Dan's Data letters #95Publication date: 11 March 2004.
Last modified 03-Dec-2011.
Heat pump fun
I wonder if you could tell me how Peltier devices are rated - how much heat can a given device shift? Will an 80W device transfer 80 joules per second? Or is it more to do with how well you suck heat from one side and feed it to the other?
I have this long-held fantasy (among others) of solar powering a few Peltiers on a big heat sink and using this as a space cooler/heater.
Your fantasy, though, will only become reality if the space you're heating or cooling is the size of an Esky and the solar panel's pretty large. It ain't gonna happen for home heating and cooling. For heating, you'd do better to just feed the current to a plain old resistive fan heater; for both heating and cooling, a reverse cycle air conditioner will achieve much better efficiency than a Peltier rig.
Reverse cycle air conditioning in heat mode can, in fact, have heating efficiency above 100%; you can get more watts of heat pumped into your house than you've expended, in watts of electricity. Peltiers aren't efficient enough to achieve this, because their hot side inescapably heats their cold side. They also make considerable heat of their own, which reduces their cooling efficiency.
The advantages of Peltiers are their "black box" solid state nature and their smallness. The down side is that they just don't work that well compared with older, simpler, bulkier technologies.
It's part of the R2-D2 unit
I have just dismantled an old LS120 drive and came across some thing that I didn't expect. Along with the magnetic read/write heads, there were some optics. At the heart of the optics was a small device:
...with HUL7001 marked on it. After consulting Google, I found out that it is a "hologram unit" (PDF datasheet).
What does an LS120 drive do with a hologram unit?
I think "hologram unit" is basically just a fancy term for "laser pickup". The LS120 and other "floptical" drives store data magnetically, but have optical tracks on their discs as well; the optical pickup's used to locate the magnetic heads more precisely than would otherwise be possible. All-magnetic systems (Zip drives, in particular) beat the flopticals for data density after a few years.
Incidentally, you may also hear about "hologram lenses" in modern optical drives. A hologram lens is a regular lens with a transmission hologram of a different lens embedded inside it. This cunning strategy lets one lens have two effective simultaneous focal lengths, so different laser colours and/or different focus distances (for DVD media) can be accommodated without the need for a rotating "turret" with two lenses on it.
I have an Asus P4P800, and after we got a lightning strike through our cable line which destroyed our router, switch and all associated NICs, it seems it also fried a chip on my motherboard.
I was looking at replacing the chip. I'm hoping it won't be a problem to replace since it appears to be a surface mount regulator of some sort. If you could tell me what it does and what it is it, that'd be a great help. The chip in question has a melted top so I can't read all that it says.
From what I can see, the text on the chip says:
The unit price of this item approaches zero, provided you can find someone who stocks them. If you can't do your own surface mount rework, though (it's not dead easy, even if you've done basic soldering before), it's likely to cost you rather more to get a new one swapped in. And there's no guarantee at all that a new HJ882 will bring your mobo back to life; there may be considerably more damage that you can't see.
If you've got a friend who's handy with a soldering iron then by all means try scaring up a new HJ882, but I suspect there'll be more that needs fixing. Lightning is a harsh mistress.
Maximum photons, minimum money
I was wondering if you know of a cheap alternative to a portable video light? I am unable to afford the optional video light for my video camera, so I was thinking of possibly building one. Can you give me some ideas? I will be filming skateboarding footage at night time so it will have to be fairly bright. My video camera is a Hitachi DZ-MV208E DVD camera.
You want a halogen downlight and a 12 volt gel cell. This is basically what regular cheap video lights are anyway, but a downlight and a standard 7Ah 12V SLA brick won't cost you nearly as much between them, and the light can be bodged onto a camera with little more than some heat-insulating cardboard and duct tape.
For something that works a bit better, you'd want to look into project boxes and, y'know, a switch, and stuff. But you absolutely can't beat the light-per-dollar of 50W halogens, and one will run happy as a clam from a 12V SLA.
If you want more than one, you might want a beefier battery, or at least a couple of 7Ah bricks in parallel; you shouldn't have a big problem toting four or more of them in a backpack, though.
I recently purchased an old IBM terminal keyboard (part number 25H2142) at a garage sale. I had the intention of using the keyboard on my PC at home, although its cable ends in an RJ-11 plug, not PS/2. To my despair, I have not been able to find any stores that sell RJ-11 to PS/2 converters, or think of a good reason why anyone besides me would want such an adapter.
Is the construction of such an adapter possible or is there a way to modify the keyboard to fit my needs? If so, how would I go about doing so?
Terminal keyboards do not necessarily use the same interface and protocol as PC keyboards. Actually, it's pretty rare for them to be compatible. This keyboard might be convertible with a simple plug adapter, but you'd need to figure out the plug's pinout to do it.
AT and PS/2 pinouts are simple, but you can't just apply a multimeter to the pins on the RJ-11 plug to tell which one's which. Opening the thing up should make +5V and ground obvious; after that you can just suck it and see with the data and clock pins, at no risk.
This is, however, only going to work if you actually do have a PC-interface 'board on your hands, and I'm not at all sure that you do. Swapping in a PC-interface circuit board from another compatible IBM 'board would solve the problem, but if you had a donor 'board handy, you would of course probably do better to just use it instead of the terminal 'board.
Cry for help department
I'm so far past wit's end I'd need a space-based radio telescope to even detect it from here.
I've been reading your letter column for some time, so I think I have a pretty good feel for the advice you give people who tell you that their system won't work for some reason or other. I've got the following hardware:
Asus P4P800 Deluxe
P4 2.4C at stock speed
2x Corsair TwinX 512MB PC3200
Radeon 9700 Pro at stock speed
Maxtor DiamondMax Plus 9 6Y160M0 (I think - definitely Maxtor 160GB SATA)
Running WinXP Pro
(All retail boxed, if it matters for anything but the Radeon)
Of late, I've been experiencing what can only be described as "general flakiness". The system will give the occasional blue screen with messages like IRQ_NOT_LESS_OR_EQUAL or similar, but the errors almost never give an actual .sys or .dll file that caused the problem. More often than not, it just clicks out and reboots on me. I've tried the following fixes in no particular order:
Virus and spyware checked with recent definitions
Swap power supply with one from old computer
Swap video card with one from old computer
Disable all onboard components (incl. audio, LAN, FireWire, etc)
Use only 1 stick of RAM at a time (each stick separately)
Disable screen saver, all Windows-based and BIOS-based power management
Reinstall WinXP from scratch, doing a full (not quick) reformat of the partition
I checked for heat problems, and had none. Immediately after a crash, I could check the hardware monitor onboard and CPU temp never peaked over 45°C, usually staying closer to 30. The case is an Antec Sonata, very well cooled by two 120mm fans.
As you might imagine, I'm at a bit of a loss. Before it got really bad, the system event log showed a few Category (102), EventID 1003 errors, never specifying an offending driver file. I've no idea what to do now, as the only thing I can think to try swapping out is the motherboard (and processor?), but I don't have a replacement for either - the old computer referenced above is a Slot A Athlon.
Can you tell me what's going on here?
I can't tell you with any certainty, and I do think the CPU and motherboard are definite suspects; problems that get worse and worse like this aren't usually software problems. But there are some other possibilities.
If the back of the motherboard's making intermittent electrical contact with the case, you can get occasional bizarreness. This happens in the most catastrophic way when someone installs too many standoffs in their case and one makes permanent contact with some solder on the back of the board, which can do nothing (if it's a ground trace being touched) or make the machine un-turn-onable (if it's an important data trace) or result in a smoke test failure (if it's a power trace).
But an extra standoff may be not quite in contact except when case vibration hits the mobo just so, or there may be something else stuck in there - a lost drive screw, a pinched-and-stripped drive power wire, a staple, some reasonably conductive crud inhaled by the cooling fans, whatever. Such mechanical-electrical problems can increase in severity over time.
Similar symptoms can be caused by a broken track on the motherboard, but that's unlikely to be easy to find, or fix; if it's a short-to-case problem then just lifting out the board, blowing it clean, checking for weirdness behind it and reinstalling it should fix it.
Similar intermittent connections can occur in other places - a fractured wire in a data cable, a yanked wire in a power plug, et cetera. Your IDE leads probably aren't at fault, but they might be; try swapping them too.
Slow deterioration can also, however, be caused by things like drying/leaking electrolytic capacitors on the mobo, and I suspect that's going to turn out to be the problem. It might also be a dying CPU, which as you say isn't the easiest thing to test without a spare to hand. If you've got a friend with a Socket 478 machine you could try swapping chips, or you could leave a deposit with a friendly computer store for a loan of an old 1.8GHz Celeron or something (or buy one on eBay).
I'd be more inclined to bet on the motherboard, though, barring extreme bad luck (your old computer's PSU being a bit marginal as well) or truly weird things like a neighbour fiddling with the subject of the next letter.
On the subject of radar jammers [at the end of this column], I am reminded of something I must pass on. Many years ago I was working in the NT for a while. Up there many many drivers are big speed demons (wide open spaces, beer, lots of roads with no speed limits, etc), and many are radar detector customers.
I was in conversation with one guy about the subject, and he told me how he went one better than merely having a radar detector, and had installed in his car a powerful home-built radar jammer he claimed was very effective in disabling police radars.
He used a device he built out of an old microwave oven. At the front of the car, concealed in a large spotlight, he placed the magnetron of the microwave, so that in theory microwave energy would be reflected out the front of the spotlight. Out the back of the spotlight led a bunch of cabling leading into the cabin of the car, into a box with all the other circuitry of the microwave (transformer, voltage converter, on/off switch).
Once his radar detector had picked up a police radar, or he saw one, he flicked the switch and pumped 1000W of microwaves in the general direction of the police radar, disabling it or frying it, and thereby letting him slip by in the general confusion, or at least stopping the police radar from getting any good readings.
Not very sophisticated, using brute force not smarts, but he claimed it was very effective.
I remain sceptical of whether the device worked (not impossible), but I thought I'd pass the story on for your delectation.
HERF guns (as these things are called) certainly do work, and various loonies unconcerned about electrocution and/or getting cooked have made them out of microwave magnetrons, but I don't know whether you could actually destroy a radar gun with one.
If you sat the radar gun right in front of the microwave waveguide, then sure, it'd die, but it's not as if a HERF gun's a MASER; intensity decreases with range, and the one you describe probably had pretty broad propagation (not necessarily all of it outside the car, either...).
The HERF gun demos I've seen usually consist of the (large, unwieldy) gun rebooting a computer at 10 metres or less; you need to be rather close to permanently damage even something as touchy as a PC motherboard.
And, of course, cops with laser speed guns can still nail magnetron road warriors.
You know how you get cans of beer, right, and then you get cans of beer with a widget in them, right? How come, when you pour a regular can of beer into a glass without tilting it (the glass) you get head everywhere (and not in a good way), but when you pour a widget containing can of beer (Guinness for example) you get a perfect head (in a good, beer enjoying, kind of a way)?
Is it because the bubbles are smaller, or don't expand until the beer is in the glass, or something else?
This has been giving me headaches for years. Or maybe it was all the tests I ran...
The "widget" beer-head-creator (technically, at least in Guinness' case, known as a "Smoothifier") works thanks to a cunning nitrogen pressurisation process. When you open the can, some beer is squirted through a small hole in the widget, which creates the head. The beer also isn't very highly carbonated. Draught Guinness is similar; it's pressurised by a mixture of CO2 and nitrogen, too.
When you agitate carbonated beverages they fizz, which is what happens when you pour freshly opened ordinary beer, bubbly wine, lemonade or whatever into a glass. High carbonation is what gives you the all-head-no-liquid problem, which is worse when the liquid's agitated a lot as it's delivered into the glass, and when it holds bubbles better (beer does, soda water doesn't). This is why the same problem happens if someone pulls a draught beer untidily.
Here's some more on the widget.