Dan's Data letters #37Publication date: 16-Mar-2003.
Last modified 03-Dec-2011.
I have a bit of an issue with a laptop at the moment. I'm Head Geek in a Victorian school (yes, it is as fun as it sounds), and currently doing a systemwide upgrade, which involves me filching everyone's laptops for a brief while.
The laptop that rocked up yesterday, though, has a cracked screen. As in, screen is hanging on by one hinge, and if I was feeling particularly game, I could wedge in finger in the panel. So, the big question would be, how high a risk of toastyness in inherent in a laptop display (an Acer 514T with the 13.3" TFT), both in terms of servicing and normal use?
While trying to sort all this out, another issue popped up. To get it fixed, simply return to Acer's Customer Service Trolls and wait 5-7 years for return. So far, so froody. But, Acer's ever-helpful Customer Service Trolls have a tendency to format hard drives, seemingly at random ("Yeah, we fixed the floppy drive, flashed the BIOS and formatted the hard drive for you..."; I've had calls like that). If there is a fair chance of Geek Roasting On An Open Fire, would this be offset by running it from the battery? (The reasoning being that the (Lithium-Ion) battery wouldn't be able to dump all of it's energy into a screwdriver, and by extension, my nervous system, like 240V of wall-mounted goodness. However, I'm always open to the possibility that I'm completely and utterly clueless. Victorian school, remember.)
You won't fry. Fiddling with bare LCD panels is not at all like fumbling around blindly inside a CRT monitor.
[DISCLAIMER: The management denies all responsibility if something happens that causes you to scream for seven straight days without taking a breath and then explode.]
The panel's got a high voltage component in it - the white cold cathode fluorescent lamp (CCFL) that illuminates it. That lamp's ballast will probably output several hundred volts when it's unloaded (when I was checking the voltage of one of the inverter boxes for the decorative CCFLs I reviewed the other day, it struck some nice little multi-millimetre sparks onto the probes of my DMM...), but that's not a big deal, mainly because the ballast's output current capacity is quite small.
Sticking your finger into the inverter in just the wrong place will be a bit more exciting than doing the same thing with a desktop air ioniser, but shouldn't leave you with anything worse than a teeny-weeny arc burn.
The ballast delivers AC at some tens of kilohertz (the higher the frequency, the better the CCFL will work, but the more power will be lost though cable capacitance), but that's not high enough to get into the radio frequency range where "skin effect" causes the current to fail to penetrate much into a conductor (like, for instance, a human). What electricity gets to you will pass through you much like lower frequency AC would. But the low current capacity still makes these things reasonably safe.
It doesn't matter whether you run the thing from its mains adapter or from battery power. Either way, the CCFL inverter shouldn't be able to do you any harm.
[Note, once again, the disclaimer above.]
I've been thinking of ways to clear up the ever growing jungle of wall warts on power boards that is evolving under my desk. I've noticed that many of the devices I'm using (net hubs, USB hubs, video transmitters, etc) have wall warts that are rated for between 5V and 12V. This being the case, I've been thinking (rightly or wrongly) that I could use one of my spare PC power supplies to run many of these devices instead of each having their own transformer. With the heat that each of these things puts out, and the fact that I hate vacuuming around them, I've started to get a little concerned about the fire hazard that they may present - you know, the "I don't think I should plug any more into this outlet, but I *REALLY* want to play with my new toy" kind of thing (logic never wins when I have a new toy ...).
I'm fairly handy with a soldering iron, so I've no prob making cables, but I've never got my brain around making new voltage lines by combining the rails or whatever, and I'm still not 100% sure that my idea isn't full of holes the size of a very large hole.
Yes, you could replace 12 and 5VDC plugpacks, at least, with a PC PSU. Modern PC PSUs can generally be counted on to keep their rails pretty close to the rated voltages even with a light load (old fashioned PC PSUs wouldn't even start up without a decent load; old PCs with no hard drive sometimes had a power resistor warming up a drive bay just to provide enough load for the PSU), so you can quite easily use them to power all sorts of 12 or 5 or 3.3 volt gear.
There are no lurking earthing or isolation issues, and the PSU should be able to deliver more than enough current; just chop off some drive power connectors and wire up some plugpack connectors instead.
You'll need to get the plugs and polarity right, of course, but just using cables from existing plugpacks will do.
Do make sure that all of the plugpacks you're replacing actually are DC plugpacks. Some devices - modems, commonly - have AC wall warts, which probably can't be replaced by DC.
Don't try combining rails to make different voltages. You don't need to do that, of course, if you just want 5V and 12V; there are regular rails right there to supply those voltages. If you wanted, say, seven volts, though, you could get it by wiring a load between +12V and +5V. Many PSUs won't like it if you do that, though; if the only load they see is between +12 and +5, they'll assume a wiring fault and not power up.
As part of a larger load, wiring between voltage lines without an earth to get odd potentials should work OK, but it's just not necessary for most purposes; it's not as if it can give you the 4.5V or 6V or 9V you're likely to need if 5V or 12V aren't suitable.
Note that many devices have some input voltage latitude - in fact, well designed devices always should have. A device with a 6V plugpack, for instance, may work from 5V. It's safe to try a lower voltage; it's not safe to try a higher one, unless you're monitoring the device's consumed current to see if it rises. If the current rises with the voltage, you're probably going to blow up the gadget. If it stays level or drops, then a (slightly) higher voltage than the device is rated for should be fine.
As I mentioned in letters #22, changing linear plugpacks over to switchmode (to switchmode plugpacks, or to one big switchmode box) will also save some power.
My girlfriend recently decided that I needed a "Woodside Biomedical ReliefBand electrically stimulating nausea reducing bracelet"! I added the exclamation point for effect.
So I've been reading up on the "studies" they post on their website, and most of them look bad. But I'm not sure if it was the "researchers" that don't know what they are doing, or the Webmaster who wants to make the bracelet look good on the Web page. There is, of course, no conclusive evidence anywhere, and several of their studies found that while they were operated double-blind the bracelet had placebo effect, but once they switched to not-blind-at-all their bracelets had enormous effects. Sheesh.
Here is the only hard data I could find on these things. From this page:
"The ReliefBand provides electrical stimulation to P6 at 10-35 mA and 31 Hz via a battery powered device."
P6 is your wrist, they just want to call it something special. I read this and thought of you (how touching, eh?). I'm not sure if you know how much power or electricity this is, I certainly don't, but I was hoping that if you have the time you might enlighten me. Is it more then a AAA battery? Could I just hold a battery in my hand and get the same effect or what? Electrical doohickeys confuse me, I'm an accounting major, so if you have time to educate the masses please do!
Daniel [no relation]
The ReliefBand is an "acupressure wrist band" with electrodes. The wrist bands are an old, though not necessarily effective, remedy for nausea of various kinds. Some people swear by them, some people think they're just a placebo, everyone agrees that they do no harm and don't cost much, so what the heck.
"ReliefBands" seem to cost at least $US70, though (or twice that, from some retailers...), so they're not that cheap a bet.
While the ideas behind acu-anything are pretty much totally incoherent, it's not ridiculous to suppose that some form of stimulation - and mild electric current is a perfectly valid form of stimulation - to some part of the body might successfully distract you from your queasy stomach, or even do something more neurologically complex to relieve nausea. Cure your brain tumour? No. Make you feel less woozy? Maybe.
Snapping a rubber band on the wrist, for instance, has been quite successfully used by people prone to panic attacks to help keep them out of a phobic reaction. Again, though, rubber bands aren't a high-dollar item, but ReliefBands are.
There are a few abstracts in Medline to do with the ReliefBand, but it doesn't seem to have been extensively studied, and I haven't read the whole papers to see if they're any good. The research backing up many "alternative" treatments is not impressive.
The ReliefBand's makers seem quite proud of the device's 510(k) approvals from the FDA. That irritates me, because a 510(k) like the ones the ReliefBand has is nothing to be proud of.
Some places with info about 510(k)s say that they require proof of safety and effectiveness; that's not necessarily so. A 510(k) (see here) is just a proof of equivalence - proof that a new device isn't different enough from an existing, legal-to-sell device that there's any need to retest it.
Now, obviously, existing legal-to-sell devices should have been tested for safety and effectiveness, by someone, but they haven't necessarily. Not by the FDA, anyway. Electro-acupuncture gear like the ReliefBand, for instance, would seem not to have been.
The ReliefBand's 510(k), you see (see the original application in PDF format here; it contains some detail about what the thing is meant to do, and how), states that it's substantially the same as a pre-1976 device which can also be legally sold. In the USA since 1976, new medical devices that weren't equivalent to something else already on sale have had to undergo FDA "premarket approval", or PMA, which is a quite serious safety-and-effectiveness test. Before 1976, there was no US Federal testing requirement for medical devices.
Trouble is, gear that was on sale before the 1976 introduction of the PMA process was "grandfathered in", and allowed to stay on sale until such time as the FDA did a PMA and proved it shouldn't be, which generally didn't happen. Heck, I don't think it's happened for pre-1976 pacemakers yet, much less anti-nausea wrist bands.
(See here for a case with some background on this.)
Such would appear to be the case for the ReliefBand. It's been proven to be as safe and effective as another device which has not been proven safe and effective, at least not to the FDA. But the makers are happy to pretend the FDA has an opinion about the ReliefBand, when it in fact has none at all.
The "10-35mA" claim on this page doesn't sound at all likely to me, for more than one reason. I can only surmise that the quoted current is a misprint.
First up, the ReliefBand is, according to its 510(k), powered by three SR/PR41 batteries. These little button cells can deliver a few tens of milliamps, but not for very long. Ask for 10mA constantly from them and you'll get a lot less than an hour of service, but they're supposed to last something like 150 hours in the ReliefBand (or longer if you leave it turned off a lot, of course), I think.
Now, the ReliefBand might be stepping down the voltage to reduce the load on the cells, but there's no way it could achieve its rated current if it did; even with the conductive gel you're meant to put under the ReliefBand's wristwatch-like body, there'd be too much resistance to get any worthwhile current through the skin (as opposed to through the gel...) at a lower voltage. If anything, it's got to be stepping the voltage up, if it wants to deliver even 10mA through its dinky little belly contacts. The contact area is too small to do it at the 4.5 volts, maximum, it'll be outputting, without step-up.
I am not just making up numbers, here. I tried pushing some current through the soft tender skin of the underside of my own wrist, to see what'd happen.
My first attempts, using just a couple of alligator clips pushed onto my flesh as contacts, were a failure. Even with about 16 volts coming from my bench power supply, with the skin wet and the probes close enough to each other that they were almost touching and shorting out, I only managed to get 2.5mA flowing.
I managed to do better than that, by making one two-inch-wide wrist-cuff electrode out of aluminium foil, and attaching the other wire from my power supply to a steel ruler that I could easily slide up and down my arm. Now, I managed 5mA (still from over 16 volts) before I wussed out. My wrist was starting to sweat and increase the conductivity of the contact (even trolls know that), so I dare say I could have managed considerably more, but I really didn't want to because 5mA was quite uncomfortable.
There were no muscle spasms or other weirdness (and since the electricity was just flowing from one part of my wrist to another, there was also no risk of nasty cardio-pulmonary complications; I'm silly, but I ain't stupid), but it still certainly wasn't an experience that you could effectively duplicate with an inert placebo. You're not going to fail to notice several milliamps through any tissue with nerves in it. 10mA would, I think, really be rather painful; I would pay a fair bit of money to avoid 35mA.
Pulsing the current at 31Hz, as the ReliefBand is meant to do, wouldn't make much difference to its painfulness, I think. DC's generally more painful than AC when it causes muscle contractions, but that's not what's happening on the wrist, where there aren't any muscles, just nerves and tendons and bones and blood vessels. Very high frequency electricity is a lot less dangerous, and less painful, because it tends to penetrate only slightly into conductors (including humans; see the first letter on this page). But you need radio frequency for that effect; 31Hz is barely even audio frequency, much less RF.
I think the ReliefBand people probably meant to say one to three milliamps, or something. With a short enough duty cycle, the thing could manage that with its button cells and not hurt the user. And maybe it even does something. No way is it delivering 10 to 35mA, though; the batteries can't do that unless the "31Hz" frequency is 31 50-microsecond pulses per second, and you might feel even those. And if you don't feel them, then all I can imagine the thing could possibly be stimulating is the advertised acupuncture point, which is imaginary.
Acupuncture/acupressure points and meridians are available in such a remarkable, and contradictory, profusion that some acupuncturist somewhere is guaranteed to believe pretty much any spot on your body is special for some reason. P6 is on the wrist? OK, whatever.
You couldn't just hold a AAA battery in your hand and get the same current. From the 1.5 volts of a AAA cell, across an amount of skin equal to the length of a AAA battery, you'd only get a couple of tenths of a milliamp flowing, even with the help of salty conductive gel. But I can only imagine that pressing a nine volt battery's terminals to your wrist would work just as well as the ReliefBand, although it wouldn't be as convenient.
UPDATE: Shortly after I put this page up, this Business 2.0 piece showed up on the CNN site. The person writing it reports that at full power, with fresh batteries, the ReliefBand is noticeably uncomfortable. So maybe it does have decent current capacity, though presumably with a very small duty cycle. But the author also describes the ReliefBand as "sanctioned" by the FDA, which is disingenuous, as I explain above. Oh, and she also seems to think that dizziness and nausea are the same thing.