Dan's Data letters #65
Publication date: 12 October 2003.Last modified 21-Jul-2012.
Goop wars
Your article about the difference between thermal compounds came up in a forum I frequent, and to our surprise a representative claiming to be from Arctic Silver responded, pretty much saying your testing is complete bollocks.
Here's the thread.
So... any response?
James
Answer:
Gee, wouldn't it have been nice of Nevin (assuming it's actually him posting
there) to tell me all this when I notified him of the review, well
over a year ago? Or, well, ever?
While he's at it, he can take the issue up with LostCircuits, who would like to talk to him about cod liver oil.
Arctic Silver sent me some grease, and I reviewed it, and apparently got it all completely wrong. Who knew, since they couldn't find time in their day to talk to me about my apparently incompetent testing of it, and went on to send me some Ceramique to play with later. You'd think they'd stop sending review product to such a lousy reviewer, but nope.
Oddly enough, I similarly don't have time to address Nevin's backhanded complaints about my review in detail, largely because they're not very coherent; I don't know where he studied the "Science 101" he mentions, but it doesn't work that way on my side of the world.
The basic point he makes, though, is that the more heat you're trying to move and the smaller the contact patch is, the larger will be the influence of the thermal interface material.
I've tested grease with small contact patch simulators in the middle of my CPU-imitating heater plate (the little patch plates are fiddly and annoying, which is why I don't usually bother using them), and so far as I can see, there's bugger-all difference then, too. CPU cooling doesn't deal with above-boiling temperatures or super-hard clamping force, so it's not actually a tremendously critical application for thermal compound.
The meaningful differences between compounds, for CPU cooling purposes, seem to largely be ease of application and durability. If you can't get a good thin consistent layer onto the CPU, it doesn't matter how good the goop is thermally; also, if it all leaks out the edge of the contact patch over six months, or turns into powder, it's no good.
Arctic Silver products have always been pretty durable and reasonably easy to apply, and I've said so.
Scanner rehabilitation
I have an aging but perfectly good Mustek 600 IIN Paragon 24 bit single pass flat bed scanner. It came with what appears to be a proprietary interface card (SCSI?) that used to sit in an ISA slot on my old Socket 7 motherboard. My new machine only has four PCI slots and one AGP, so I can't plug the card in anywhere.
The card in question is labeled AB306N I/F Card Rev A. The plug for the scanner appears to be different too, so I'm concerned that even if there are SCSI type adapter cards that could sit in a PCI slot, the plug may not fit.
I'm allergic to throwing otherwise good equipment out, and will only look at a new scanner if the cost/benefit of getting a new interface card isn't worth it.
Any thoughts?
Kerry
Answer:
I think you're out of luck. If
this page
and a couple of other references I hunted up are anything to go by, that
card is indeed highly proprietary - a mutant parallel-plus-power
device, not SCSI or anything else standard. I strongly doubt any PCI equivalent
exists.
So you could junk the old scanner (its resale value is likely to closely approach zero, these days) and get a new USB one, or you could buy or build an ISA-equipped machine to use as a scanner server. This isn't as daft as it sounds, if you've got the space for another PC; you could use it as a print server as well if you liked, and also as a quick and dirty backup box. A second hard drive in a different physical machine, connected via Ethernet, is considerably better than nothing as a backup medium. Since it's so convenient, you're likely to actually use it, for a start.
An old Pentium II or Slot 1 Celeron machine would be perfectly adequate for this task - running Win98, or even Win2000 or WinXP if you load it up with RAM. The latter choices would be better if you want the second machine to be on all the time.
The storage swindle goes to court
You rant a lot about the difference between advertised memory and actual memory on hard drives and other storage devices. What's your opinion of the current lawsuit against Apple, Dell, Gateway, and others?
Alex
Answer:
Since the real-versus-advertised small-print lie factor is getting worse
and worse as drives get bigger and bigger (5% when you're talking megabytes,
7% for gigabytes, 10% for terabytes...), a case can be made for dealing
with it sooner rather than later. I imagine the actual result will just
be bigger small print, if anything, plus the usual $5-and-a-cup-of-coffee
payout for class action members and buy-the-whole-Porsche-showroom fees
for the lawyers.
On the other hand, as drives get bigger and bigger and cheaper and cheaper per megabyte, the lie factor matters less and less. People do keep thinking of things to take up all the space with, but it's hard to keep pace; my whole backup of everything I've written since well before Dan's Data existed, including pictures, is only 5.7Gb.
If you're only being ripped off for $2.50 worth of drive space that you're not going to need for three years anyway, it's not a big deal. It matters more for smaller capacity, higher price per megabyte devices like MP3 players and digital camera memory cards, but the severity of the problem's fading there too.
At least now I've got a simple thing I can link to when I say "This "256Mb" device has 244Mb real capacity...", rather than explaining the swindle over and over!
Chargus interruptus
If one buys a brand new device which runs on NiCd rechargeable batteries (Motorola Talkabout T5720) and interrupts the initial charge for a couple of hours before the recommended 16 hour period has passed, will this adversely affect the batteries?
Greg
Answer:
No. Rechargeable batteries may or may not need a couple of charge cycles
before they deliver their full capacity, but there's nothing else special
about the first charge.
Phantom errors
I have a 60GB Maxtor on my IDE RAID controller. One day I noticed after a power failure that when my system did its HDD check before starting Win2K Pro (FAT32 BTW, it's a long story why I am not using NTFS), that the CHKDSK summary said I had bad sectors. 1.4Gb of bad sectors!
Well needless to say I was concerned for my drive, so I downloaded Maxtor's MaxBlast drive diag utility, and ran it. Lo and behold, no errors, no bad sectors! So I ran a few more checks. The Windows error-checking utility shows no errors also, but CHKDSK (run in a DOS window) still insists that I have a huge chunk of bad hard-drive. Unfortunately, CHKDSK in Win2K always wants you to use the Windows disk checking thing to repair problems.
I keep my computer in tip-top shape as far as maintenance goes; up-to-date drivers, BIOS revisions, etc. The drive runs like a top, and I have never lost any data, so should I just continue to ignore this?
Josh
Answer:
Sure, it could just be a harmless software glitch, but I'd still use it
as an excuse to upgrade.
While the drive's still behaving itself, get yourself a nice cheap 120Gb drive, hook it up too, and use the disk-cloning software of your choice to dump the contents of the one onto the other. Then demote the 60Gb drive to paperweight, doorstop and/or fridge-magnet-donor duty.
Maybe the thing'll keep working fine for years. But given how cheap commodity drives are these days, you might as well upgrade.
Cramming 'em in
I'm trying to figure out whether the Lian Li PC-76's four front intake fans can be swapped out for 80x80x38mm Deltas. This is true of the PC-68. I can't exactly tell from the PC-76 screenshots whether that case has the same amount of extra space that the PC-68 does.
Stephan
Answer:
I'm not sure, but I think it'd be OK. I know that you could easily
move the bottom drive cages back a few millimetres if you had to, and re-fix
them with whatever technology suited your degree of laziness:
1) Chewing gum
2) Double sided tape
3) New mounting holes (easy to drill, since it's only aluminium; you could
do it with a hand drill), and little screws and nuts
4) Nicely drilled and tapped holes
The colours, the colours...
I figured I should shoot you this bit of info on the Sony DSC-F828 you mentioned the other day. The new Sony will be using their new RGBE sensor. This means that the extra bit of green information will make for better color balance. So far Sony is the only company that produces this type of chip.
Simon
Answer:
Well, Sony say the RGBE sensor's a big deal, but it's just another
kind of
colour filter array pattern. There are several camera sensor filter
schemes out there, as well as other enhancements like Fuji's
SuperCCDs,
particularly the new SuperCCD SR with one high sensitivity cell and one
low sensitivity cell per subpixel, for extended dynamic range.
Every digicam company with a distinctive flavour of sensor can be counted on to yammer on about how great it is, but the proof is in the actual independent photographic tests, which usually find that the new innovation isn't nearly as impressive as the manufacturer promised.
The SuperCCD is a great example of this. It does produce higher resolution results than its sensor cell count suggests, but not anything like as much higher as Fuji claim. The SuperCCD SR also seems to be pretty much a dud, if its one appearance in a retail product so far is anything to go by.
Maybe the RGBE filter has distinct advantages, but it's not even the first four colour filter. CYGM (Cyan, Yellow, Green, Magenta) filter patterns have been around for ages.
Interface archaeology
I've read on multiple Web pages that the interface between a floppy drive and the motherboard is not IDE. I've only seen it called IDE by places that sell cables, and the pages I saw didn't give it a specific name either. What is it really?
Jordan
Answer:
The floppy interface is actually much older than ATA/IDE. It's called SA-400;
Shugart developed it in 1978. It was used by the first popular 5.25 inch
floppy drive, and it later mutated into the late and unlamented ST-506 hard
drive interface.
SA-400's stayed pretty much the same for 25 years. The pinout of modern pin-connector drives is different from the old edge-connector one, but the interface itself is basically unchanged.
Subatomic hair care
Recently on US TV we've had adverts for a special hair blow dryer type thingy that allegedly pumps ions into the air stream to condition your hair. The first time I saw the commercial for this I almost had a blood vessel in my forehead burst. I have almost gotten used to the ads for putting vitamins into your hair by applying them via shampoo. My feeling has always been that since hair is dead, I might as well coat my dear deceased grandmother with vitamins as well. The effect would be the same.
Would she be "fuller and more vibrant"? I'm thinking it couldn't hurt, right? Don't know if she'd look much better though.
Anyway. My assumption is that they charge more for these suckers, due to the ion thing. The corollary to the assumption (if there can be such a thing) is that airheads will spend more for the ion models.
You aren't James Randi, but could you debunk this whole concept? Or do I need one RIGHT NOW?
Rich
Answer:
The things are
all over the Web, too, aren't they? It's amazing that a primpy little
clotheshorse like me didn't know about them until this moment.
Since there's not a strong crossover between people who care about hair dryers and people who know about high voltage devices, it doesn't entirely surprise me that there's little to no analysis of what the heck these things can actually do on the Web.
Well, there's this, but I don't think anything that "breaks down the water molecules" (ionisation, electrolysis, it's all the same, ain't it?) is going to do your hair a great deal of good, unless gaseous hydrogen and oxygen are needed in there.
The fact that people selling them can't decide whether they work their wonders by shooting positive or negative ions at the user (most say negative, but this one says positive if you click a "more detail" link) does not inspire confidence.
Many of the claims made for ionic dryers have to be bollocks - "Ions surround the hair shaft, making it smoother, shinier and more manageable, taming frizziness and static electricity", eh? "The negative ions attach to the individual shafts of hair, to provide a protective coating that keeps moisture locked in, penetrating deep into your follicles", what what?
They all seem to be supposed to reduce "frizziness", but as you say, that's not nearly the end of the claims.
If you electrically charge your whole head, your hair will stand up, because the similarly charged hair strands will repel each other. This is the classic hand-on-a-Van-der-Graaf-generator look.
I can see how this might be useful for hair drying purposes, because as the hair dried it'd tend to separate, making it easier to dry the rest of it. You'd probably also end up with a fluffier (frizzier?) result, even after the charge dissipated - which it would, in quite short order, after you stopped delivering new charge.
I doubt these ioniser dryers do this, though; blow an ion wind over a head and you'll charge the hair more than the head itself, with the result that the hair will actually stick to the skull. With a strong enough ion wind, this effect has been described as feeling "like you've been dipped in vegetable oil".
Charging your head this way would, presumably, help to stick down freshly dried hair and stop it frizzing up. If you then spray on leave-in conditioner or something, the hair will dampen up a bit again and, probably, stay where it's put.
I am unconvinced of the value of the ionic dryers versus an ordinary dryer followed by a damp comb, though.
Bijou tankette
Got this photo from an article on the situation in Afghanistan on the Economist website.
I know the Germans can be funny people in the weird sense of the word, but just what the hell is that toy that poor soldier is driving?
Matthew
Answer:
It's a
Waffenträger Wiesel (Weasel Weapon Carrier), Mark 1, distinguishable
from the Mark 2 by its three sets of road wheels, versus four on the later
version.
(Info Auf Deutsch on the Mark 1 here, hilariously translated version here; info on the Mark 2 here, translated here, lots of pics here.)
Not exactly a match for any proper tank, but pretty bloody formidable in the supermarket car park.
Help request
Hi,
I have a xp tower pc. What is the best way to break it without opening it up. It is covered against accidental damage. Therefore I will get a new computer if it breaks.Your help will be highly appreciated.
Thanks
Alex Richardson
Answer:
Well gee, Alex Richardson, I'm not
sure. For some reason, Alex Richardson,
I always have trouble helping people like you, who decide that
this page means I must be running some kind of PC
Fraud Assistance Service to help people break their own computer in order
to get a new box, or cash, from retailers or insurance companies.
That's a victimless crime, of course, because there's no way that it could possibly make everybody else's hardware and insurance more expensive.
Fortunately, Alex Richardson, I'm sure that some of my readers will be happy to tell you what you can do with your computer.
(Yes, I've e-mailed Alex to make sure it's actually him sending the message, not some reply address spoofer. He replied and confirmed it, so maybe it's his mischievous kid or something, but what the heck. An e-mail address at somewhere like CATos.co.uk would not seem likely to be owned by someone this clueless, but I'll be darned if Alex doesn't seem to own the domain!)
