Dan's Data letters #69Publication date: 30 October 2003.
Last modified 03-Dec-2011.
Recently I tried to upgrade my BIOS. Whoops - I flashed the wrong BIOS file. So my motherboard seems to be pretty much cactus. All part of the learning experience, it's an old computer.
I've read that you can buy pre-burned BIOS chips, and found a company, BIOS Recovery Center, that sells them.
Do you know of any Australian BIOS chip sources? And does this really work?
Yes, it works. You can just swap an appropriately burned chip of the right type into your computer and recover instantly from a bad flashing experience, provided you have a vaguely recent machine with a socketed BIOS.
You can make your own new BIOS at home, too, if you have access to another computer with the same motherboard. You power that machine up, remove its BIOS chip while it's running (which gives you 101 ways to destroy the second computer, but let's move on...), insert your own mis-flashed chip, and re-flash it. Many motherboard manufacturers even have handy-dandy Windows utilities to do this, these days, though the utilities don't necessarily work with older boards.
After the re-flash, you don't even need to swap the chip back again; since the donor machine now has its own freshly-burned working BIOS, you can just stick the donor PC's BIOS into the first PC's motherboard, and away you go.
This can even, sometimes, be done with motherboards of a different model that have the same kind of BIOS chip, provided the flash program lets you flash the "wrong" ROM image. Then, of course, you do have to swap the chip back again afterwards.
I didn't know of any Aussie BIOS chip sources (after I put this page up, a reader pointed me to OZFLASH), but some on-the-ball computer stores will do this for you, probably via the above method.
If you've got an old machine, though, they probably won't have a donor system available, and so probably won't be able to help you. And even if they can, you'll probably end up paying more than BIOS Recovery Center would charge you, even taking international shipping into account.
It's also worth checking with the company that made the motherboard; they may be happy to sell you a new chip cheaper than BIOS Recovery Center will!
The motherboard companies and BIOS Recovery Center, by the way, are not making new BIOSes by the second-motherboard method. They'll be using a proper EPROM burner, which is much simpler. A computer-connected burner should be able to create arbitrary BIOSes from the regular downloadable manufacturers' files, but most computer stores aren't hip enough to have one of those handy.
I'm interested in these supposedly 5.1 channel capable headphones. I would be truly interested in them, if they work, but I'm quite skeptical.
Multiple transducers in one set of headphones is not a daft idea - it's a shortcut to headphone surround that doesn't require fancy DSP diddling of the signal to insert fore-and-aft signals into a simple two-channel setup.
I want to put a PC CD-ROM drive in my car, but I'm not sure how to do the wiring. The car has a regular 12V output for the radio, and a ground, that makes three wires (+,-,G). The CD-ROM has a 12V and a 5V input and two ground wires (black), which makes four wires. I assume the black wires are minus-wires. I connected the car's power to the 12V input and in parallel I connected an adapter (12V in, 5.7V out). I connected the adapter's output to the drive's 5V input. I connected the ground of the car to the black cables of the CD-ROM. Then the CD-ROM started to smoke.
Could you give me the proper connection of the wires?
First up - every "negative" wire in a normal negative-ground car is a ground wire. They're all connected together, and to the frame of the car.
There are dedicated PC power supplies designed to work from 12V (and the 13.8V that car electrical systems deliver when the engine's running), with a full ATX wire harness on the other side. They're considerably more expensive than regular PC PSUs, though (an inverter and a normal PSU may cost you less!), and overkill for your situation (an inverter and normal PSU would be even more ridiculous, of course).
I'm not sure why you smoked the CD-ROM drive, but it could have something to do with 13.8V, and with the 5.7V (or maybe higher, under low load...) output of the other adapter. Properly regulated 5V and 12V should work fine; an electronics kit-builder could knock together a circuit to do this elegantly without too much trouble.
If you only want the CD-ROM drive for audio playback, though (and without a computer, that's all it's going to do), you'll probably find just buying a cheap car CD player to be a considerably better solution.
At some point in the past few weeks, I've noticed that if I run a command prompt window full screen, my monitor will go into standby mode. Switch back to Windows mode, and it pops back to life. I'm running WinXP on an Compaq Athlon system, with a NVIDIA GeForce2 MX400 (Personal Cinema).
I haven't done anything recently that I think would cause this, but then again cause and effect under Windows is always a little hard to guess.
This is probably because the computer's sending the monitor a refresh rate and resolution combination that it can't handle.
Full screen DOS text mode should be 720 by 400 pixels at 70Hz. If you've twiddled your display settings to, for instance, lock every possible resolution to 100Hz or something, then you may have locked DOS mode to the same refresh rate and your monitor may not be able to handle it.
Generally speaking, modern multiple-sync monitors can handle higher and higher refresh rates as the resolution drops, but there are often odd anomalies down at the bottom end of the resolution chart.
I think there's a good chance that Display Properties -> Settings -> Advanced -> Adapter -> List All Modes won't actually list 720 by 400 at all; it doesn't on my computer. RefreshForce may know about it, though.
After I sent Jason the above reply, he got back to me and said that RefreshForce didn't solve the problem, but did show him that the computer thought it had two monitors, a Compaq MV940 (which it really did have) and Default Display, which XP obstinately refused to stop using.
In reply to which - here's a possible way to kill dumb extra phantom monitors that can't be deleted by normal means. Edit the registry ( Start -> Run -> type "regedit" ), and navigate to HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Enum\DISPLAY ; that's where the keys for monitors are kept. Provided XP hasn't decided they're all in use and untouchable, you should be able to delete any extras.
Since you mentioned media player classic in Letters #68 I thought I'd shoot you a line to tell you that Media Player Classic (6.4.09.1128) is already installed on all Windows XP systems. It's called mplayer2.exe; you can just RUN it, and/or you can set it as default on most of the media types.
MPLAYER2.EXE is the old pre-skins pre-plugins version of Windows Media Player, and it's a perfectly handy program in its own right, but Media Player Classic took the idea and ran with it. It's got playlists, QuickTime and RealVideo support (provided you've already installed QuickTime and RealPlayer, or got the codecs some other way - MPC uses their codecs without sticking you with their lousy interfaces), DVD playback, far better keyboard shortcuts, better zoom and aspect ratio control - and the list goes on.
I can't seem to be able to open your site with Mozilla Firebird. Up until two months ago or so it was working fine, so maybe you can give it a moment or two and fix it!
When I tried the latest Windows build (v0.7+) myself, the first time I tried to load my index page it mis-rendered - everything's there, but the main text is in a narrow column on the right-hand side. If I press F5 to refresh, though, it gets the page right. This behaviour is repeatable.
This suggests to me that it's the Mozilla devteam that needs to fix a bug, not me.
(A reader's now told me that turning off the "pipelined HTTP" option in Mozilla's preferences solved the problem for him.)
Just wondering if you know of anywhere localish that sells the Swiss-Tech Utili-Key, preferably in New Zealand where I live, but Australia or anywhere else that will charge less than $US35 for shipping would be good.
They look very good, but I was wondering - in a world where nail files get broken off nail clippers to thwart the would-be toe nail cutter hijacking gang, who's stopping people taking their key-tool-equipped keys onto a plane?
I didn't know who sold these things down here in the Antipodes, but some readers clued me in; apparently Kathmandu has them, though they charge a big premium, even after you take exchange rates and shipping into account. If you don't buy at a half-price sale, you might as well get your gadgets from the USA.
Coincidentally, though, I just bought a Utili-Key, and a few other Swiss-Tech widgets, from this eBay dealer.
(Obviously, I was shopping for Marilyn Monroe purses and got distracted.)
Their prices are really good, but like many such dealers they soak you on the shipping and <air quotes>handling</air quotes>. The total price for five tools was only $US42.80, which is miles below retail, and the actual postage to Australia was another $US11, which was fine, but they pumped that up by another $US19 in "handling".
That reduced the quality of the deal from "remarkable" to merely "OK".
I doubt they'd sting you for $US30 for one tiny tool, but who knows.
In answer to your second question: Usually, nothing is stopping people taking the one-point-five-inch blunt-pointed Ultimate Death Blade of the Utili-Key on a plane, or indeed into other "secure" areas. The Utili-Key's quite popular as a Pocket-Knife Of Last Resort, and has reportedly gone unnoticed not only by minimum-wage airport security, but even the US Secret Service!
The FBI's wise to it, though; it's listed in their Guide to Concealable Weapons 2003 (in PDF format here)!
Is temperature the only factor that limits the output current of a transformer, say, one that drops 240 volts down to 12 or so? If so, then is any good reason not to try water cooling it, since the copper wire is all insulated?
Assuming you could stop the iron from rusting, and keep the water flowing around the copper wire pretty well, then could it be possible to draw several amps continuously from a one amp rated transformer?
By and large, at low frequency (the 50/60Hz of mains power counts as low frequency), heat is the limiting factor. The impedance of the transformer also plays a role - the more current, the more voltage drop - but heat is the biggie for low frequency.
Retrofitting water cooling to an existing transformer is easier said than done, because the wire you can see on the outside is far from being all of the wire, and heat from the inner windings has to make it to the outside.
You certainly can liquid-cool transformers, though, by simple immersion in a suitable fluid or with some kind of heat exchanger coils. The big "pole pig" transformers on power poles, for instance, are oil-cooled (and the oil, famously, often contains polychlorinated biphenyls, which cause some waste disposal adventures).
Here's an informative page about all kinds of coils and transformers.
I have an opportunity to purchase a Kenwood LVD-K590V laserdisc player with a bunch of discs for about AU$50.
I was just wondering, what is your opinion on laserdisc?
It's the format of the future!
But seriously, folks - it's fine. If you don't have a double-sided player then you have to flip and/or swap discs, just like VCD/SVCD users today (I don't know the K590V's specs, but I presume it's double-sided), and laserdisc quality isn't up there with DVD. But it beats the heck out of VHS!
You may find the Laserdisc FAQ interesting.
The boss swears by it, but I've rarely found it to be at all useful. It certainly does run a large battery of tests, but very few of them are actually needed. The main ones that are, are the CPU, memory and HDD tests. They seem to do a pretty comprehensive job, but there's been a number of times when I've used it to test dying RAM and HDDs and it hasn't found a thing even though they're clearly cactus.
Where it might be useful is tracking down those bastard motherboard problems where you get completely random resets etc, but to date it hasn't for me.
My favourite part of it is that it runs off a boot floppy, and the main program only allows them to be created a fixed number of times. I don't have to tell you how sucktastic floppies are at holding their data. And no, you can't get around that.
Having spent enough years of my life trying to figure out why junior's P-200 can't boot, I'm now very much a professor of It's Boned, Get A New One University.
I'm a computer/ATM/printer/bank/general stuff technician. I notice you mentioned Micro-Scope diagnostic software in one of your letters.
I am one of the people who has been using this software, weekly, for several years now. it is a pretty invaluable tool when diagnosing faulty hardware. It's mainly used to test hard drives, system and video RAM, and motherboard components (keyboard controller, DMA controller, parallel port, etc).
You're right though, it won't get every fault. I've had sticks of RAM (usually Acer or HP crap) that test fine but cause all the Windows problems under the sun, yet mysteriously replacing the RAM with new stuff makes the system stable (well, as stable as can be expected from a Windows desktop).
And it won't reproduce, as you said, "the system load of playing a 3D game", but likely if that's a fault, it's going to be a driver issue or software (OS or game) patch issue. At least you can use Micro-Scope to see whether the video RAM is faulty and run a test on the hard drive. Maybe the game is creating temp files over a stuffed section of the drive. Every tool has its place!
It is quite expensive, but definitely worth it to all the businesses I have worked at who had used it. It comes on a floppy disk (yes, a 1.44Mb 3.5 inch floppy) and you can make 10 copies, using the special copy utility on the disk. I have tried every disk imaging program under the sun to make copies of the original and of the copies (as floppy disks are wont to die when they sit in my work case in the sun in my work vehicle while I am doing things elsewhere) but to no avail. If you run out of master copies, you have to buy 10 more licenses from micro2000.
As an ex-user of POST-Probe and Micro-Scope I'd just like to say that they did pay for themselves very quickly in saved time. One time, with a faulty batch of Gigabyte motherboards, POST-Probe correctly identified the fault as being the clock generator chip (a batch of 25 motherboards at that!). I only wish I'd managed to snaffle the kit out with me when I left the employ of that particular firm. It must have been at least six or seven years ago that I was using it, and it surprised me to see that Micro2000 is still going.
I'm a moderator for a couple of the forums at [popular Web site]. Recently I've noticed a couple of members selling ex-review hardware at quite inflated second hand prices. My concern is with the legality of selling the review units.
When you are given a sample for review, are you given full ownership of that sample, or does it remain the property of the original supplier (be that manufacturer or distributor)? If you gain full ownership of it, do you get a full, partial or no warranty?
I'm asking you, as I trust your judgement, and would like to protect [the site's] members from getting stung by these profiteering reviewers, and possibly winding up with bug-ridden early release hardware with no warranty and no comeback.
I Am Not A Lawyer, but as far as I know free review product is the reviewer's property. Review software generally carries a Not For Resale sticker, and review hardware might, but if they gave it to you (as opposed to just loaning it to you; most high-priced review gear is only loaned out, and retailers that provide things for review generally won't give away anything with much of a resale value), then it looks as if it's yours to me. Sell it, set fire to it, give it to a friend; whatever.
Some review product, especially loaner gear from major corporations and their PR companies, comes with elaborate review agreements which make the reviewer financially liable for the gear, should it be broken or lost or otherwise annihilated. Those agreements probably say something about selling it as well, if it's not loaner gear, though I've never seen a review agreement bundled in with giveaway gear.
In any case, I'm not really an authority on those documents, since I always just tear them up unread - I'm damned if I'm going to pay for a $10,000 video projector that just decided to die while I had it. If I drop it or stick a screwdriver through it, then the only price I am prepared to pay is that I never get any more review gear from that source.
So that's the legal side of it, to the best of my knowledge.
The moral side, if one may laughingly apply the term "moral" to the practice of journalism, is another matter.
Selling your review product, even if it's legally perfectly fine for you to do so, is generally accepted as a pretty low act - if only because it creates an obvious conflict of interest. Not to mention arguably taking a sale away from the people that provided the gear to you.
(Selling ex-review product at inflated prices is, of course, even lower!)
I've given various ex-review things to friends, and I've accepted compensation for a couple of them (Remember: It Doesn't Count As Selling It If Someone Just Gives You A Bottle Of Scotch), but by and large ex-review gear just piles up here, until I find a friend to foist it upon.
The provider may give the original reviewer warranty support, despite the fact that no contract exists (if you haven't paid for it, no contract exists between you and the seller, and they have none of the usual responsibilities - though I think you'd still have a case if the product burned your house down or ate your dog). The providers aren't obliged to do this, but product reviewers often get amazingly good customer service nonetheless, for the same reason that restaurant reviewers do.
If you buy ex-review gear from someone, though, the only person you have a contract with is the person you bought it from, and if you manage to get warranty support for it (which you may, if the item has some kind of no-docket-needed manufacturer's warranty), then it's only because they haven't bothered to create a "review product" category in their serial-number database.
If they have created such a category, or otherwise manage to identify the widget you return to them as being ex-review by some particular person, then they may give you warranty service out of the goodness of their hearts, but put a black mark against the reviewer's name...
As regards defects in early release hardware - that's possible too, but not much of a problem with computer hardware, in my experience. Yes, now and then there's a mobo whose version 1.0 has some problem a BIOS patch can't solve, or some other analogous genuine manufacturing defect. But most of these sorts of issues turn out to be fixable with software or firmware updates, and many of the rest can be worked around, and most of the remainder just never get fixed - which could be grounds for a warranty return for refund, of course.
Genuine pre-release alpha software is common enough, but unpatchable alpha hardware isn't, unless it's a special non-retail pre-release engineering sample, which the manufacturer is likely to have taken special care over anyway. Manufacturers don't normally let hardware that ain't fully cooked yet out of the building.
When people buy things second-hand, they need to expect not to get warranty support from anybody but the dealer they actually bought it from. That dealer generally won't offer anything much beyond the usual goods-must-be-as-described sorts of common-law things.
The auction sites are full of doofuses who don't seem to realise this and bid stuff up to 95% of the retail price and beyond, of course - but whaddayagonnado.
I noticed in the papers recently that Media World Communications is back with its "Adams Platform" thing.
I can't believe you haven't had a look at this before, but I couldn't find an article on your site. The closest I saw was your general piece on video compression from a while back.
The Media World Communications site has a bit of a description on the Adams Platform; there's been a piece in the Sydney Morning Herald too. Basically, full motion video and sound over a 14.4 modem.
They could show us how, but then they would have to kill us, etc etc, so just trust me and hand over the cash without worrying about an independent test (I think I can see a reason why the technology has still not yet been commercialised since 1998, unless of course it doesn't work).
Actually, they say they have had it tested (I think by The Tolly Group, but no sign of the promised summary) and they seem to have got some cash (yet to be raised from the market) from a tech company that used to be in mining. Clearly they're well qualified to identify a good streaming technology.
Here, by the way, are four hints from a corporate perspective "that your claimed technology may be 'doubtful'":
One, a company with $340K in the bank signs a deal to pay you $81 million.
Two, only $11 million of that is cash, the rest in millions of shares "valued" at many times their current trading price.
Three, all of the $11 million is to be raised from the market (generally as part of a bigger capital raising, often involving the sale of shares currently owned by the people that signed the deal with you).
Four, the deal is conditional an all of the above, as well as approval of "shareholders" (the people that signed the deal with you and who want to exit the company with a bag of cash before the technology is proven to be a violation of the laws of mathematics).
Anyway, since Media World Communications are getting some press, and since it is all so secret, I wondered if you might be able to work out what this technology would have to do to work?
I am certainly no expert, but I believe a fair bit of thinking has gone into compression at MPEG or DivX level, and a lot of clever tricks are required to get it to where it is. While the mark of true invention is that the solution seems obvious, there are a couple of questions that should be able to be worked out.
First, by working backwards, how many bits of information can the modem handle per second? No matter what the data is, the modem can only send so much stuff. It should be possible to reverse engineer the maximum number of bytes of information that can be expected, and convert that into the number of unique numbers that can be sent per second.
Wouldn't that tell us how much compression will be needed? Presumably we get something like each photo (claimed 1000 to 1) will need to be transmitted in 5000 unique numbers. Divide by the number of frames per second required to have "DVD quality" (ignoring for a minute the benefit of video that doesn't change, like grass growing on a still day) and each different frame of video has to be sent in 250 units of time (whatever that calculates out to be).
So, broadly speaking, the job in front of them will be to get each picture down to x unique numbers and each frame video down to x/28 or whatever.
Next, making some wild assumptions about a compression algorithm, what would it have to do to achieve this? If I devised a big enough "lookup table" can I send full motion video over a 14.4?
At the simplest level, I guess if I have a copy of Casablanca on my hard drive all I have to stream down the 14.4k modem is the instruction to play it. If it is cut into 1000 sections, they send 1000 "play" instructions over the 1 1/2 hours - a 14.4 can handle that no problems. Maybe they have come up with the ultimate MMFPS game - every possible scene in the world stored as a lookup table and they just send the codes to show it. As long as I watch a movie that has scenes for which I have a tileset, and if I don't mind an animated Sean Connery, I will be fine.
Ah, The Adams Platform. "Failing To Impress People Who Know Anything About Information Theory For Over Five Years!"
There's more information about this amazing, but rather slow-to-market, invention on adamsplatform.com.au. All these years, and they still haven't been able to just patent it.
(It's now 2008, another five years since this page first went up; I'm giving this page a little update to fix broken links and such. The all-conquering Adams Platform has still strangely not managed to make it to market, but there are strange things on the Web site now.)
Compressing video is like compressing luggage. Yes, you can fit more clothes into a suitcase without damaging them by being cleverer about how you pack them, and yes, you can squish your clothes harder and harder to fit them in whether or not you've packed them cleverly, but beyond a certain point you cannot compress further without losing quality in the de-compressed result.
Lossy compression is, of course, meant to throw away data (whereas packing is not meant to throw away clothes...), but the total amount of data in a given image, sequence of images, tune, text file, whatever, can be mathematically expressed. You absolutely cannot compress the data, by any means (wavelet, fractal, whatever), beyond that point. If you end up with a smaller file, then you've thrown away more data.
Really clever data-throwing-away schemes that can reduce the information content of the signal without affecting its appearance to the viewer/listener/whatever are where compression's at the moment. Different kinds of data have different kinds of redundancy, to which lossy compression can be tuned.
Here's more fun.
And then there was Pixelon, of course...
Oh, just go here.
Your string-of-play-commands idea is, indeed, how many of these schemes work; the usual magic-compression scam-artist technique isn't quite that advanced, though, as it usually amounts to "press play on the hidden VCR".
The "giant lookup table" concept, though, can't work, because of the utterly, horrifyingly, brain-detonatingly mega-giantness which such a table would have to have. Douglas Adams ain't in the race, here.
As I mentioned in my old digital video piece, there are 16.8 million to the power of 307,200 possible images that can be displayed on a 640-by-480, 24-bit screen.
Even if you figured out a way to store a whole video frame in one atom, you'd need an uncountably vast number of universes worth of atoms to store them all.