Dan's Data letters #144
Publication date: 11 April 2005.Last modified 03-Dec-2011.
Better than my ThinkPad 600X, that's for sure
I gave your Inspiron 9200 review a good read, and I was wondering how you felt it ran in the gaming department?
I'm really struggling to find a decent review of how well the 9200 can run, say, HL2, Doom 3 and Far Cry.
Here are 4 prefab answers for you:
1) Sweet! Nice FPS. - high detail.
2) Okay - Medium settings - average FPS.
3) Just - Low Settings, crap FPS, reminds me of my 486.
4) You're freaking kidding, right?Joel
Answer:
This isn't entirely relevant any more, since the 9200's now been replaced by the similar-but-different 9300 (US
page, Australian
page). The major difference is that you now have the choice of a 128Mb
Radeon X300 video adaptor, or a 256Mb
GeForce Go 6800 (for significantly more money). The X300 will
perform much like the 9200's Radeon 9700 (OK, maybe more like a 9600, but it's not a big difference. Note that the
cheap-ish base model only has 64Mb of video RAM, though; any gamer's likely to want 128Mb). The Go 6800 performs like
a real 6800.
They've also switched to DDR2 memory (a bit more expensive, but not exciting), and Intel's three-letter-designated Pentium Ms (the cheapest P-M 730 is the same 1.6GHz chip I got in my 9200). The price is much the same on a bang-per-buck basis - though as I write this the minimum RAM and hard drive specs have been raised, along with the price. And, as usual, special offers and discounts commonly apply.
Oh, and the 9300's nominally a WinXP Media Center machine, but there's still no integrated TV tuner, a choice on Dell's part which has attracted a certain amount of opprobrium.
Regarding your prefab answers: If you get a 9300 with a 128Mb X300, you're talking 2, or maybe 2.5. If you get the WUXGA screen, the X300 will have a very hard time pushing all of those pixels in 3D mode.
With the Go 6800, you're talking 1. Again, though, the 1920 by 1200 screen will give the adapter a workout.
As with all modern LCDs, though, the 9200/9300 screens handle lower resolutions quite smoothly, so you could just run lower res with no FSAA (because the resolution scaling will smooth stuff out for you anyway), and you should be fine.
Hey, why not cesium?
I read your rant on short-lived Li-ion batteries. As a cardiologist, I deal with pacemakers with some sort of Li batteries that last 12-15 years, even longer.
What's their technology?
Robin
Answer:
They're not lithium ion; they're just non-rechargeable lithium batteries. Maybe lithium-manganese dioxide batteries,
with a liquid electrolyte and solid manganese dioxide cathode; that's the most common lithium battery type, but there
are a couple of others (lithium iron disulfide is what's used in lithium AA {and now also AAA} cells; lithium thionyl
chloride is used in ultra-long-life low drain backup batteries; that'd be my second guess at the pacemaker battery
chemistry).
Anyway, ordinary lithium batteries as used in cameras have a quoted shelf life of ten years; it's not too hard to stretch that out a few years more for an expensive implantable battery, and the current draw of a pacemaker is low enough that the batteries probably last for pretty much their whole shelf life.
Lithium is a highly reactive metal that does not play well with others; it ended up in these batteries because it's very light, and has very high standard potential - much higher than more manageable metals like iron, nickel and cadmium. This means batteries using it have very high energy density - lots of energy per gram.
There's not a lot of point making rechargeable batteries for pacemakers unless you've got a wireless charging system (possible, using an inductive arrangement, but not to my knowledge ever actually implemented). If you have to plug the battery in to charge it, you might as well just swap it for a new one while you've got the patient open.
Implantable pacemakers were developed a while before the lithium battery was, though; some early models therefore used plutonium batteries, kin to the radioisotope thermoelectric generators that power some space probes. Plutonium makes lithium look like a fuzzy little pussycat.
Won't you guide my sleigh tonight?
I'm interested by a new device aiming to cure nasal congestion by phototherapy; see here and here.
On the second site the characteristics of the LEDs are given - red light with 660nm wavelength, intensity of 6 mW by diode; aren't these standard high luminosity LEDs, meaning that such a device can be build for a few dollars?
JP
Answer:
Yes, they are. Not all "red" LEDs are 660nm, but 660nm high intensity LEDs are common, and cheap. You can get them
from any electronics store; add a battery box, a series resistor and a switch and you've got basically the same thing
as the "Bionase", for maybe a tenth of the price.
Of course, there's no point investing even that much money and time in the thing if it doesn't bloody work.
There is, so far as I can see, exactly one clinical study supporting the reality of this thing's effect. You can find it in full in PDF format here.
The study gives the probabilities of the various outcomes, which are generally very impressive. The whole study's standard for significance is the usual clinical one, P=0.05; an event with a probability of 0.05 is one that has a 5% probability of occurring by chance, rather than because of any actual significance of whatever you're doing. The results the experimenters got, though, are generally much better than that - 0.004 for cough, for instance (which you wouldn't think would be affected much by anything you did to the nasal membranes, but there you go) and a stunning 0.0004 for rhinorrhea (known to normal humans as a runny nose).
There are two important things to remember here, though.
One: The study was of only 50 people. Not a tiny sample size, but not a huge one either. For some reason (maybe because they coin-tossed the treatment and placebo groups) only 21 of the patients received the actual treatment. This is enough to do real science, but it's not enough to be very sure of your end results, especially when you start seeing P=0.0004 results.
Two: To my knowledge, nobody's replicated this study since it was done in 1997. That's a killer. You can find this study referenced on zillions of Web sites selling the Bionase (none of which name the "major journal of immunology" whose "findings" allegedly say the thing works; it took me a moment to look it up...), but nobody anywhere seems to have been able to repeat the results. Replication is an essential part of the scientific process; the more replication there is, the less chance there is that the results can be ascribed to fraud or incompetence.
You can find one study that says that anything works, up to and including squirting coffee up your bottom to cure cancer. Replication is key, and if nobody else can duplicate someone's results (cold fusion, anyone?), it's best not to waste your time on the idea.
Especially when it's an idea that has a long and undistinguished history. This is the case for "phototherapy", which has been around since the first quack put some coloured glass in a sunbeam and charged people to sit in the light. It got to be quite a big business once electric lights were invented.
Today there are a few tightly defined phototherapy-like treatments that're actually supported by evidence. Ultraviolet therapy for some skin complaints, for instance (generally in conjunction with drug therapy, though if you're a jaundiced baby, I think they still basically just put you in a tanning bed), and "light boxes" to help people with Seasonal Affective Disorder. That's about it, though; umpteen people are shining LEDs and diode lasers on people these days to treat this, that and the other, but none of that stuff has any value beyond placebo, as far as I know.
Option B: Night-vision headset
I've got a big arse 4Ah, 6V, 1kg SLA battery which powers my bike light (10W globe and 2.4W globe). I use the light for night mountain bike riding. My issue is that when I'm racing, I'd like to carry as little weight as possible. Can I replace the SLA battery with 5 1.2V 2400mah NiMH AAs? How would I go about it (I've never made a "home made" battery before)? Would I get significantly more run time too, seeing as the AAs have a total of 12Ah?
Some Guy
Answer:
Yes, the little cells would work. NiMH AAs can handle delivering that much current reasonably cheerfully.
The 2.4 watt lamp, if it's accurately specified, is drawing only 0.4A; plenty of AA-powered flashlights draw that much, so you could use four AA alkalines (nominal 1.5 volts each) and get at least a couple of hours of decent brightness. The 10 watt lamp is drawing about 1.7 amps, though; D-size alkalines or a six volt alkaline lantern battery could just about handle that (lantern batteries have four cylindrical "F" sized cells inside them), but smaller non-rechargeables wouldn't be happy.
NiMH cells, though, should be fine for a couple of amps.
The current's not enormous, so to assemble the pack you could take the easy way out and just put the cells in a standard plastic battery holder from any electronics store. Alternatively, buy "tabbed" cells, which have pre-attached solder tabs on them, and you'll be able to easily solder up your own pack (without scuffing up the ends of the cells to get solder to stick, and running the risk of baking them if you heat them for too long). That pack won't, of course, be easy to disassemble if you want to charge the cells in a normal separate-cell-bays charger.
You wouldn't get more run time from the AAs, though, because you only get more capacity if you wire the cells in parallel - in which case you don't add their voltages.
Five 1.2V 2.4Ah AAs in series gives a 6V 2.4Ah battery; less than an hour and a half of run time at 1.7 amps. Five 1.2V 2.4Ah AAs in parallel gives a 1.2V 12Ah battery, which'll run your light for ages, but only very very dimly.
If you need more run time, you might like to look into higher capacity NiMH cells. Five 5Ah D cells in series should weigh quite a bit less than your SLA battery, and give longer run time - but they'll cost a lot more than the SLA did, too.
Moving parts reduction
In I/O 43 you have this picture. It looks like a Flash based 44 pin IDE drive. Do you know where you can get one of those in Australia (that doesn't cost the earth)?
You linked to an article that mentions VME Systems' Compact Flash to IDE adaptors, but they don't seem to have them on their website anymore.
Chris
Answer:
That picture's actually of a CompactFlash card with its casing off. That is, of course, a basic ATA drive,
if you can only plug it in.
Similar-to-identical things aren't too hard to find locally.
They're basically just pin adapters with a power input, so in a pinch a dedicated hacker could make one with a couple of connectors and hook-up wire.
Next stop, kittens
After reading your reply about the magnetic field required to move coins, I was wondering if you were aware of the demonstrations of magnetic levitation of non-magnetic objects. Most famously, the levitation of the frog.
Apparently, if you can make a field of about 16T, you can move stuff regardless of its lack of inherent magnetic properties.
Daniel
Answer:
The levitating frog and water-drop float because of
diamagnetism, which can be demonstrated domestically
with a few water tricks.
You could indeed push around a coin made from an appropriate material (copper would work) with a multi-Tesla magnet, but so far the only such magnets we have are electromagnets that're at least water cooled, and often superconducting and cooled with liquid nitrogen (MRI machines, for instance). You could maybe make a small low-duty-cycle model that you could carry, running perhaps from a capacitor bank or a bunch of NiCd cells you didn't like very much, but you'd have a bugger of a time making the switch not go "bang" when you activated the thing, thereby alerting your audience.
If we ever manage to make 16 Tesla permanent magnets, people who try to do magic tricks with them will be identifiable by the fact that they become firmly attached by their sleeve to the first refrigerator they walk past.
A modest proposal
I would like to suggest if you could make your RSS feed full text.
I understand that you do receive some revenue from click-throughs on your pages, but you could also include those ads in a full text feed.
It is just much more easier to read your posts through aggregators than through a browser.
Geito
Answer:
Dude - my site shifts more than a hundred gigabytes of traffic per month already. If I turned my little 4k dansdata.xml
file into something that contained just the single last article - say, 35k of HTML plus far more in pics, less whatever
pic-caching people's aggregators do - I shudder to think what'd happen to the server load.
Full-text RSS can work OK for blogs with regular small posts with zero to one pictures per post; traffic can still be surprisingly high (thanks to people who leave check-every-hour aggregators running all day), but most blogs of course don't have that much traffic.
Dansdata.com got more than 1.3 million pageloads in March; more than 370,000 of those pageloads were dansdata.xml. Do you see the problem?
(UPDATE: Traffic for October 2007 was 1,685,768 pages served. 705,831 of those were dansdata.xml. When I quote my traffic to people I always make sure to not include the RSS feed, on account of its amazing popularity. And yes, I'd be happy to add ads to the RSS feed and have someone else serve it, but no suitable candidate has come along yet.)
Sliding currency, redux
Regarding your letters #143 - I vaguely remember learning that iron, nickel, and cobalt were all magnetizable in the basic sense. I've never magnetized a nickel-denomination coin, but I figured it was due to the latter day lack of its eponymic element.
And during WWII, US pennies were made with steel. Apparently the few copper pennies in a particular year are rather valuable.
Cole
Answer:
It's true that nickel and cobalt are magnetic, but coins containing those metals aren't, necessarily. The most common
stainless steels are made out of iron, nickel
and chromium; two out of three of those metals are magnetic, but the alloy isn't.
Ah, memories
The letter on wine cooler reminded me of one of my childhood experiences: spilling an entire glass of orange juice into the keyboard - and thus the heart - of my Amstrad CPC 6128.
After the initial panic, and the furtive glances around to ensure that Mother Dearest was not aware of her 8 year old son attempting to destroy her $4K computer, I took the cursed thing outside and shook it upside down, ooh, for about ten minutes straight. Once it had stopped dripping and the worrying rattling sounds had ceased, I took it back inside and left it alone for a day. Turned it on the next day with great fear and trepidation, and was not consoled at all by the disturbingly treacle-like action of the on-off switch, nor the pungent citrus aroma that wafted out from the keys.
The screen printed all sorts of garbage at first (Basic 1.yurgle, it seems), and beeped alarmingly. Power off, power on. Repeat. Seven or eight cycles later, the old girl was back to normal, albeit a definition of normal expanded to include EVERY DAMN KEY sticking down when pushed, a habit that was to last for ~4 years in the case of the lesser-used keys.
That computer continued to be used more or less daily until I was 19, when - get this - the MONITOR gave out, after having been used as my Amiga viewer-of-choice for the past five years.
Who says the British don't make decent 'puters?
Damian
Answer:
This one time, the office I was working in sprung a roof leak
directly above an Amiga 3000, which got well and truly waterlogged and started yellow-screening (the not-often-seen
"early bus error" indicator).
So I stripped the machine completely (not really difficult with a 3000, just annoying, 'cos everything's on top of everything else), and put all of the damp bits under an X-hundred-watt halogen flood light to dry.
Only after it'd been baking for, oh, maybe a day, did I notice that there was an EPROM on the board whose light-protecting sticker had floated off in the water, and which had therefore had its erase window facing the floodlight for all of that time. I don't think that light still had its glass UV-stopper in place. I fully expected the computer to, therefore, be toast.
It worked perfectly. The bugger probably still works today.
