyboard set for the last ten years.

Oh, and wireless input devices "eat batteries like a cop in a donut shop".

If current wireless-desktop gear is only as good as the devices I reviewed in 2001, this means police officers have listened to their cardiologists and reduced their consumption of doughnuts to maybe one every two months, tops. Good for them!

He has also written... a thing... about Linux. I wouldn't call it an article. I'm not sure what it is.

(This piece has the brilliant subtitle "Dave Thompson gives his take on Google's new search engine, Chrome", but that's probably the work of a subeditor, not Dave. Mr Thompson tried his best, though, complaining about the usage-tracking feature of Chrome without figuring out that you can turn it off any time you like, and don't have to turn it on in the first place.)

The end of Dave's wonderful battery article says Dave "runs a computer-services company in Christchurch, New Zealand". I think this is it. I wonder if his workmates have some stories?

I don't know how Dave's managed to end up with the ideas he's got. Mere incompetence is common enough in all branches of journalism, but Dave's version of it is odd. Perhaps he just fixes his opinion of every computer technology when he plays with version 1.0, and assumes that 20 years later it'll still be the same. Who knows.

(Oh, and here his battery article is on the Stuff.co.nz site, in case the Herald/Age people do another of their embarrassing-article disappearing acts. Here are other sources, from a Google search for a string from the article.)

Posted in Electricity, Nerdery, Strange Tales. 40 Comments »
« Older posts
Newer posts »