That's a good point, that voltage rating is an important issue. But it isn't because of expense of high voltage insulation, it's because THAT part of the power handling is safety-critical (electric shock hazard).
When you've converted to nonlethal voltages (under 50VDC, usually) the remaining safety concern is does-it-start-a-fire, and simply limiting the current completes that task (fuses, fusible resistors, circuit breakers, some kinds of active 'foldback' voltage regulator schemes, all are available strategies).
First, power requirements for common peripherals (small hard disk drives) exceed the standard capacity of USB (0.5 A at 5VDC); there are some laptop plugin drives that come with TWO USB cords, because it needs power from both ports to spin up.
Second, high power variants of the USB port (Apple tried this, on the cube, for some high-power speakers) give rise to odd incompatibilities. Breaking the standard is a bad idea. Trust me. I've diagnosed/dealt with it and don't care for a repeat of THAT.
Third, there are devices that need other voltages (like EIA-232 serial ports) and the 'universal' +5V is just plain wrong. Converters are used, of course, but the converter isn't notably simpler than an AC power brick; you save on cables, not on hardware. Cheapo converters for EIA-232 are energy inefficient, but there isn't much energy required, so that's OK.
Fourth, it's cheap to make high voltage parts and expensive (in terms of chip area and yield from a semiconductor processing plant) to make high current ones; if you knew you were gonna convert the DC voltage, your choice of input voltage would be higher, 48VDC (about like telephone company power handling) or the new automotive standard, 42V. Power-over-Ethernet is standardized at 48VDC (negative voltage) for this reason. The price difference makes little impact on the customer, but some pennypincher engineer will always choose for you. Then, the marketing department won't show the brick in the pretty boxtop picture. Firewire does use higher voltage power (12 to 24V), with similar current (so the wire isn't stiffer than USB wires). As a result, Firewire power DOES support a hard drive with appropriate conversions inside the peripheral.
I'd like to see The Snarkout Boys and the Avocado of Death on the silver screen, myself. Or, if that's too challenging, how about Yobgorgle, Mystery Monster of Lake Ontario. Imagine, cinematically, the chicken-suit car purchase and a pink submarine lurching onto shore in search of a really GOOD roast-beef sandwich. Daniel Pinkwater is da bomb.
Favored books don't always work on screen (Dune didn't, IMHO, and before that I was disappointed in the anime version of Lensman). I figure I'm due for one that DOES work in film.
To 'shield from magnetic fields' is generally the same thing as 'generate an opposing magnetic field'; that means the shield materials ARE affected, and are in fact somewhat magnetized, to create the shielding effect.
The best shield materials are superconductors (which only exist at low temperatures). The most common magnetic shield materials are soft iron alloys (Permalloy and Mu-Metal are brand names). Shielding from rapidly-changing magnetism is easier, most electrical conductors will do this (but superconductors do it for constant magnetic fields as well as changing ones). A weak shielding effect is called diamagnetism, and is interesting in its own right. Did you know that water is repelled from a magnetic field? Water is diamagnetic (weakly). Brass is more highly diamagnetic.
Yes, it should be possible. There are positive-temperature-coefficient resistors that 'switch' current off when heated. You could make a crude relay with heater resistors as the input and a PTC resistor as the output conduction channel. Negative-temperature-coefficient resistors are also available, you could do something like CMOS using both.
I've seen some thermal effect in silicon chips, related to heat changing the input current of (for instance) op amps. It makes a big effect, but only at low speeds (two cycles per second, in the case that bit me). Silicon is a very good heat conductor, if you were to make thermal gates on such a substrate it'd be one per chip OR you'd have to get such high switch rates that the heat 'leakage' current was negligible.
But crude logic IS still useful; there are intrinsically-safe electric heaters using a ceramic that stops heating when it is just under the ignition temperature of common home furnishings (using a positive-temperature-coefficient heating element to lower the power consumption when the target temperature is achieved).
It's actually beneficial that a single 'gate' element can perform AND, OR, and INVERT functions all in one stage. The early TTL won over other logic designs in part because the basic gate used multiple emitters on the input transistor to get an AND function, and multiple input transistors to get the OR function. That meant that the delay and complexity character of AND and OR were the same, and that the complex function of AND/OR/INVERT was available as a fast multiplexer, with the same characteristics as a simple NAND. There was a brief attempt to use expandable gates (making the connection point after the input transistor available on an external pin, which was NOT TTL-logic-level compatible), but it didn't catch on.
CMOS, on the other hand, had input impedance and delay differences in the AND and OR and other gates, so the whole 4000 series CMOS logic family only became trouble-free to use AFTER THEY BUFFERED THE WHOLE FAMILY with an extra inverter (and consequently extra time delay). Buffered (4000B series) is the common small scale CMOS you see today, the unbuffered (4000A series) has been sidelined.
From a circuit-design viewpoint, the AND/OR/INVERT is a very good starting element, for a lot of reasons that only show up when some poor engineer is perspiring over his timing budget...
The clear problem here, is that a President can direct a large number of civil servants, but CANNOT ethically use those civil servants as proponents for some political position or the other. The academic community dislikes this particular chief executive because of multiple instances of pressuring academics to adhere to a Bush world view,
When pressed for an explanation, the official word is that the manipulation of published works is 'part of a normal review process". Of course, the NORMAL review process under other administrations was 'peer review' and this administration applies 'political officer spin doctor' review. Our president also thinks Intelligent Design should be taught in schools, which gives you an hint of what our scientists in government agencies are dealing with...
This argument, "one big problem with radioactivity is that people can't see it" is insidious and evil. It's the sort of fall-back statement that CAN NEVER BE COUNTERED, but has no other merit, logically, whatsoever.
It's like the old story of an argument between a white politician and a black one,
Bystander to reporter: "What's happening?"
Reporter to bystander: "He's still black."
It is evil to think energy policy depends on the limitations of human sight and taste as applied to sensing radioactivity, just as it is evil to think a politician is at fault for his race (or any other non-negotiable accident of birth). Both arguments are flawed in the same way, and both do clarity an injustice.
It should be pointed out that the persistent cookie mainly is undesirable because folk who bulk-buy surplus computer hard drives could search for nuggets of sensitive info and be guided (as to the interests of the original user) by those cookies.
Or folk who bulk-burgle hardware bits.
If all the cookies are from Microsoft or Yahoo or music fan sites, it's not likely that candid comments on foreign policy from government insiders are on THAT hard drive. But a drive with cookies from NSA and CIA, might indicate some value of further data mining efforts.
It's a minor indication, certainly, but plugging small leaks is still worthwhile. So, I don't see a policy intended to benefit the public, per se, but rather a policy intended to obscure the historic record left on a filesystem. You have to recall that NSA isn't the only data security threat out there... then it all makes some more sense.
It's said that president Lincoln often composed letters to or about folk he was upset at, then carefully filed them away and never sent them. I'd like to see the hard drive text files for some contemporary Republicans to see how their habits mirror those of the party founder. Heck, a love-letter from Bush senior to Saddam could plausibly be on a discarded disk drive now. Check your disk deadpile!
I've replaced backlights in powerbooks, both with official Apple parts (it has been some years since THAT was available) and with generic lamps from third-party suppliers.
Firstly, remember the lighting uniformity is HARD TO ENSURE. For the thinnest displays (like modern Powerbooks) it's unlikely you can even FIND a source for the lamp (2mm tubes are common, the available units are usually 4mm or the wrong length or both), and if you do find it, getting the foam/mylar/backplate sandwiched after replacement is going to be a chore. Expect a splotchy result.
The most recent display I had any good luck with was one of the Powerbook 500 series (about 1995 vintage). That required a little work with a Dremel tool to fit the available lamp (which was about 5mm too long), and took a bit of care during disassembly. The plastic display bezel got brittle with age by the time the unit was both out of warranty and failing to light up.
The 'supported' solution is to ship the unit to the manufacturer for rebuild. Presumably, they DO have the right parts and a lot of patience. My advice: pay the $400 (or whatever).
There's a lot less toxin in the lamp than in your last filling. More worrisome is the wiring to the tube, which wants about 2000V to start. The wires are funny, rubbery things, possibly a hightech silicone.
One very useful tool was a spare power supply; if your backlight was on a connector, you can test it on known-good power that way, and an inverter replacement is a LOT easier than lamp rebuild.
Apple is a VERTICAL organization. The quality of the components relates to their INTEGRATION, both with hardware and software/drivers.
One post commented on the 'generic RAM and crappy video card'--
Apple's RAM is higher quality than much commodity RAM (I repaired lots of Macs with bad RAM, usually not the Apple-supplied stick); but it doesn't show when you open the box.
Apple's video card supplies power and USB to their monitors; no PC vendor has a monitor powered without an AC cord, nor a USB hub conveniently on the monitor, because the designers of the units never sat down together. Vertical organizations can support this kind of coordination, and the Wintel stack-of-blocks teams are at best gonna stay a few years behind.
You can castigate the 'crappy video card' all you want, but the imaging model in the MacOS fully USES the graphic programmability of that card, for ALL applications, and supports it (virtual video memory) in ways the Wintel boxes don't. This happens because the OS designers and the card programmers all talk to each other. Before going into production.
Booting from USB disks and from network volumes and Firewire disks, and hotswap mice and keyboards, and target disk mode (a crashed computer can be rebooted so as to serve its disks out as Firewire volumes to a known-good computer) are all Macintosh features the Wintel world can only copy as an afterthought (if at all).
You can castigate the 'crappy video card' all you want, but the imaging model in the MacOS fully USES the graphic programmability of that card, for ALL applications, and supports it (virtual video memory) in ways the Wintel boxes don't. This happens because the OS designers and the card programmers all talk to each other.
Apple's upcoming x86 boxes may lack some PC essentials (PS/2 keyboard ports and onboard floppy controllers), but the quality that matters, the good functional integration of features, will still be there. And people won't buy Apple machines to run Wintel software; that would be pretty pointless. The full advantage of buying the Apple machine lies in the integration of the parts, not in their chrome plating.
It's a great plan, but there's a major safety issue. House wiring is self-extinguishing for AC arcs across the wires. It isn't similarly safe for DC arcing.
DC fuses are harder to make safe, as well. At low voltages (12V like in a car) that's not too bad. At higher voltages (28V like in aircraft and trucks) it's a well-studied and solved problem. Even at telco (telephone company) voltages (nominally 48V, range 36-72V is the typical design limits), it can be handled.
But, the practical UL-approved devices in your home have started to have little switching power converters, that take the 120VAC and/or 250VAC and convert it to your low voltage right inside the appliance, and that's nearly as convenient as having multiple power grids and WAY more flexible in terms of total power available and voltage-output flexibility.
Like you, I've thought of backup DC power on a separate bus (and use some fixtures like trailers and RVs use that just run on my backup battery). But, it's too much trouble to do safely. And, WAY too much trouble to do any other way.
The only way data centers can benefit is because they have UPS issues (like telco does) and the battery backups and their DC layer has to be common to the whole roomfull of equipment anyhow. Might as well run power-distribution wiring with the low-voltage DC; the wires only stretch one room in distance, the wire-resistance losses aren't prohibitive.
There are multiple uses for a slowed signal; you can combine it with the un-delayed signal and make filters, like is done with SAW filters (but those use surface acoustic waves, and are not silicon-compatible). You can also make some kinds of shift register VERY simply by sending the signal out into the delay and picking it up when you need it. And a delay of a clock signal often makes a computer more reliable (designing high speed compute devices, this is OFTEN a vital consideration).
The split/multiple delay/combine scheme for (for instance) radio signals is a very powerful tool, and is why a complicated-looking antenna can work so well. And, why a rabbit-ear antenna can take a lot of tweaking to get your idiotbox to receive Red Green.
For major processing of data, it was common practice in the old days to tweak the interconnect wiring to make the correct time delay. Seymour Cray reported (of the Cray-1 supercomputer) that the interconnect in the central core of the computer was hand-wired by (slender women) assemblers who used cut-to-measure lengths of twisted pair, so that all the signals had the appropriate settling time before the clock arrived and latched the data. The computer was a cylindrical hole with draped wiring all over its interior, with spokes out that housed the cooled ECL logic modules.
To keep the Cray quick, the cylindrical core was as small as feasible. The assemblers knew a LOT of the common computer language of their profession, i.e. profanity.
But... while lots of USB drives come formatted in DOS style, and require a reformat, there shouldn't be any reason a USB2 drive won't boot a Mac Mini. It sure would boot my old iMac. After putting a HFS extended file system on it, and (OS 10.3+) enabling journalling, of course.
Grounding has to be ABSENT when working on high voltage (monitors or potentially backlights for LCDs), so the 'put grounding mats everywhere' approach is flawed. Keep ground straps around, and when reseating RAM use at least the left-hand-on-chassis precaution. No real need to be anal about it. While MOS inputs are sensitive, there are some ameliorating circumstances:
(1) except for RAM and CPUs, the removable parts of a computer have I/O pins (with both a MOS input and a transistor OUTPUT connected together) and that means very little sensitivity to static. Video cards from the plastic bins at generic-salvage-mart usually work fine.
(2) since the 1970s when static charge clobbered lots of expensive hardware, the manufacturers have learned to protect their products. It's normal to pass a 2000V static test (but they still don't recommend it).
(3) many materials (cardboard, real wood, most kinds of paint) are static dissipative anyhow, and the exceptions (shoe soles, nylon carpet) react well to some coatings (spray-on Downy fabric softener is a common recommendation). Here in Seattle, we have humidity which helps, too.
When a CDROM is packed in a sleeve with an antistatic precaution label on it, you KNOW something is over-the-top in the warnings about static. It's like the magnet I bought, that had a little pamphlet explaining it should always be used with eye protection... Electronic parts suppliers put warnings everywhere about static, just like machine tool suppliers put warnings everywhere about eyeshields. Those warnings often exceed the underlying dangers.
From a few years spent tech-ing, I'd say there's no more frequently used tool than frosty Scotch tape. Rip off a couple of inches, bend over a bit at the end (for a handle) and you have a removable label. Stick it on the components you yank, and write on it with a Sharpie (felt-tip fine point marker). Bingo, no more confusion over which memory you have tested and which you haven't. Remove the labels when the job's done, or leave 'em on the dead bits so your junk pile is documented. When the user wants to see the dead part, you're so ORGANIZED, even your mom would be proud.
An early step is going to be visual inspection, so a small paint brush (to loosen the dust) and a vacuum (to remove it without putting your lungs at risk) are key. Get a jeweler's loupe. Seeing a scorch mark, a burn blister, or oozing capacitor is quicker than functional testing. In case of intermittent faults, ANYTHING is quicker than functional testing.
We used Wiremold brand power strips (with one AC socket per 6 inches) on the leading edge of an above-bench shelf, always accessible. That was nearly enough.
Keep some chemicals on hand; Caig Labs stuff for contact enhancement, methanol (or denatured ethanol) for most plastics, Windex (low-residue waterbase cleaner) for just about everything, threadlocker for hardware that isn't supposed to come loose, maybe MEK (lacquer thinner) to be applied with a small cotton swab away from plastic parts... And I presume any tech knows to have freeze mist and WD-40 and canned air (or a compressor and blowgun) around. Low-residue wetting agent (Kodak Photo-Flo) in few-drops-in-a-quart-of-water solution, and water soluble solder flux (apply with an artist's camel-hair brush) are useful. One odd item was brake fluid; it's formulated to be very friendly to rubber items, makes a great cleaner for rubber rollers and was VITAL in the old days with impact printers.
Waterless handcleaner and a toothbrush will brighten up a case. Work it in, then let sit a few minutes before you wipe up. For a smoker's computer monitor, I've done wonders with a tiny bit of lye in soapy water (not for the metal or electronic parts). Gives you something to do while listening for hard disk misbehavior.
Boxes or ashtrays or muffin tins for small parts, and a tool organizer/rack for all the most frequently used tools. Don't hide tools in a box unless they're used infrequently, or you need to drag 'em onsite.
Finally, we used a database to keep our notes, including the computer's serial number and customer contact info. It was a very fine way to look into the past history of a machine, and anyone who answered the phone could relay the job status after checking the database record.
As a two-box solution, consider getting EyeTV (it plugs into/powers from firewire port).
The only real media-machine weakness of the mini is the disk drive; you might want to invest in a network hard disk (park it in the attic and run it via Airport). A modest month or two of haven't-watched-yet TV can suck up gigabytes fast!
Technically, that's not the whole truth. Carbon in graphite form is a semimetal at room temperature (has nearly zero temperature coefficient of conductivity); at lower temperature it is a semiconductor, at higher temperature it is a conductor. The high temperature behavior is similar to a semiconductor 'going intrinsic' (Germanium does this at an inconveniently low temperature, Silicon goes to several hundred Celsius first).
No one used graphite semiconductors much because the lattice is VERY strong in one dimension and weak in another; diffusion of impurities is quixotic. Also, fracture can occur easily. Nanotubes are different in interesting ways from graphite in these regards.
It's common (or used to be) to use the abrupt resistance rise of a carbon resistor to implement dipstick functions for liquid nitrogen containers. It's also possible to dope graphite for very high electrical conductivity (I've heard it can exceed that of copper). Both these are semiconductor electronic applications, though crude ones.
Re:Summary of tech advantages?
on
DECnet Isn't Dead
·
· Score: 3, Informative
Excellent question! Piddling little things like performance and does-my-router-know-it aside, the DECNet wasn't just a protocol, it was an enhanced user experience.
Instead of just transferring files, you could refer to a file on a foreign computer by name (a facility similar to our DNS (domain name system)). The network access was transparent.
So, every file open of "file" opened the file in your current default directory. Open "directory/file" and you can get the file in a subdirectory Open "disk:directory/file" and you get the file on a specific
rooted filesystem (other disk drive, usually) And, open "outofstate::disk:directory/file" and you have
access to any known node (other computer) whose disk and directory are readable (permitted) through the network.
The beauty of it is, there's no need to recompile the program, just to feed it the string (filename and other info all go into the same OPEN command).
Everyone using the internet with named URLs (universal resource locators) and DNS (domain name service) has similar capabilities nowadays, but DECNet users had it two decades ago. And they had it in ALL cases of file access. You could tell the help utility to read helpfiles from Stanford's SSRL physics lab, or tell the print output to go to a teletype in Maine.
And DECNet used (originally) mainly LAT networking protocols, not TCP/IP, because it predates the internet; I have a short stack of LAT network boxes that don't know TCP/IP, but they'd be hard to replace this week (and they're all 10base2 or somesuch, which is another issue...). There's nothing intrinsically LAT-based about the DECNET, it's just the historically original pairing; I presume DECNet and TCP/IP are mainly cooperative these days.
No, much of the attention directed at the power supply is at the filtered-input end (an extra RF filter at the power cord would be a similarly useful addition, of VERY limited utility). The 'better diodes' just make the filtered power a few dB quieter, the subsequent switching and linear regulation stages do that work better and aren't much affected.
And, an op-amp isn't noisy or distorted because it's cheap, it's noisy or distorted because it was chosen incorrectly for the application. Buying an expensive op amp is just making a new (uninformed) choice. There aren't any low signals at the analog end, the signal is HUGE coming out of the DACs, and no special amplifier is required to suppress the noise (because there isn't much, and inexpensive IC amplifiers don't add enough to matter).
The most important effect such tweaks could have: better stereo separation (I doubt you'll hear it).
Ahh, the goes-black symptom! That silly connector to the deflection yoke, that rarely soldered securely to the main video board-- we resoldered and replaced LOTS of those back at the computer maintenance shop.
Sometimes the bad connection was tolerated for years before it got to a shop; I've seen the heat buildup turn the (white) nylon connector shells black. It's an extra $5 for a replacement connector...
And my old 128k is still workable; it got the MacPlus upgrade (with a SCSI bus), then a CPU upgrade, and internal fan and hard disk... I kept it going over 10 years before I replaced it with a newer model.
If a secret, undocumented feature WERE to be used by, for instance, Microsoft, from an Intel CPU, this would be prima facie evidence of collusion in restraint of trade.
It's illegal in this country for one or two companies to get together in an attempt to freeze out (for instance) Linux or any other 'competitor'. Intel giving extra bennies to Microsoft (and excluding Lindows) is worth a lot of attention from federal prosecutors.
Similarly, if a card maker communicates his requirements to one or more 'other' companies, he HAS to give similar information to ALL other companies,
So, the legal department insists (quite properly, in their limited view) that the tech departments NOT give out info except for a strict known package.
What IBM did with the original PC, that made it popular, was to give out EVERYTHING in that known package, the Technical Reference volume, of the Personal Computer Hardware Reference Library.
And what made the PS/2 machines such a commercial flop, was that it wasn't similarly supported. No one could make cards except by 'partnering' with IBM.
Macintosh users like the user interface, the 'insanely great' inclusion of features that you can't get in a Wintel box, and will do just fine with whatever CPU comes out next year. Macs will boot from the iPod if asked to, and Wintel boxes won't. Mac users like that sort of thing.
Developers weren't stumped when the 68000 family expanded, had no difficulty getting good functionality from PowerPC and were even getting the hang of G4/G5 vector processing. Adding Pentium target CPUs is going to be mainly a matter of telling the compiler to be prepared for another variant. It's not likely to be a deal-breaker.
I suspect C code will recompile and cover the next machines just fine.
What folk like the games developers REALLY care about is assurance of their library support; a game delivers its video to Aqua or OpenGL nowadays, not to a CPU.
Alas, while battery CELLS are quite similar, the mechanical packaging, connectors, and charging systems are dissimilar. Many laptop batteries have their own charge controllers built in, and Macintosh batteries have been known to need reset procedures and firmware updates.
There aren't many success stories of third parties that make competitive replacement batteries for laptops.
The design of a battery socket can be protected by patent, and WILL be if the computer manufacturer thinks it will make them a nickel. Batteries will become reliable enough to last a lifetime before they will be allowed to become generic, I suspect. No smiley.
Remember, OS X has that old Mach kernel that worked on the Next with multiple CPU targets on all the compilers.
Mach uses data-passing protocols that specify endianness. It's intended at the ground level for systems even including multiprocessor with some big-endian and some little-endian all passing stuff to each other.
If you have a problem keeping track of which files were written how, that's a metadata issue; on a Mac, you're pretty well covered.
If you have a routine that wants to pass other-endian data, the CPU can do end-alterations and the kernel has the habit of keeping track of end-ness. So, on a Mac, you're pretty well covered.
Slashdot Mirror
That's a good point, that voltage rating is an important issue. But it isn't
because of expense of high voltage insulation, it's because THAT part
of the power handling is safety-critical (electric shock hazard).
When you've converted to nonlethal voltages (under 50VDC, usually)
the remaining safety concern is does-it-start-a-fire, and simply limiting
the current completes that task (fuses, fusible resistors, circuit breakers, some
kinds of active 'foldback' voltage regulator schemes, all are available strategies).
First, power requirements for common peripherals (small hard disk drives) exceed the
standard capacity of USB (0.5 A at 5VDC); there are some laptop plugin drives that
come with TWO USB cords, because it needs power from both ports to spin up.
Second, high power variants of the USB port (Apple tried this, on the cube, for
some high-power speakers) give rise to odd incompatibilities. Breaking
the standard is a bad idea. Trust me. I've diagnosed/dealt with it and don't
care for a repeat of THAT.
Third, there are devices that need other voltages (like EIA-232 serial ports) and
the 'universal' +5V is just plain wrong. Converters are used, of course, but
the converter isn't notably simpler than an AC power brick; you save on cables,
not on hardware. Cheapo converters for EIA-232 are energy inefficient, but
there isn't much energy required, so that's OK.
Fourth, it's cheap to make high voltage parts and expensive (in terms of chip area
and yield from a semiconductor processing plant) to make high current ones; if
you knew you were gonna convert the DC voltage, your choice of input voltage
would be higher, 48VDC (about like telephone company power handling) or the
new automotive standard, 42V. Power-over-Ethernet is standardized
at 48VDC (negative voltage) for this reason. The price difference makes little impact on
the customer, but some pennypincher engineer will always choose for you.
Then, the marketing department won't show the brick in the pretty boxtop picture.
Firewire does use higher voltage power (12 to 24V), with similar current (so the wire
isn't stiffer than USB wires). As a result, Firewire power DOES support a hard drive
with appropriate conversions inside the peripheral.
I'd like to see The Snarkout Boys and the Avocado of Death on
the silver screen, myself. Or, if that's too challenging,
how about Yobgorgle, Mystery Monster of Lake Ontario.
Imagine, cinematically, the chicken-suit car purchase and a pink submarine
lurching onto shore in search of a really GOOD roast-beef sandwich.
Daniel Pinkwater is da bomb.
Favored books don't always work on screen (Dune didn't, IMHO, and
before that I was disappointed in the anime version of Lensman).
I figure I'm due for one that DOES work in film.
To 'shield from magnetic fields' is generally the same thing as 'generate an
opposing magnetic field'; that means the shield materials ARE affected,
and are in fact somewhat magnetized, to create the shielding
effect.
The best shield materials are superconductors (which only exist at low
temperatures). The most common magnetic shield materials are soft
iron alloys (Permalloy and Mu-Metal are brand names). Shielding
from rapidly-changing magnetism is easier, most electrical conductors
will do this (but superconductors do it for constant magnetic fields
as well as changing ones). A weak shielding effect is called diamagnetism,
and is interesting in its own right. Did you know that water is repelled
from a magnetic field? Water is diamagnetic (weakly). Brass is more highly
diamagnetic.
Yes, it should be possible. There are positive-temperature-coefficient
resistors that 'switch' current off when heated. You could make a crude
relay with heater resistors as the input and a PTC resistor as the
output conduction channel. Negative-temperature-coefficient
resistors are also available, you could do something like CMOS using both.
I've seen some thermal effect in silicon chips, related to heat changing
the input current of (for instance) op amps. It makes a big effect, but
only at low speeds (two cycles per second, in the case that bit me).
Silicon is a very good heat conductor, if you were to make thermal
gates on such a substrate it'd be one per chip OR you'd have to
get such high switch rates that the heat 'leakage' current was negligible.
But crude logic IS still useful; there are intrinsically-safe electric heaters
using a ceramic that stops heating when it is just under the ignition temperature
of common home furnishings (using a positive-temperature-coefficient
heating element to lower the power consumption when the target temperature
is achieved).
It's actually beneficial that a single 'gate' element can perform AND, OR, and INVERT
functions all in one stage. The early TTL won over other logic designs in part because
the basic gate used multiple emitters on the input transistor to get an AND function,
and multiple input transistors to get the OR function. That meant that the delay
and complexity character of AND and OR were the same, and that the complex function
of AND/OR/INVERT was available as a fast multiplexer, with the same characteristics
as a simple NAND. There was a brief attempt to use expandable gates (making
the connection point after the input transistor available on an external pin,
which was NOT TTL-logic-level compatible), but it didn't catch on.
CMOS, on the other hand, had input impedance and delay differences in the AND and
OR and other gates, so the whole 4000 series CMOS logic family only became
trouble-free to use AFTER THEY BUFFERED THE WHOLE FAMILY with an extra inverter
(and consequently extra time delay). Buffered (4000B series) is the common small
scale CMOS you see today, the unbuffered (4000A series) has been sidelined.
From a circuit-design viewpoint, the AND/OR/INVERT is a very good starting element,
for a lot of reasons that only show up when some poor engineer is perspiring over his
timing budget...
The clear problem here, is that a President can direct a large number of
civil servants, but CANNOT ethically use those civil servants as
proponents for some political position or the other. The academic
community dislikes this particular chief executive because of
multiple instances of pressuring academics to adhere to a
Bush world view,
When pressed for an explanation, the official word is that the
manipulation of published works is 'part of a normal review process".
Of course, the NORMAL review process under other administrations
was 'peer review' and this administration applies 'political officer spin
doctor' review. Our president also thinks Intelligent Design should
be taught in schools, which gives you an hint of what our scientists in
government agencies are dealing with...
This argument, "one big problem with radioactivity is that people can't see it"
w w.amazon.co.uk/exec/obidos/ASIN/0878430024&e=9797
is insidious and evil. It's the sort of fall-back statement that CAN NEVER BE
COUNTERED, but has no other merit, logically, whatsoever.
It's like the old story of an argument between a white politician and a black one,
Bystander to reporter: "What's happening?"
Reporter to bystander: "He's still black."
It is evil to think energy policy depends on the limitations of human sight and taste
as applied to sensing radioactivity, just as it is evil to think a politician is
at fault for his race (or any other non-negotiable accident of birth).
Both arguments are flawed in the same way, and both do clarity an injustice.
For a good (though depressing) treatment on the subject at length, check
out _The_Fight_Over_Nuclear_Power_ by Fred Schmidt et al. The foreword is VERY
insightful.
http://www.google.com/url?sa=U&start=8&q=http://w
It should be pointed out that the persistent cookie mainly is undesirable
because folk who bulk-buy surplus computer hard drives could search
for nuggets of sensitive info and be guided (as to the interests of
the original user) by those cookies.
Or folk who bulk-burgle hardware bits.
If all the cookies are from Microsoft or Yahoo or music fan sites, it's not
likely that candid comments on foreign policy from government
insiders are on THAT hard drive. But a drive with cookies from NSA and
CIA, might indicate some value of further data mining efforts.
It's a minor indication, certainly, but plugging small leaks is still worthwhile.
So, I don't see a policy intended to benefit the public, per se, but rather a policy
intended to obscure the historic record left on a filesystem. You have to
recall that NSA isn't the only data security threat out there... then
it all makes some more sense.
It's said that president Lincoln often composed letters to or about folk he
was upset at, then carefully filed them away and never sent them. I'd like
to see the hard drive text files for some contemporary Republicans to
see how their habits mirror those of the party founder.
Heck, a love-letter from Bush senior to Saddam could plausibly
be on a discarded disk drive now. Check your disk deadpile!
I've replaced backlights in powerbooks, both with official Apple parts (it has been
some years since THAT was available) and with generic lamps from third-party
suppliers.
Firstly, remember the lighting uniformity is HARD TO ENSURE. For the thinnest
displays (like modern Powerbooks) it's unlikely you can even FIND a source for the
lamp (2mm tubes are common, the available units are usually 4mm or the wrong length
or both), and if you do find it, getting the foam/mylar/backplate sandwiched
after replacement is going to be a chore. Expect a splotchy result.
The most recent display I had any good luck with was one of the Powerbook 500 series
(about 1995 vintage). That required a little work with a Dremel tool to fit the
available lamp (which was about 5mm too long), and took a bit of care during
disassembly. The plastic display bezel got brittle with age by the time
the unit was both out of warranty and failing to light up.
The 'supported' solution is to ship the unit to the manufacturer for
rebuild. Presumably, they DO have the right parts and a lot of patience.
My advice: pay the $400 (or whatever).
There's a lot less toxin in the lamp than in your last filling. More worrisome is the
wiring to the tube, which wants about 2000V to start. The wires are funny, rubbery
things, possibly a hightech silicone.
One very useful tool was a spare power supply; if your backlight was on a connector,
you can test it on known-good power that way, and an inverter replacement is a LOT
easier than lamp rebuild.
Apple is a VERTICAL organization. The quality of the components
relates to their INTEGRATION, both with hardware and software/drivers.
One post commented on the 'generic RAM and crappy video card'--
Apple's RAM is higher quality than much commodity RAM (I repaired lots
of Macs with bad RAM, usually not the Apple-supplied stick); but it
doesn't show when you open the box.
Apple's video card supplies power and USB to their monitors; no PC
vendor has a monitor powered without an AC cord, nor a USB hub
conveniently on the monitor, because the designers of the units
never sat down together. Vertical organizations can support
this kind of coordination, and the Wintel stack-of-blocks teams
are at best gonna stay a few years behind.
You can castigate the 'crappy video card' all you want, but the imaging
model in the MacOS fully USES the graphic programmability of
that card, for ALL applications, and supports it (virtual video memory)
in ways the Wintel boxes don't. This happens because the OS
designers and the card programmers all talk to each other.
Before going into production.
Booting from USB disks and from network volumes and Firewire disks,
and hotswap mice and keyboards, and target disk mode (a crashed
computer can be rebooted so as to serve its disks out as Firewire
volumes to a known-good computer) are all Macintosh features the
Wintel world can only copy as an afterthought (if at all).
You can castigate the 'crappy video card' all you want, but the imaging
model in the MacOS fully USES the graphic programmability of
that card, for ALL applications, and supports it (virtual video memory)
in ways the Wintel boxes don't. This happens because the OS
designers and the card programmers all talk to each other.
Apple's upcoming x86 boxes may lack some PC essentials (PS/2 keyboard
ports and onboard floppy controllers), but the quality that matters,
the good functional integration of features, will still be there.
And people won't buy Apple machines to run Wintel software; that
would be pretty pointless. The full advantage of buying the Apple
machine lies in the integration of the parts, not in their chrome
plating.
It's a great plan, but there's a major safety issue. House wiring is self-extinguishing
for AC arcs across the wires. It isn't similarly safe for DC arcing.
DC fuses are harder to make safe, as well. At low voltages (12V like in a car) that's not
too bad. At higher voltages (28V like in aircraft and trucks) it's a well-studied and
solved problem. Even at telco (telephone company) voltages (nominally 48V,
range 36-72V is the typical design limits), it can be handled.
But, the practical UL-approved devices in your home have started to have
little switching power converters, that take the 120VAC and/or 250VAC and
convert it to your low voltage right inside the appliance,
and that's nearly as convenient as having multiple power grids and WAY more flexible
in terms of total power available and voltage-output flexibility.
Like you, I've thought of backup DC power on a separate bus (and use some fixtures
like trailers and RVs use that just run on my backup battery). But, it's too much
trouble to do safely. And, WAY too much trouble to do any other way.
The only way data centers can benefit is because they have UPS issues (like telco does)
and the battery backups and their DC layer has to be common to the whole roomfull
of equipment anyhow. Might as well run power-distribution wiring with
the low-voltage DC; the wires only stretch one room in distance, the wire-resistance
losses aren't prohibitive.
There are multiple uses for a slowed signal; you can combine it with the un-delayed
signal and make filters, like is done with SAW filters (but those use surface acoustic waves,
and are not silicon-compatible). You can also make some kinds of shift register
VERY simply by sending the signal out into the delay and picking it up when you
need it. And a delay of a clock signal often makes a computer more reliable (designing
high speed compute devices, this is OFTEN a vital consideration).
The split/multiple delay/combine scheme for (for instance) radio signals is
a very powerful tool, and is why a complicated-looking antenna can work
so well. And, why a rabbit-ear antenna can take a lot of tweaking to
get your idiotbox to receive Red Green.
For major processing of data, it was common practice in the old days to tweak the
interconnect wiring to make the correct time delay. Seymour Cray reported (of the
Cray-1 supercomputer) that the interconnect in the central core of the computer
was hand-wired by (slender women) assemblers who used cut-to-measure lengths of
twisted pair, so that all the signals had the appropriate settling time before the clock
arrived and latched the data. The computer was a cylindrical hole with draped wiring
all over its interior, with spokes out that housed the cooled ECL logic modules.
To keep the Cray quick, the cylindrical core was as small as feasible. The assemblers
knew a LOT of the common computer language of their profession, i.e. profanity.
But... while lots of USB drives come formatted in DOS style, and
require a reformat, there shouldn't be any reason a USB2 drive won't
boot a Mac Mini. It sure would boot my old iMac. After
putting a HFS extended file system on it, and (OS 10.3+) enabling journalling,
of course.
Grounding has to be ABSENT when working on high voltage (monitors or potentially backlights for LCDs), so the 'put grounding mats everywhere' approach is flawed. Keep ground straps around, and when reseating RAM use at least the left-hand-on-chassis precaution. No real need to be anal about it. While MOS inputs are sensitive, there are some ameliorating circumstances:
(1) except for RAM and CPUs, the removable parts of a computer have I/O pins (with both a MOS input and a transistor OUTPUT connected together) and that means very little sensitivity to static. Video cards from the plastic bins at generic-salvage-mart usually work fine.
(2) since the 1970s when static charge clobbered lots of expensive hardware, the manufacturers have learned to protect their products. It's normal to pass a 2000V static test (but they still don't recommend it).
(3) many materials (cardboard, real wood, most kinds of paint) are static dissipative anyhow, and the exceptions (shoe soles, nylon carpet) react well to some coatings (spray-on Downy fabric softener is a common recommendation). Here in Seattle, we have humidity which helps, too.
When a CDROM is packed in a sleeve with an antistatic precaution label on it, you KNOW something is over-the-top in the warnings about static. It's like the magnet I bought, that had a little pamphlet explaining it should always be used with eye protection... Electronic parts suppliers put warnings everywhere about static, just like machine tool suppliers put warnings everywhere about eyeshields. Those warnings often exceed the underlying dangers.
From a few years spent tech-ing, I'd say there's no more frequently used tool than frosty Scotch tape. Rip off a couple of inches, bend over a bit at the end (for a handle) and you have a removable label. Stick it on the components you yank, and write on it with a Sharpie (felt-tip fine point marker). Bingo, no more confusion over which memory you have tested and which you haven't. Remove the labels when the job's done, or leave 'em on the dead bits so your junk pile is documented. When the user wants to see the dead part, you're so ORGANIZED, even your mom would be proud.
An early step is going to be visual inspection, so a small paint brush (to loosen the dust) and a vacuum (to remove it without putting your lungs at risk) are key. Get a jeweler's loupe. Seeing a scorch mark, a burn blister, or oozing capacitor is quicker than functional testing. In case of intermittent faults, ANYTHING is quicker than functional testing.
We used Wiremold brand power strips (with one AC socket per 6 inches) on the leading edge of an above-bench shelf, always accessible. That was nearly enough.
Keep some chemicals on hand; Caig Labs stuff for contact enhancement, methanol (or denatured ethanol) for most plastics, Windex (low-residue waterbase cleaner) for just about everything, threadlocker for hardware that isn't supposed to come loose, maybe MEK (lacquer thinner) to be applied with a small cotton swab away from plastic parts... And I presume any tech knows to have freeze mist and WD-40 and canned air (or a compressor and blowgun) around. Low-residue wetting agent (Kodak Photo-Flo) in few-drops-in-a-quart-of-water solution, and water soluble solder flux (apply with an artist's camel-hair brush) are useful. One odd item was brake fluid; it's formulated to be very friendly to rubber items, makes a great cleaner for rubber rollers and was VITAL in the old days with impact printers.
Waterless handcleaner and a toothbrush will brighten up a case. Work it in, then let sit a few minutes before you wipe up. For a smoker's computer monitor, I've done wonders with a tiny bit of lye in soapy water (not for the metal or electronic parts). Gives you something to do while listening for hard disk misbehavior.
Boxes or ashtrays or muffin tins for small parts, and a tool organizer/rack for all the most frequently used tools. Don't hide tools in a box unless they're used infrequently, or you need to drag 'em onsite.
Finally, we used a database to keep our notes, including the computer's serial number and customer contact info. It was a very fine way to look into the past history of a machine, and anyone who answered the phone could relay the job status after checking the database record.
As a two-box solution, consider getting EyeTV
(it plugs into/powers from firewire port).
The only real media-machine weakness of the mini is the
disk drive; you might want to invest in a network hard
disk (park it in the attic and run it via Airport).
A modest month or two of haven't-watched-yet
TV can suck up gigabytes fast!
Technically, that's not the whole truth. Carbon in graphite form
is a semimetal at room temperature (has nearly zero
temperature coefficient of conductivity); at lower temperature
it is a semiconductor, at higher temperature it is a conductor.
The high temperature behavior is similar to a semiconductor
'going intrinsic' (Germanium does this at an inconveniently
low temperature, Silicon goes to several hundred Celsius
first).
No one used graphite semiconductors much because the
lattice is VERY strong in one dimension and weak in another;
diffusion of impurities is quixotic. Also, fracture can occur
easily. Nanotubes are different in interesting ways from
graphite in these regards.
It's common (or used to be) to use the abrupt resistance rise
of a carbon resistor to implement dipstick functions for
liquid nitrogen containers. It's also possible to dope
graphite for very high electrical conductivity (I've heard it
can exceed that of copper). Both these are semiconductor
electronic applications, though crude ones.
Excellent question! Piddling little things like performance
and does-my-router-know-it aside, the DECNet wasn't
just a protocol, it was an enhanced user experience.
Instead of just transferring files, you could refer to a
file on a foreign computer by name (a facility similar
to our DNS (domain name system)). The network access
was transparent.
So, every file open of "file" opened the file in your
current default directory.
Open "directory/file" and you can get the file in a subdirectory
Open "disk:directory/file" and you get the file on a specific
rooted filesystem (other disk drive, usually)
And, open "outofstate::disk:directory/file" and you have
access to any known node (other computer) whose disk
and directory are readable (permitted) through the network.
The beauty of it is, there's no need to recompile the program,
just to feed it the string (filename and other info all go into
the same OPEN command).
Everyone using the internet with named URLs (universal
resource locators) and DNS (domain name service) has
similar capabilities nowadays, but DECNet users had it
two decades ago. And they had it in ALL cases of file
access. You could tell the help utility to read
helpfiles from Stanford's SSRL physics lab, or tell the
print output to go to a teletype in Maine.
And DECNet used (originally) mainly LAT networking protocols,
not TCP/IP, because it predates the internet; I have a short
stack of LAT network boxes that don't know TCP/IP, but
they'd be hard to replace this week (and they're all 10base2 or
somesuch, which is another issue...). There's nothing
intrinsically LAT-based about the DECNET, it's just the
historically original pairing; I presume DECNet and TCP/IP
are mainly cooperative these days.
No, much of the attention directed at the power supply
is at the filtered-input end (an extra RF filter at the power cord
would be a similarly useful addition, of VERY limited
utility). The 'better diodes' just make the filtered power
a few dB quieter, the subsequent switching and linear
regulation stages do that work better and aren't much
affected.
And, an op-amp isn't noisy or distorted because it's cheap,
it's noisy or distorted because it was chosen incorrectly for
the application. Buying an expensive op amp is just making
a new (uninformed) choice. There aren't any low signals
at the analog end, the signal is HUGE coming out of the DACs,
and no special amplifier is required to suppress the noise
(because there isn't much, and inexpensive IC amplifiers
don't add enough to matter).
The most important effect such tweaks could have: better
stereo separation (I doubt you'll hear it).
Ahh, the goes-black symptom! That silly connector
to the deflection yoke, that rarely soldered securely to
the main video board-- we resoldered and replaced LOTS
of those back at the computer maintenance shop.
Sometimes the bad connection was tolerated for years
before it got to a shop; I've seen the heat buildup
turn the (white) nylon connector shells black. It's an extra $5
for a replacement connector...
And my old 128k is still workable; it got the MacPlus upgrade
(with a SCSI bus), then a CPU upgrade, and internal fan and
hard disk... I kept it going over 10 years before I replaced it
with a newer model.
If a secret, undocumented feature WERE to be used by, for
instance, Microsoft, from an Intel CPU, this would be
prima facie evidence of collusion in restraint of trade.
It's illegal in this country for one or two companies to get
together in an attempt to freeze out (for instance) Linux
or any other 'competitor'. Intel giving extra bennies
to Microsoft (and excluding Lindows) is worth a lot
of attention from federal prosecutors.
Similarly, if a card maker communicates his requirements
to one or more 'other' companies, he HAS to give similar
information to ALL other companies,
So, the legal department insists (quite properly, in their
limited view) that the tech departments NOT give out info
except for a strict known package.
What IBM did with the original PC, that made it popular,
was to give out EVERYTHING in that known package, the
Technical Reference volume, of the Personal Computer
Hardware Reference Library.
And what made the PS/2 machines such a commercial flop,
was that it wasn't similarly supported. No one could make
cards except by 'partnering' with IBM.
Macintosh users like the user interface, the 'insanely great'
inclusion of features that you can't get in a Wintel box,
and will do just fine with whatever CPU comes out next year.
Macs will boot from the iPod if asked to, and Wintel boxes
won't. Mac users like that sort of thing.
Developers weren't stumped when the 68000 family expanded,
had no difficulty getting good functionality from PowerPC
and were even getting the hang of G4/G5 vector processing.
Adding Pentium target CPUs is going to be mainly a matter
of telling the compiler to be prepared for another variant.
It's not likely to be a deal-breaker.
I suspect C code will recompile and cover the next
machines just fine.
What folk like the games developers REALLY care about
is assurance of their library support; a game delivers its
video to Aqua or OpenGL nowadays, not to a CPU.
Alas, while battery CELLS are quite similar, the mechanical
packaging, connectors, and charging systems are dissimilar.
Many laptop batteries have their own charge controllers
built in, and Macintosh batteries have been known to
need reset procedures and firmware updates.
There aren't many success stories of third parties that
make competitive replacement batteries for laptops.
The design of a battery socket can be protected by patent,
and WILL be if the computer manufacturer thinks it will
make them a nickel. Batteries will become reliable enough
to last a lifetime before they will be allowed to become
generic, I suspect. No smiley.
Remember, OS X has that old Mach kernel that worked
on the Next with multiple CPU targets on all the compilers.
Mach uses data-passing protocols that specify endianness.
It's intended at the ground level for systems even including
multiprocessor with some big-endian and some little-endian
all passing stuff to each other.
If you have a problem keeping track of which files were
written how, that's a metadata issue; on a Mac, you're pretty
well covered.
If you have a routine that wants to pass other-endian data,
the CPU can do end-alterations and the kernel has the habit of
keeping track of end-ness. So, on a Mac, you're pretty well covered.