Dan's Data letters #5Publication date: 25-Oct-2002.
Last modified 03-Dec-2011.
I am going to make a prototype for a vehicle. For the wheels I want to use hard drive disks on their sides, so they spin like wheels normally do. On your site I read about the voice coil design; could you help me come up with a way to use one to cause the disks to spin and result in a moving vehicle?
I plan on making the body of the prototype out of a light weight plastic with wheel wells that accommodate the tall disks (wheels). I figure I can tinker with an axle system at some point in the future, but if I could get these wheels to spin by somehow imposing a magnet force on them like you described it would make an interesting little vehicle.
What do you think about using the NIB magnets in conjunction with the voice coil design you described on your site? Better yet, enclosing ferrofluid in a tube or some 3D ring that would act as a wheel and superimpose a magnet on it causing it to rotate?
Hard drive voice coils do not turn the disks. There's a brushless motor that does that, as you'll see when you open the drive.
Direct magnetic drive for the disks would require you to spin a magnet next to them, which is unlikely to be what you're interested in; you might as well connect the magnet-spinning motor to the wheels directly. You could maybe make a multiple-pulsed-coils setup work, but it'd be a lot more trouble than it's worth, if it's even physically possible, and it wouldn't be unless you attached something to the disks for the magnets to pull or push on; plain aluminium drive platters won't spin if you just pulse coils around them.
The practical solution is a conventional motor with a gearbox. Hard drive motors don't have enough torque and are meant to spin far faster than you want - a 7200RPM drive motor spinning one of the 3.75 inch platters you'll generally find inside a "3.5 inch" drive would, if it could get to full speed, propel a vehicle using that platter at about 35.9 metres per second, which is almost 130 kilometres per hour. It'd never actually do that, though, because a drive motor has barely enough torque to even start a little car rolling.
You could use a drive motor with monstrous gear reduction, like a worm drive, if you were really hot for the idea - but it'd need driver hardware. You can't just connect DC to a hard drive motor and get it to spin, like a computer fan will; it's more complex than that.
My mate's just purchased an ASUS V8420 video card and has been having problems with his computer switching off for no apparent reason! His motherboard is a MSI K7T266 Pro 2. I was wondering if you have come across a problem mixing the two together? If so, can you help?
I doubt it's an incompatibility issue; my money would be on either damage to something else (he static-zapped something while installing the video card), or a dodgy PSU which just happened to go dodgier at the same time he did this upgrade. Bad drivers could also have something to do with it; I've seen software problems that manage to turn a computer off plenty of times. They're definitely one of Windows' most appealing quirks.
If he's not running the Nvidia reference drivers, he should try them before he tries anything else.
I wonder if, as a broadband man, you remain interested in dial-up stuff?
I sent the following to Maestro the other day and have yet to receive a reply. Surprise, surprise.
I've just moved house. I used to dial my ISP via an exchange about 100 metres down the road, and now I'm well and truly rural, with the exchange many kilometres away.
With my old LT Winmodem I used to get solid V.90 connections at 53600 line speed with 5 to 6 kilobyte per second downloads; now I get an extremely sluggish 26400 initial connection with about 1 to 2kb/sec, frequently interrupted, downloads. By disabling V.90 and forcing V.34 I get reasonably solid 28800 connections with about 3kb/s download speed; only occasionally do I get a (solid) 26400 connection.
With the Lucent chipset, ATi11 reports: Rx -22dbm, Tx -9dBm, Noise -77/-78dBm. At least I think that's what the noise "number" represents, it doesn't specify. The line sounds quiet, there's no obvious interference. I have the facilities to measure absolute signal to noise ratio but haven't done so yet... but it seems that 6dB or so improvement might be all I need to get V90 connections...
I am therefore thinking of buying a Maestro Woomera, but I'd like an honest opinion as to whether it is likely to give a significant improvement, bearing in mind that I don't know what my line impedance is and 28k isn't too bad anyway.
I realise you don't have a crystal ball, but I would appreciate your opinion. I'd rather not spend the money and find that I'm still connecting at 28k.
Now, the "Woomera" incorporates fuzzy logic and eight impedance matching options, so it's quite possible that it could give me a signal to noise improvement. It can also automatically adjust Tx levels, presumably to the lowest level (that still works) to reduce crosstalk.
In short, Maestro reckon it's the best thing for rural dial-up since the Stelvin Seal started to appear on bottles of red. But I'm personally tired of forking out money to pay for marketing hype; I'd like to try one myself. Only the local distributors don't stock them, they'll only get one in if you order it. And I hate going through all the palaver of returning something that doesn't come up to expectation.
Telstra tell me I'm not on a pair gain line; do you think the Woomera would help me?
I'm guessing that you've moved somewhere that does have pair gain phone service, regardless of what Telstra say. You're served by a multiplexer that's stuffing a couple of phone lines into each wire pair back to the exchange, to save money. This is common in the bush, and in new housing estates too. Small pair gain Rural Access Multiplexers are twin-line systems that cause pretty much the exact symptoms you're reporting. Some poor buggers in the sticks are on four or eight line multiplexers.
This issue has contributed significantly to all of the protests and political point-scoring here in Australia around Telstra, privatisation, rural service guarantees, blah blah blah.
If you're on pair gain, you're screwed. No power on earth will get you a significantly faster connection than you've got now, unless your phone line's upgraded to a normal un-multiplexed one. You also will not be able to get DSL service, at any speed, until such an upgrade is done. A funkier modem is unlikely to make any difference at all.
No, wait. I don't know what I was thinking. The fact that Telstra say you're not on pair gain settles it. Telstra are never wrong.
Is it possible to make a standalone CD player out of leftover parts such as a CD-ROM drive, sound card and power supply?
It's possible, but it won't be very user-friendly. You can just use the CD-ROM drive's analogue output, and connect it to anything that takes line level input, and then use the buttons on the front for control. No sound card needed; just plug the thing into a PSU (ATX PSUs need to be jimmied by connecting pin 14 (the green wire) on the ATX connector to ground if you want them to turn on when they're not connected to a motherboard) and you're away.
If you want anything fancier than that, though, then it can't be done. Don't expect a display, easy forward and backward track skipping, and so on.
I just bought a Lian Li PC-9300 and I'm trying to find a suitable P4 compatible PSU. I know it's SFX form factor, but I'm not sure any SFX PSU will fit. The Lian Li website doesn't have any information. Do you know if any of the following are OK?
And in your review, you fitted a long thin PSU in. Which form factor was that?
As you suspected, any old SFX PSU won't fit in the PC-9300, because there's more than one SFX PSU box size.
The SFX spec allows for 50 by 100 by 125mm PSUs with the back of the PSU (the side with the IEC input socket) being a 50 by 100mm face, and for 63.5 by 100 by 125mm units with the back being a 63.5 by 100mm face. The PSU that fits the PC-9300 perfectly is, I think, a 50 by 100 by 125mm unit.
The FSP PSUs you mention are 76 by 125 by 100mm units, which seem to have the IEC input socket on one of the 125 by 76mm faces, which is no good. The Yeong Yang ones are 63.5 by 100 by 125mm, the second standard SFX size, but that's no good either.
Pretty much anything with a 50 by 100mm back panel should fit; there's a decent amount of room for a longer box, if you can find one. These dinky little PSUs are annoying to source, though.
The long thin PSU that fits, more or less, in the PC-9300 is a flexATX unit from AOpen. It's 175 by 85 by 65mm, and the back is one of the smaller faces, so it pretty much fits.
I read your review of the Benq VP150X projector and finding a cheaper alternative, and the part about smaller (more inexpensive) projectors not having image flipping capabilities made me wonder - if I have one of these with a VGA input, are there any applications/utilities/options that can be used to make a regular PC do all of the flipping for you?
There are a couple of joke programs that do it, but I think they all do it randomly, which is of limited utility. Apart from that, there aren't many options - though Nvidia's Detonator 40 drivers, in beta as I write this, have a new feature called "NVRotate". That seems to do exactly what you want, assuming you've got a suitable NVidia-chipset graphics card, of course.
There's another way around the problem, though; to mirror an image, use a mirror! Aim the projector 90 degrees away from the screen (up or down, for vertical mirroring; left or right, for horizontal mirroring) and use a mirror to reflect the image onto the screen. Bingo - one reversed image.
I didn't think that any search engine would ever turn up someone who made their own camera battery pack AND gave directions on how to do it! Very briefly, I have a Canon PowerShot S10. I was thinking of wiring a cut-off AC adapter end right to a 6 volt rechargeable lantern battery. Will it fry?
It probably wouldn't fry, and it'd probably work pretty well.
A non-rechargeable lantern battery, on the other hand, wouldn't work well. I don't have proper PowerShot S10 specs to hand, but a quick Usenet search (see this, for instance) suggests that a lantern battery should be fine, but only until its terminal voltage dropped to about 5.7V. That's not good, though, because at that point the battery will still have a lot of its capacity left. Check out the PDF-format datasheet here, for instance; the discharge curves on the second page make clear that a 5.7 volt cutoff point will leave from around 80% to well over 90% of the battery's capacity unused, depending on the load.
A six-volt rechargeable battery, on the other hand, would be good. All of the rechargeable battery chemistries hold their output voltage close to their nominal voltage quite well as they discharge, and only slump substantially towards the very end of their charge.
You can get six-volt sealed lead acid (SLA) rechargeable batteries in the lantern format, and in a few other shapes and sizes. These batteries are heavy, but compact, and can easily be carried in a belt pouch of some kind; they should give you well over 90% of their rated capacity before they drop to 5.7 volts. A roughly 600-gram (1.25-pound) 2.8-amp-hour 6-volt SLA battery and a decent wall charger to suit it shouldn't cost you more than $US30, and should give you easily four times as much run time as the standard PowerShot S10 battery.
My roommate and I are pretty into car stereos, both of us have two amps in our cars to power speakers and subwoofers. Our question is about adding a capacitor to the setup.
My roommate has a 1997 Subaru Outback sport wagon, running about 320 watts of continuous power. The car has the stock 85 amp alternator with a normal size car battery. His headlights dim whenever he has the volume up on his stereo. He's thinking of adding a capacitor to help solve this problem.
I, on the other hand, have my stereo in my 1988 Chevrolet Blazer running about 280 watts of continuous power on a 115 amp alternator and large truck battery. I don't notice any light dimming or gauge dropping until I turn on all my accessories with the volume up.
My question is: what exactly would adding a capacitor do for systems like these, and would it be a good idea for my roommate, me, or both of us to add capacitors to the systems?
It won't help.
Adding a cap to the power supply side of any circuit will smooth out dips in the supply voltage caused by the power source - the alternator, battery and supply wires, in this case - not being able to keep up with the instantaneous current draw of the load. The cap charges when the power demand is lower, and discharges when it's higher, but the capacity of even the hugest super-ultra-monster-caps is trivial compared with the actual power demands of even a modest car stereo. A one farad cap charged to 13.8 volts can deliver an average of one amp for one second into a suitable load. That's 13.8 watt-seconds, which will power your stereo for a bit less than a twentieth of a second.
If the lights dim, but the alternator is still managing to keep the battery charged, then I presume the lights and the stereo aren't separately wired from the battery - everything's coming back to taps on one positive cable from the battery, and/or the negative cables aren't too well connected to the chassis. If your friend gives the headlights and other car systems their own separate parallel connection to the battery, instead of splicing everything into the one cable-set, he ought to solve the problem. Wiring everything up with ultra-low-resistance super-fat show-off one-gauge hyper-cable is optional.
If the problem is that the alternator can't deliver as much current as the stereo (and other systems) can draw, on the other hand, then your friend needs a heftier alternator. The stock one ought to have more than enough capacity, but it might not be living up to its rating. If it isn't, then your friend will be able to run the battery flat while driving around, and only the fact that the stereo isn't on all the time when he's driving is saving him from not being able to start the car again afterwards.