Dan's Data letters #190Publication date: 20-Sep-2007.
Last modified 03-Dec-2011.
Recently, I decided it would be a good idea to start investigating this whole "Wireless USB" thing.
I did the Google searching (even the "I'm feeling lucky" button...) and checked your site for any possible information. All I've been able to come up with so far is a news release from Belkin, and the ability to purchase such an item from a few well-known online retailers.
What I've failed to locate is any review of the performance, reliability, functionality, etc. of wireless USB devices.
If you've heard anything about these, or even used them yourself, I would love to hear about it.
As you've noticed, Wireless USB gear barely even exists in the consumer market at the moment. Seeing as the only way to connect anything via "WUSB" so far is by plugging a normal USB device into a WUSB hub, this is not entirely surprising. There just aren't many situations where a plain old cable doesn't work just as well.
It'll probably trickle down into the regular market at some point; I'm guessing that WUSB-equipped cameras will be a big selling point, if we don't see mainstream WiFi-equipped cameras first (Kodak's EasyShare One doesn't seem to have led to anything much, and it is my considered opinion that the EOS-1D Mark III doesn't qualify; its WiFi module apparently works very nicely, but camera plus module will cost you $US5500).
Until that happens, though, the gear's rare enough that I don't even know who to ask for review product here in Australia, and it's boring enough that I don't much care.
(The existence of multiple competing WUSB "standards", as explained on the Wikipedia page for the technology, doesn't help. I'd guess that the "Certified" version is the one that'll survive, but right now I'd give the whole thing only about a 50% chance of catching on at all. It's not as if WiFi and Bluetooth don't already have the high- and low-bandwidth markets sewn up, after all.)
Your review of the Australian-wired power meter was great, and linked to the "Cable That Should Not Be" you made to run your fridge from a light socket. I'd read it before, but I frequently re-read your older articles because they're so darned amusing.
In re-reading your cable-that-should-not-be blog entry, your explanation of hot-vs-neutral problems using a toaster made me jump. You wrote:
But the primary purpose of earthing is to protect you from being zapped by a faulty appliance. [...] If the toaster's earth wire is connected, the above failure will trip an RCD in milliseconds [...] With no earth wire, the small load created by a dying human body will not bother your house's breaker box at all.
The reason I jumped was because such a toaster would be a violation of the electrical codes here in North America. A toaster has its electrical element exposed, so an idiot user with a knife can touch it if they try. Any such exposed-element appliance must be double-insulated here - you're not allowed to ground the chassis, because then the idiot knife-wielding moron, who is almost certainly steadying the chassis of the toaster with their other hand, is far more likely to kill themselves by shorting across their body from the element to the chassis.
So, I gotta ask: Are Australian toasters really allowed to have a grounded metal chassis?
I think it's technically impossible for a normal toaster to be "double insulated". That, by definition, means all energised components must be completely enclosed by two layers of insulation. I don't think a toaster with an element you can touch with a knife can qualify.
(I suppose you could still make a double-insulated toaster if you give the elements quartz-tube enclosures or something. Toasters are still likely to use old-fashioned bare nichrome-wire elements, but toaster ovens today are likely to use "Calrod"-type sealed elements, which look like a short thick piece of bent metal rod, but contain a nichrome coil in an insulating ceramic binder, which does not electrically contact the outer metal shell. Electric jugs use these kinds of elements, too.)
Most modern toasters have a plastic outer casing (the "plastic" parts of old toasters are likely to be Bakelite), but there's usually still chassis metal accessible around and between the slots. And yes, just to make sure I wasn't hallucinating, I just checked my toaster. That metal definitely is earthed.
While this does indeed make it possible for the user to become part of the circuit if they hold onto the toaster (touching some bit of metal somewhere) and probe around in the slot with a metal implement, it's actually very difficult to do this without your implement also touching the earthed metal at the top of the slot, whereupon you won't need any fancy safety switches to protect you - the circuit breaker will trip.
We're presuming that you're poking the implement in there because you're trying to fish out some stuck toast, after all. So you won't be carefully lowering your knife "Operation"-style into the middle of the slot. You'll be sliding it down the side, and it'll touch metal.
I don't know exactly why it is that toasters in the USA and Canada aren't grounded, excepting the fact that they never have been in the past. There are still quite a few old houses in the USA that don't have three-pin power sockets, but you'd think you'd at least be able to buy both kinds. But electrical regulations commonly include all sorts of ancient grandfathered-in weirdness, so I'm not at all surprised that, in North America, you can't.
There's a good page about the many ways in which toaster-like appliances can be grounded, or not grounded, here.
As Charles pointed out to me in a later message, the Electrical Wiring FAQ, among other documents like this US government PDF, is adamant that any appliance with exposed conductors must never be earthed.
The Sci.Electronics.Repair FAQ has something to say about it, obliquely, as well; it mentions that a GFCI ("safety switch") is desirable on US kitchen outlets into which something like a toaster may be plugged.
The explanation given for this initially mystified me, though. Jamming a knife into an unearthed toaster and shorting the element to the casing will not, in itself, trip a ground-fault detector or any other kind of protective device. All it's doing is electrifying the casing, which isn't connected to ground or to neutral or to anything else, so as far as the house wiring can see, nothing out of the ordinary is happening.
The safety switches they're talking about, though, work by detecting a difference in current between what flows into the appliance on the active wire and what comes out on the neutral wire. Any significant discrepancy may be passing through a person, so the switch trips. And if you touch the live chassis of your shorted toaster with one hand and something that's reasonably well earthed, like the sink, with your other hand, that'll create a live-versus-neutral current discrepancy that'll trip the switch. If your body is connecting live to neutral, though, the current in each will remain identical - just rising by however much is passing through you - and the switch won't notice a thing.
In this same situation, a grounded toaster would have long since popped the breaker or safety switch for the circuit it was on.
Perfect reproduction of humorous vocal SSB imitations a plus
I work for a company that specializes in fire and water damage restoration. My job is to monitor all fire department radio traffic from Houston, Texas and the surrounding areas.
I basically sit in a room with a bunch of computers and a bunch of speakers and listen to people talk on radios all day.
Now, the speakers we have do a fine job... music-wise (eight sets of Altec Lansing VS2220s). But when push comes to shove (i.e., multiple major incidents happening at the same time), they can make my job harder. Fortunately, we are in the process of upgrading our equipment and I've been asked to provide some input into the matter.
So, what should we be looking for in speakers so we can get the best voice reproduction? Money isn't necessarily that much of a concern, but I'm pretty sure the company would smile more upon a not-so-expensive option.
The main limiting factor for intelligibility of radio and phone voice traffic, I think, is the restricted audio bandwidth of those transmission systems. Standard telephone audio only goes up to 3.3kHz, which is significantly less than optimal when you're trying to tell apart words that only differ in consonant sounds most of whose energy is higher in the spectrum than that.
There's a whole branch of acoustics having to do with speech intelligibility - see this page for a quick primer. The take-home message, though, is that the inherent intelligibility of speech transmitted via lots of common low-bandwidth communication systems is lousy, and there's not much that can be done with a speaker to make it better.
Most ham radio and commercial CB speaker setups use a single small widerange driver in a crappy plastic box. But that awful little speaker already has more response than it needs.
That said, the computer speakers you're using aren't necessarily quite right for the job. It's possible, especially if the radios you're monitoring have better audio quality than the average, that you'd get better results from some basic small "monitor"-type speakers. They'll be clearer at medium-to-high volume, if nothing else.
Before you rush out and buy new gear, I suggest you tote an ordinary hi-fi amplifier and any old wood-cabinet bookshelf speakers into the office and try them out. If there's a definite intelligibility advantage, then some basic self-powered monitor speakers should be just as good.
There's no need to spend a lot of money. Something cheap (like a "Clones-R-Us" Behringer product...) will probably do the job.
Note that typical "mini system" or boom-box speakers are likely, if anything, to make speech less intelligible. A lot of computer speakers have the same problem. Monitor speakers are meant to sacrifice total frequency response for flatness over what response they do have.
I read with interest your article on the reason to only install 3Gb RAM due to "memory hole" issues.
Is this true for Intel-CPU Macs? I assume it is, because you mentioned that it applies to Linux as well.
- me not brilliant, just CAD guy
- had nth Windows fatal crash on various machines over 15 years
- was running RAID1, just in case
- RAID1 useless in event of software failure
- me pay money (again) to fix Windows
- me use excuse to buy MacBook Pro
I'm running WinXP Pro on my 2GB Mac under Parallels so that I can use AutoCAD and zoomy 3D exploration software (NavisWorks JetStream to design check/comment on process piping plants).
It's working better than I expected but still slow, and a bump up to 4Gb RAM would cost a rather pricey 600 Canadian dollars.
This price might be worth it if there is a significant speed-up during model fly-thrus, and I wonder if 4Gb would make a difference when running two OSes at the same time.
I haven't looked into it in depth, but I think Intel-chip Macs, at least the ones that aren't based on Xeons, do have similar memory-hole problems to PCs.
Very shortly after I put this page up, two readers wrote to inform me that the very latest collection of Macs that use the Santa Rosa revision of the Centrino chipset are, genuinely, free of memory-hole problems. That seems to make sense. Someone with more clue than me is still welcome to correct me on the following, which I'm now about 70% sure only applies to Macs before the newest generation:
Old PowerPC G5 Macs have some kind of pseudo-64-bit thing going on, like PAE, that apparently lets them avoid most memory hole issues.
Pre-Santa-Rosa Intel Core and Core 2 Macs - everything except Mac Pros - have memory-hole problems. They don't have the old DOS-compatibility memory holes, and their video adapters only have at most 256Mb of memory and so only make a hole that big, but that hole is still there. And probably a few other smaller ones as well.
I think this is why, for some time, the only Core-CPU Mac you could order with more than 3Gb of RAM was the MacBook Pro. Then the iMac was the same, and I think for the same reason - it has to have matched memory modules, and only has two slots.
So, on those pre-Santa-Rosa Macs, all 4Gb was not available. But the video adapter was only a 256Mb unit, so the shortfall wasn't that terrible.
I don't think the 64-bit-enhanced OS X Leopard will be able to fix these hardware limitations, on the pre-Santa-Rosa Macs.
Mac Pros are Xeon-based, and have always been able to take up to 16Gb of RAM with, I think, no significant memory holes.
So, in answer to your question: 4Gb is likely to be quite a bit better than 3Gb if you really are running out of physical RAM, and with a Santa Rosa Mac there should be no drawbacks at all. With a previous non-Xeon Intel Mac, though, I don't think you'd be able to use all 4Gb.
Your readers might like to hear your thoughts on the "Linear Array Transducer" or LAT, their "first radical redesign of the subwoofer "in almost 80 years".
I am skeptical.
I remember seeing this when it was widely reported, a year ago; I don't know whether it's made it to any actual products yet. Your Stereophile link informs me that Tymphany have been sending out press releases about the LAT since 2004, and not much seems to have come of it yet.
UPDATE: After I put this page up, a reader pointed out this Alpine car subwoofer, and its smaller sibling. This rather breathless reviewer reckoned the bigger one was quite good. A couple of the customer reviews here are less enthusiastic, but the ones for the smaller sub, which is made for tight spaces, are better. Neither seems to have made much of a splash.
UPDATE 2: Here's a dealer selling a bare LAT unit, which Tymphany themselves still list as an "OEM" product, for only $US58.
And now, back to what I said when this page was new:
Given that the Tymphany spiel about the product is that it can "bring great bass to the new generation of consumer lifestyle products", though, we're probably talking about pretty low values of "great".
Here's the kind of thing Tymphany expect the LAT to be used in - a slimline "sound bar" for the top or bottom of a big-screen TV.
It beats the pants off crappy not-really-a-woofer-at-all options for this application, but it still only takes bass down to about 70Hz, which is far from real subwoofer territory.
The LATs that Tymphany have datasheets for are presumably larger units than the soundbar version, though. They're alleged to have a resonant frequency of 28Hz.
There's unquestionably room for improvement of the basic cone loudspeaker design, which hasn't changed in essence since the end of the 19th century. Usually, improved speaker designs just use ordinary cone drivers installed in some new and exciting configuration, like the isobaric dual-driver configuration or the Bose boxes-full-of-little-drivers and long tuned enclosures. There's no reason why a genuinely different design couldn't be better, though, and the LAT doesn't look ridiculous on its face.
If they really were that great, though, you'd think they'd have turned up in more than one, indifferently received, commercial product by now.
A few days ago I read Mark's review of LSK's F6s. Since then I've been becoming more and more interested in building a couple of kit speakers myself.
Since I'm in the states, LSK isn't an option. One of the kits I've been reading about is GR Research's AV-3 kit.
What I'm curious about is the capacitors. As an optional upgrade, they offer SoniCap capacitors for $64.
My understanding of speaker crossovers and capacitors is limited at best, so I just don't understand if expensive capacitors could actually make a noticeable difference in sound quality.
Is there real gain to be had in fancy caps or is it just more audiophile rubbish?
[yet another] Dan
It's audiophile rubbish.
There is some reason to opt for fancy expensive capacitors in speaker crossovers, because crossovers require high-value capacitors that're going to be sitting in there for years, probably with little to no prospect of ever being replaced. Your basic cheap "wet" aluminium electrolytic capacitor will dry out over time and change its value, and will have a less accurate value in the first place.
It took me a while to figure out what kind of capacitor the "Sonicaps" actually are, because the manufacturers rave on about "warmth" and "detail", and even fantasise about capacitors having "noise", without remembering to tell their customers what kind of bloody capacitor they're selling.
They're described as "film" caps elsewhere on the SonicCraft site, which together with other Web pages led me to the conclusion that they're polypropylene-substrate metallised-plastic-film caps. Those do indeed have better value accuracy and long-term stability than cheap electros.
A 10 picofarad 200V "Sonicap" has a $US17.60 list price. The same thing from a normal retail electronics store is about $US10.
Find further reading about the uncomfortable intersection of audiophile opinions about capacitor characteristics, and what anybody can actually repeatably detect, here.
Are you aware that JPS Labs has an advert on rotation at your site?
I didn't think you would be too happy with that, as some people may see it as some kind of implicit approval of their products.
I presume it's a Google ad. If it isn't, it may be surprisingly difficult for me to block it; every now and then someone complains about a particularly seizure-inducing pop-up, but even if they tell me the name of the entity being advertised and/or the Web site the ad clicks through to, I still can't figure out which one it is in the ad management console thingy.
(This is often because Burst Media, the purveyors of all of my most annoying ads, sometimes sublet ad locations to other ad companies. If Burst puts an ad supplied to them by Company X on my site, and Company X is another ad agency that has its own client list including some horrible fake-error-box ad for Registry Munger Pro, I'll only see Company X in the ad management console, not Registry Munger Pro.)
I don't think it's really very likely that people will think I'm endorsing audiophile voodoo, as practically everybody who uses the Web today has seen many examples of contextually-delivered ads that're humorously opposed to the message of the page on which they appear.
I'm also inclined to lean towards the happy-to-take-their-money end of the spectrum. But I would say that, wouldn't I?
I'll block ads that're really obnoxious (or just not take them in the first place), and I'll also block ads from companies I specifically talk about in a negative way, like the nutballs at Life Technology.
So, yes, I've now blocked Google ads for jpslabs.com. But this page will probably attract some other audiophile flim-flam ad.
Thought you might like to make fun of the Riddex Plus. (Warning, annoyingly loud auto-playing Flash video on the homepage.)
It goes beyond the usual useless ultrasonic pest repellent, and somehow makes your electrical wires repel bugs.
But naturally, one unit can only make ONE floor of your house pest-free. Somehow the power wires in the floor below don't get the benefit.
Even if this somehow made power wiring vibrate in an ultrasonic frequency, which I don't think is possible but is the only way I can imagine it being meant to work, I don't think I'd particularly want my electric wiring constantly subject to shaking.
The testimonials are great. Martha's been using this product for eight years. A product that only just started being advertised!
Somehow words like "digital", "Motorola", and "microprocessor" just make everything better.
Ah, but you see, the non-Plus "Riddex", and numerous products with other names, have been around for years. Why, right here on the FTC's site the manufacturers are agreeing to stop selling the things until they can substantiate their claims.
That was in 2003, as was this complaint: PDF
And here they are selling them again!
So they must have substantiated their claims, and now be totally on the level!
I saw your humorous warning in one of your reviews re: boiling crowbars with amplifiers, and did some back-of-the-envelope type maths with assorted constants pulled from Google:
Density of steel: 7.8g/cm^3
Standard crowbar dimensions: 96cm long, 18mm diameter
Based on this, an average crowbar weighs 1.3kg (I know this seems a bit light, but run with it...)
Specific heat capacity of steel: 450 J/kg*degrees Celsius
Boiling point of steel: about 2700 degrees Celsius
Therefore we need to add about 1,600,000 joules of energy to a crowbar to raise its temperature sufficiently to cause it to boil.
At this point I arbitrarily decided that this energy would need to be delivered in at most one second to really call this "vapourisation". Which makes the calculation of amplifier output power quite easy.
Result: 1.6 megawatts.
That's not an amplifier to laugh at.
Assuming a decent PA amp can provide 2.5kW or so per channel, and is (hopefully) stereo, we'd need 320 of them in parallel, bridging the channels.
Assuming the crowbar's resistance is approximately 0.002 ohms (estimated based on resistivity of steel, cross-sectional area of crowbar and length), we would need 56 volts at 29,000 amps (at minimum) to achieve this.
Of course, I'm ignoring all kinds of real-world things like the fact that as the crowbar heated up its resistance would increase, that crowbars aren't in fact perfect cylinders, heat losses to the environment, complete and utter disregard for mixing initial values for different kinds of steel, etc....
Now point out where I've gone horribly wrong. (I think a good starting point would be "attempting to figure this out to begin with".)
Seems kosher to me.
And I don't think there's anything at all wrong with attempting this important calculation.
Us nerds don't get any respect at all if people don't know we spend time figuring out exactly how many grams of TNT a 200-litre garbage bag full of a stoichiometric mix of oxygen and acetylene mix is worth.
There are some good YouTube videos on that subject, by the way.
They demonstrate the difference between mere deflagration and a true explosion quite well.
To the best of my knowledge, nobody has ever actually vaporised, or even melted, a crowbar-sized chunk of steel with any kind of audio amplifier.
People have, however, occasionally dropped - or at least SAID they've dropped - crowbars across the busbars of high-capacity electricity generation, distribution or use facilities. Aluminium refineries, for instance.
I believe the result cannot, strictly, be described as vaporisation. Most of the steel ends up as small puddles relatively close to the site of the incident, not condensing from a gaseous state all over the facility.
I'm also given to understand that one should procure a very long piece of wood, and use that to lower the crowbar into place.