Slashdot Mirror
X-Rays Of A TiBook's Interior
A reader writes: "A fine application of expensive medical equipment: producing neat desktop pictures by taking an x-ray of the guts of a PowerBook G4. Guy Mullins has the details." The actual photos are on a separate site.
Don't xrays wipe drives? I think I can see the hard drive in the xray - in the middle at the bottom - the CD drive is on the left, and the batteries are on the right I reckon.
Anyone know what effects XRays have on magnetic media? I always used to post floppies with a 'magnetic media, do not xray' sticker on em?
Fantastic images tho.
I've had a couple of assembled battery packs apart because off the shelf was not available and each time it was an assembly of standard batteries soldered together.
OTOH I'm sure there are some custom batteries out there.
Generally yes, though some are using prizmatic cell shapes but they are more costly than the good ol' cylindrical cell. If you go with a non standard cell size or shape it costs more per unit due to lack of volume. The standard cell sizes are also available from many different manufacturers. That means you have second and third sources available in case your manufacturer of choice fails to meet your demand for some reason.
TiGutz in Blue
;p
TiGutz plain
The 3m and 9m files will half to wait for later
Sexy stuff.
Computational Madness in a round package.
The reason for this is that a cell puts out about 2 volts. To get a higher voltage, you have to connect several cells in series to make a battery. For a car, you need six cells to make twelve volts. For a telephone exchange, you need 24 cells to make about 50 volts. You can also connect several batteries together in parallel to get more current or more power.
...
No. X-Rays are just light. If it caused a problem, you wouldn't be able to take your computer on a plane. Nothing in a computer is affected by x-rays.
I once had to obtain a new battery for P75 laptop and that battery could not be had from anywhere. However, the cells were in Batteries Plus' catalog and they were able to rebuild the battery for me.
I used to work as a technician for a firm that rented environmental instrumentation and we recelled batteries all of the time. It is a common practice for more than just laptops.
And of course a 9V battery is also just 6 cells underneath the outer covering. Your car battery is 6 cells but it is a lead-acid battery which produces ~2V per cell whereas normal batteries (AAA, AA, C, D) use a dry cell which produces ~1.5V. NiCad cells are ~1.2V.
The TiXray.orig file is in DICOM format, and I imported it into GraphicConverter. If anyone's interested, this is what GraphicConverter put in the comments:
Image Type: ORIGINAL\PRIMARYStudy Date: 20010424Acquisition Date: 20010424Image Date: 20010424Study Time: 154340Acquisition Time: 154532Image Time: 154531Accession Number: TiModality: CRManufacturer: Lumisys Institution Name: Institution Address: Referring Physician's Name: Referring Physician's Telephone Numbers: Station Name: OPACS_SENDERStudy Description: Name of Physician(s) Reading Study: Operator's Name: Administrator Admitting Diagnoses Description: Manufacturer's Model Name: Lumisys LS135 Patient's Name: PowerBook^TitaniumPatient ID: Apple Patient's Birth Date: 20010101Patient's Sex: O Other Patient IDs: Ethnic Group: Additional Patient History: Body Part Examined: Device Serial Number: clt35403.datCassette Orientation: PORTRAITCassette Size: 35CMX43CM Relative X-ray Exposure: 1713View Position: Study ID: c0a865080gq5m8Series Number: 1 Image Number: 1 Photometric Interpretation: MONOCHROME2 Pixel Spacing: 0.172\0.172 Study Priority ID: MED Requesting Physician:
$x='S24;r)>63/* h@<5+oZ)32"5cz';$me='phroggy'x$];
$x=~y+ -xz+\0-Tx+;print$_^chop$me for split'',$x;
Figuring out why, and going back to figure out what other comments we might have missed, is one of our priorities for this coming week.
To the trolls who started this meme, if you are interested in getting these problems fixed rather than just raising a fuss, the next time you find something like this, please submit a SourceForge bug. Thanks. Meanwhile, extended discussion of Slashdot's bugs on a story that's not about testing the Slash code is offtopic and should be moderated as such.
When I was in college, the probability was closer to 1 in every 1000 scans. These days, the airport xray machines are much stronger than back then. I can't cite a new probability, because I don't know the new radiation strength, but it must be lower than 1 in 1000.
.2um technology (that tells us that the transistor channel will be .2um x .4um in EEPROM memory cells), then your bulk EEPROM is probably using a .5um or 1um process. That means that not only does your bulk memory have far more EEPROM transistors, but that each one is 6.25 to 16 times as large. For this reason, we can ignore the probability of getting a one bit error in your CPU when discussing the likely places for errors to occur. But a rigorous discussion would require that we pay attention to all EEPROM cells in a computer.
The critical question is, "Will you notice the damage?" We're talking about a one bit error, here.
Take your average PC. You've got pretty much three firmware chips in there: one for the system BIOS, one for the hard drive firmware, and one for the video controller. No matter which firmware suffers the error, there's first a question of whether or not the error is in a memory cell that ever gets accessed, and second whether or not the error changes the value in a way that makes a difference. If we're talking about one of those lame bootup graphics that a lot of manufacturers like these days, you'll probably never notice a one bit error. And if we're talking about code that drives a piece of hardware that you don't have installed (or dont use) you won't notice that error either. But if you get an error in your POST code, you're dead. System BIOSes these days are pretty fluffy. Lots of extraneous stuff in there. Video firmware is also pretty fluffy, but not nearly as bad as system BIOS. And drive firmware is quite tight. Almost any error you get in there will be in some code that gets executed.
But on the other hand, the Intel opcode set is full of lots of unused bits. There are a whole lot of examples where flipping a bit from 0 to 1 doesn't change the opcode or operands. And recall that we're talking about errors that can only change a 0 to a 1; there is no way that xray damage could change a 1 to a 0. If the bit that gets damaged is already a 1 (~50% chance of that) then your dead memory cell is still functioning exactly the way you want... until it comes time to do a firmware upgrade.
So that brings us to the last two issues: given that damage occurs, and given that its noticeable, then how long will it take for you to notice it, and will you ascribe the damage correctly to the airport xray machine? Many users are perfectly happy to ascribe crashes and corruptions to Windows. And while Windows certainly accounts for more than its fair share of errors, on a machine that crashes once every 40 hours, are you going to notice and correctly assign a failure that makes it crash every 39h30m?
Lastly, we're not talking about a bit suddenly changing from 0 to 1. Normally a buried gate transistor will hold its charge for around 150 years (each transistor will be different). No insulator is perfect. You zap it with a single photon, and maybe you've chopped a year off that. Of course, a xray machine isn't going to output a single photon. It's going to bathe your machine in a tremendous number of photons. So maybe one airport xray machine will drop you from 150 years to 130 years (I'm pulling that 20 year figure out of my butt, pretty much, but it's within an order of magnitude of being right). Do it again, and we're down to 110 years (it'll always reduce the life by the same number: a single xray photon opens a single ionizied path in the silicon dioxide, and that single ionized path will saturate at a small fixed current). So after 8 hits in the airport xray machine, you've probably got a number of one bit errors. And then the discussion above about whether or not you'll notice the error and whether or not you ascribe it correctly comes into play.
PS: these days, most microcontrollers, PICs, CPUs, MMUs, and other assorted large chips also have EEPROM cells on board. But in any particular computer, the technology used to design the EEPROM cells in functional chips is usually 5 to 7 years ahead of the technology used to create EEPROM cells in memory chips. That means that if your CPU is using a
-- Nolite audere delere orbiculum rigidum meum.