Dan's Data letters #80Publication date: 13 December 2003.
Last modified 03-Dec-2011.
The MHz maze
I'm about to build a new PC, which I haven't done for a while now (my last PC was a 300 MHz K6-2, with the very amazing Voodoo 2!), so I'm reading up on AMD machines, trying to work out this darn new-fangled technology ;).
I've searched for "front side bus" and "DDR" on your site, and I think I've got it mostly worked out now, but it'd help if I could run a few questions by you.
I'm looking at getting an XP 2200+ or so (it doesn't need to be super fast, and the prices for this CPU seem very reasonable), and am looking at the ASRock K7S8X motherboard. The CPU needs a 266 MHz FSB, and the ASRock claims to be able to do 333 MHz, so I figured this leaves me some room to upgrade in the future. But the manual for the motherboard only lists jumper configurations up to 200MHz FSB. What's up with that? I'm guessing that it's something to do with clock doubling?
Also, RAM. The price difference between PC266 DDR and PC400 DDR isn't much, but will the RAM also run at the front side bus speed, meaning that I don't actually need RAM that's faster than 266MHz?
Or is there any advantage to running the RAM faster? The machine's going to be a file server, so lots of DMA happening between disks and network here...
All (pre-Athlon 64) Athlon and Duron CPUs have a CPU-to-chipset bus that runs at twice the nominal FSB speed. This gives rise to all sorts of confusion about whether their FSB is actually 100 or 200, or 133 or 266MHz, or 166 or 333MHz, or 200 or 400MHz. Marketing people like to quote the faster speed; CMOS setup programs and manuals always list the slower one. Now that the doubled speed of the first Athlons and Durons is the same as the pre-doubling speed of the fastest Athlon XPs, it's even more confusing.
Yes, your motherboard can do a 400MHz CPU bus speed; you'll get it if you set the FSB to 200MHz.
By default, the RAM will run at the FSB speed, but most motherboards let you run it faster. There's little performance gain to be had by doing that, though.
You could still drop the few extra bucks for the faster memory, though, so that it'll be more useful in the future.
I read your review of the Xitel INport. I was wondering whether this device would work well with a microphone input (or two)?
My girlfriend likes to digitally record voice and instruments via a microphone. She has a nice notebook, except mic recordings through the book's line-in have rather unacceptable noise. The noise persists with different microphones and different recording software. So I thought that the INport might be a good solution to both her recording problem and my Christmas present problem.
An INport would only be useful with a mic that has line level output. Normal "mic-level" mics will be no good with the INport, unless you add a pre-amplifier to boost the microphone signal. This is perfectly doable as well, of course. Pro-grade mic pre-amps start from around $AU300, but something as cheap as an old stereo pre-amp with a mic input will do, and there are short form kits from about $AU20 ("short form" means there's no case or power supply included...), if you know which end of a soldering iron to hold.
Back from the dead
I regularly deal with ex-rental laptops which have rather dead lithium ion battery packs.
What I'm interested in is this page. It has some useful information regarding the "resurrection" of Sony InfoLithium batteries.
Would this product be helpful? Since most of the laptops don't seem to want to charge the semi-dead Li-Ion packs, I should think a standalone charger would be useful. Also, if the "Deep Conditioner" cycle works with Li-Ion batteries (it only specifically says it works with NiCds) it would be easer, and more effective.
Yes, it does look as if some "dead" lithium ion batteries can be resuscitated, especially if you hack up a discharge resistor rig so you don't have to try to flatten them with some "smart" appliance which may refuse to operate because it believes the battery to already be flat.
Resistor-discharging any multi-cell battery, though, is something that you shouldn't do in a set-and-forget way; you have to come back and remove the resistor well before the battery makes it to zero volts, if you want to avoid "reversing" the weaker cells and doing more harm than good.
It's also worth mentioning that most "dead" batteries of any kind really are worn out, and there's nothing you can do for them. Lithium ion batteries don't last for as many cycles as NiCd or NiMH, and laptop battery chargers are often depressingly dumb, and may beat their batteries to death needlessly quickly.
The charger/discharger you linked to is made for NiCd and NiMH packs. If you managed to connect a lithium ion pack to it, you might well end up dealing with a fire.
Jaycar, like many other electronics stores, sells plenty of "battery conditioning" chargers, including various models that automatically discharge before charging, every time. Generally speaking, all of these things are bad products. Battery discharging is practically never necessary for normal battery applications (super-high-current R/C racing aside).
And a clunking noise from the diff...
I have this not-quite-so old 30Gb Maxtor hard drive that has served me well since I bought it, and still to this day. A couple of weeks ago though, it started making this high pitched and kinda loud whining noise (the kind of noise they could use for an alarm clock). Its speed and data integrity haven't been affected yet as far as I know, but this PC being on 24/7 with it's side panel off (crappy circulation + stock AMD HSF + XP2400+ = H34T) makes it audible pretty much everywhere in the apartment.
I'm guessing this noise comes from somewhere in the spinning mechanism (it's also not one of those Maxtor hard drives with the liquid bearing feature) but I couldn't know as well as you! Where do you think this noise comes from, and can I make it stop without unplugging it or replacing it (it's my boot drive, but WinXP is on a separate drive)?
Sorry; it's time for a new drive.
A whine, in the absence of any other drive problems, probably indicates a reasonably benign problem - probably a dying bearing. "Reasonably benign", in this case, means that the drive can be expected to survive for a few weeks. But a dying bearing can go bad enough to finish the drive off at any time, so you'll be wanting to upgrade to a new drive soon.
There's a little collection of Scary Drive Noises that can be found in various places on the Web, by the way; here's a nice page with MP3 versions:
These noises are all great alarm clock sounds for geeks!
Is it possible to use the Windows command prompt "net send" command to send a message to an IP address?
Yes, provided the target machine is running the Messenger service.
Many people turn off or firewall Messenger nowadays, though, because spammers discovered it a while ago now, and started sending pop-up-window messages about all the usual spammer crap.
Better than Sea Monkeys
Perhaps you have read the recent Scientific American Top Sci/Tech Gifts 2003 article. It has some pretty nifty gadgets and toys in there. One in particular caught my eye, which is the Small 4 Inch EcoSphere. As soon as I saw this I knew that I wanted it, and knew exactly where in my office it was going to go. However, my excitement was short lived when I found that the company selling them won't ship overseas (I happen to be in Melbourne). I decided that it was probably not the best idea to have one shipped from overseas anyway, as apparently they should not be shaken, shouldn't go without sunlight for too long (something like 60 hours), and they also appear to be rather fragile.
Several Google searches and queries to IRC folk followed, but nothing turned up an Australian supplier. Do you know where they can be found in Australia?
H. J. Farnsworth [probably not his real name]
Ah, yes - the Cruel God Toy.
The Eco-Sphere home page is here, by the way.
Sealed terrariums may be easier to find, but they usually don't contain any macroscopic animal life.
Satisfying gaming experience #2673
I have a Radeon 9700 Pro left over from a previous system, so it's my temporary vid card. The power supply is a Thermaltake TruBlue 480 watt. The mobo's an MSI K7N2 Delta. RAM is Crucial 512 PC2100 unbuffered, cause that's the only RAM the nForce Ultra 400 chipset will take.
I installed the latest Catalyst drivers from the ATI website. While playing any game demanding even a moderate amount of power, the game kicks me out and sends me an error message saying my VPU has reset all of my graphics controls.
So I formatted again, and tried to backdate drivers, but had the same problem. I tried a different ATI card, an 8500, and things went smoothly. The 9700 works fine on any other computer, but it hates this one.
The peculiar form of the errors you're seeing is thanks to the VPU Recover feature, which was new in the v3.8 Catalyst drivers. Assuming VPU Recover or some other driver feature isn't what's causing the error (which it could be), then it's probably saving you from seeing some nastier crash - possibly a full system hang.
Unfortunately, there are lots of things that can cause this sort of problem. Since the video card seems fine in another machine, and a swapping it for a lower power video card fixes the problem, my first guess would be that it's a power supply issue. But that is only a guess.
Suspiciously cheap toys
When I did an eBay search, though, I found cars that look very similar to the genuine Kyosho Mini-Z but only cost $US19.95. They couldn't possibly be the same quality - but is it worth spending the extra $US70?
Yes, those Hei Pao cars are knock-offs.
I've never seen one of them though. If the dealer doesn't soak you on shipping and handling, they may turn out to be decent value, but you shouldn't assume they'll be compatible with any genuine Mini-Z parts.
I'm writing in reply to the "Cheap DSLR Alternative" in Letters 73. I'd like to try to convince you that buying a film scanner is not as bad an option as you make it out to be. :-)
A film scanner is easier to make than a digital camera, and I think a film scanner costing $100 (used) can create images which compare favorably to a $1000 DSLR.
I'm in the USA and all amounts are US dollars.
Image Size: A low-end film scanner will have a resolution of around 2700 or 2800 pixels per inch, creating images around 3900x2600 from a "35mm" (actually closer to 36x24) negative. The Canon EOS 300D produces images that are 3072x2048 pixels.
Advantage: Cheap Film Scanner
Color Data: The film scanner will record an RGB triple for each pixel, producing a "true" 3900x2600 image. The digital camera will use a Bayer grid, in which each pixel records only ONE color - in every set of four pixels, there are two green, one red, and one blue. Thus, the 3072x2048 image which is produced already has interpolation in it.
Advantage: Cheap Film Scanner
Bit Depth: A low-end film scanner will probably have 12 bits of data per pixel per color. The Canon EOS 300D also has a twelve-bit ADC.
Upgradeability: If you buy a better film scanner later, you can re-scan your best pictures and take advantage of the higher resolution and/or better color. With a DSLR, buying a new camera won't let you get anything new out of your existing pictures.
Advantage: Cheap Film Scanner
Labor: With a film scanner, you have to develop (or send out for developing) your film, and then get it back, and scan it. For half of the images, the automatic color balance will probably fail, and you'll have to intervene manually to get reasonable results. You will need to carefully blow the dust off your negatives prior to scanning, and then digitally remove dust spots which remain in spite of this. With the DSLR, you don't have to do any of this.
Now, let's look at cost. I pay $5 per roll to have my film developed at a professional lab (with no prints). Add $2 per roll to buy the film and the total cost is around $7 per roll. If you want to save money, however, you can go to a "minilab". I have a friend who does this. After some initial confusion (because they couldn't remember anyone ever asking them to develop only without prints before), he got them to develop each roll for about $2, for a total cost of about $4 per roll. Cheap labs will put more scratches on your picture, but this won't usually reduce final quality - just increase labor by making you spend more time removing the scratches with the rubber stamp tool.
Thus, you can take pictures with a film SLR and a cheap film scanner which are not only as good as a DSLR, but better. You pay an incremental cost as you go, but it's small. At $4 per roll, how many rolls would it take to add up to the cost of the DSLR? If we assume $800 difference between the film scanner and the DSLR, that's 200 rolls of film, or 7200 pictures. I don't know how many pictures you take, but that would be about five years for me... and in five years, you'll probably be able to buy today's DSLRs on eBay for $150.
The real cost, however, is in labor. It takes me about 3 hours to scan one roll of 36 exposures. If you're being paid for your time, the DSLR justifies itself an order of magnitude faster than it would based on film costs alone. If you're a hobbyist, however, you're often willing to put in time for free in order to save costs in real money, and this is the situation where the film scanner is ideal.
If you want to look at some of my actual pictures, to make up your own mind whether I know something or I'm just talking out my arse, you can see some on the web here.
Anyway, that was quite the rant, and if you got to the end and are seeing this, thanks for reading it and considering it. :)
You make a good case, but you've left out one of the major features of digital cameras - immediate feedback. The ability to see what you've shot on the camera screen right away and on the computer screen pretty much as soon as you can get back to your desktop or laptop is not just a convenience feature for people who do a lot of boring product photography (like me), but also makes it more likely that you'll get the shot you want, and makes it much easier to take really great pictures. Well, provided you don't spend too much time "chimping" the screen of your camera and thereby fail to notice photo opportunities passing by.
The ability to take a big burst of pictures without pausing to change film, or caring that most of the shots will be forgettable, is another large advantage. Even if the subject's completely frivolous - my cat doing something cute, for instance - I can go into full paparazzo mode and shoot 200 exposures in five minutes if I like, then prune 'em down in five more minutes on the PC to the three that were worth keeping. Total expense, ten minutes of my time, which I probably would have spent staring at the cat anyway.
Regarding film scanner versus DSLR resolution - you're right that even quite cheap scanners can extract more data from a 35mm negative than a six megapixel DSLR will capture, but it should be noted that a lot of images have considerably less real resolution than the camera sensor, or film, can theoretically capture. Focus, camera shake, lousy lenses, or just plain fuzzy subjects; I ramble on about this in my old D60 review here.
You're right about the per-pixel colour versus interpolated colour filter array issue as well, although there are a few filter patterns out there besides the straight RGBG Bayer pattern. Even if a DSLR user shoots in RAW mode and does the interpolation on the PC, they're still doing interpolation.
Unless they've got a Foveon X3-sensor camera, that is. There are now - don't be frightened - two of those on the market; not just the very weird Sigma SD9, but now also the almost-as-weird Sigma SD10. Those cameras aren't a real option for most humans, and I'm particularly irked by the fact that the SD10 is now advertised as a "ten million pixel" camera, despite the fact that its (mildly tweaked) sensor has the exact same 2268 by 1512 resolution as the SD9. Sigma are alleging that full RGB on each pixel means tripled resolution, when all it actually means is tripled chrominance resolution, on average. The SD9 could mix it with six megapixel conventional digicams, and the SD10's resolution is exactly the same.
Your point about being able to re-scan film with better hardware is valid. Digital photographers who shoot in RAW mode can take advantage of better RAW-to-RGB conversion software in the future, too, but it's not the same thing.
Note, however, that there isn't necessarily anything for higher resolution scanning to reveal in many pictures besides a more accurate view of the film grain.
At four US bucks a roll, it would indeed take a lot of rolls of film to add up to the price of a DSLR, but if you're shooting a whole lot of pictures then you can hit the crossover point surprisingly quickly. Digital photographers tend to be very trigger-happy, and film's wider dynamic range means you actually need to shoot fewer pictures. Still, though, I'd have a hard time using fewer than four rolls per review I write; your five years to break-even would be more like six months for me.
It should also be noted that lenses and other camera accessories aren't free; $US1000 is not at all a high price for a modern 35mm camera lens. The price of a DSLR body can vanish into the noise for professional or gear-happy amateur photographers.
"And now, Mr Bond..."
Have you seen these before, or do you at least have any good theories on why they would or wouldn't work?
Should it be categorized with those other peculiar electronic devices that do precisely nothing at all, which you seem very astute at identifying?
No, I hadn't seen the "ItchZapper" before. I'm relieved, though; the URL made me think it might have been something to do with the odious Hulda Clark, whose "Zapper" contraption is claimed to cure pretty much everything that ails you, especially if your problems don't actually physically exist. I was glad to find something a lot more benign.
The page you linked to talks about killing bacteria to instantly stop an itch, which is great, except that it's completely wrong. Insect bites may inject bacteria or other pathogens into you, but that's not why they itch. Ordinary bloodsucking-insect bites generally itch because you're allergic to the anticoagulant and other proteins the insect's injected into you.
The nearest thing to an official site for the ItchZapper I could find is astoundingly, itchzapper.com. The how-it-works page there tells you nothing about how it works, but does tell you you're supposed to apply it "immediately" after you've been bitten. One would hope that this doesn't mean you're meant to use the ItchZapper before a bite has actually started itching.
Anyway, the "The Science Behind..." page there is more useful. It says the ItchZapper heats your skin enough to "activate" its pain response, but not long enough to actually burn you. I can believe that lightly scalding the skin on top of a bite could numb the area, or at least turn an itch into an ache.
A reader's now told me that you only need to heat mosquito spit to 54 degrees Celsius to thermolyse it into a non-irritant state (and that you can do the same trick with a lit cigarette, if you're careful...). Briefly heating your flesh to that temperature won't do it any harm, so the ItchZapper might work that way. You'd think one of the sites selling the darn things would say as much, though.
There doesn't seem to be any unbiased information about this gadget on the Web. There's a near-useless TV transcript here, but that's it.
At $US20, the ItchZapper approaches the status of an impulse buy (provided you don't have to order it from overseas, of course), but given that there are a few pretty effective chemical itch/sting neutralisers out there already, I don't see a big future for the electric version.