Slashdot Mirror
Why Haven't UPSes Been Integrated w/ PC Power Supplies?
antiher0 asks: "A while back, while connecting the batteries to 20 UPS's at work, I had a revelation. Well... maybe it wasn't that subtantial, but I thought of something that noone I talked to could answer. UPS's work by storing energy in a standard battery (undergoing a transformation from AC to DC in the process). When the power goes out or a brownout occurs the energy stored in the battery gets converted into AC and is sent down the power cable to the end device where may be converted back to DC again (as in the case of a standard PC). These conversions between AC and DC are not very efficient. Why not just have a system that places the battery and logic board within the power supply of the PC? (or at least hooks up to the power supply, the battery need not physically reside in the power supply) By doing this, you streamline the whole process and make the whole thing a bit more convenient. There are, of course, more issues to deal with, but I think it'd be feasable." This is a good point. If PC power supplies were external, they could be integrated with the battery/UPS and maybe we could save a bit of energy by performing one less AC-to-DC conversion which could potentially save some wear on the battery. Is there any reason why something like this couldn't work?
Aside from PC Power & Cooling, the PC power supply tends to be 'cheaper, cheaper, cheaper'. Adding anything more than what a basic ATX power supply needs to operate will just drive up costs and help only a small percentage of users whilst everyone else complains. That, and it won't exactly be cheap or easy to design... greater complexity, more regulatory aprovals needed, more parts to fail, etc.
I would imagine that if such a combo UPS/power supply were to be introduced, most folks and reviews would still prefer to purchase and use their own favorite standard models. Folks like a choice, even if it means taking more wattage from the wall. Just look at modern x86 CPUs. Someone could make a modern all-in-one workstation based around a newer rev of the Crusoe or StrongArm and some cots chips to accelerate video, etc. Have the chip makers and your own design teams understand the need for power conservation. Could easily end up with a highly efficent, cool-running, and still powerful workstation if it was a well-designed and well-managed joint effort. But... there is no way it would be cheaper than building your own PC piecemeal and just getting a beefy power supply.
More (electrically) efficent hardware would be nice. Powerful hardware is always nice. All-in-one is neat. Choice is good too. A cheap price would be perfect.... but they just don't go together too well, plain and simple.
I guess one of the problems with desktops is that the monitors still require AC power, so it'd be useless to have a battery integrated in to the box if it didn't have a DC-AC converter to supply power for the monitor. Then the power wasted to power the Monitor would probably make the savings from skipping the DC-AC-DC conversion for the computer not very worth while.
If you had a DC powered LCD screen, on the other hand, then it would probably make sense.
Because, I want to be able to replace a fubar power supply without paying for a new UPS. And if I upgrade to a new power standard (P4) I don't wnat to pay for a new UPS.
What's that you say? Integrate them but make the powersupply and battery components seperate? What's the point? It's easier to manufacture powers supplies for every consumer from one or two basic designs (desktops/laptops). KISS
To paraphrase Robert Heinlein, whenever someone asks "Why don't they . . ." the answer is always "Money".
The PC market is extremely cost competitive. The power supply is one of the least important components to most computer buyers.
Also, most computer users in general would turn all electronic equipment off in a storm. If you're on a reliable power grid, a storm is about the only time you might lose power. I can only think of two or three times in the last decade when I lost power when there wasn't a storm.
An added cost part that many people will never see the use of is a very hard sell in such a cost competitive product.
Here's the answer I found in the telecom digest FAQ:
i ve s/new-readers/frequent-ask-questions-97
Q: Why use a negative charge (-48 volts) for Ring instead of a positive charge (such as +48 volts)?
A: The reason for doing this is galvanic corrosion protection. A conductor with a negative charge will repel chlorine ions, as Cl (chlorine) ions are negative also. If the line were to have a positive charge, Cl ions would be attracted.
This form of corrosion protection is called cathodic protection. It is often used for pipelines, bridges, etc. Such protection was very important in the days of open wire transmission lines.
From:
http://mirror.lcs.mit.edu/telecom-archives/arch
(Take out the spaces)
The UPS is such a big chunk of hardware because it tries to output 110 volts 60 Hz. Putting it in the PC's switcher eliminates all those frequency and voltage constraints, and makes it vastly smaller, lighter and cheaper. Plus, you have other markets open up:
- People who want to put a computer in an RV, and don't want to run an inverter all the time they're away from mains power.
- People who use solar or wind power, ditto.
Your point about the power supply market having big economies of scale is a good one, but the number of consumers in California, New York and other areas which have a > 50% likelihood of being hit by blackouts in the near future makes a pretty good market by itself, and enough volume to justify a power supply aimed at their needs. If you could get it into Dell as one of their options, they'd sell like hotcakes.--
spam spam spam spam spam spam
No one expects the Spammish Repetition!
Scientists restrict study to entire physical universe; creationist
Dan Berstein asks a similar question and finds one or two answers. It seems like the only (readily found) solutions are American Advanced Power and Amsdell.
--
Except for a few special breeds (think ferroresonant transformers, motor-generators, or other solutions involving huge chunks of iron), all surge protection (including the much-touted APC stuff) consists of inexpensive varistors.
They operate by shunting to ground when voltage potential rises above a certain point (usually ~250VAC), which hopefully absorbes the spike until a circuit breaker or fuse pops, and hopefully even after that, should there be sufficient potential to arc across the fuse/breaker.
After a number of cycles of this, they eventually fail.
All of them.
And there is no method (at least, none that I've seen) which can accurately determine when they're no longer capable of conducting, short of physical examination (hint: if it's blown into bits, it doesn't work) or destructive tests (ie, hit it with 300VAC and see what happens).
That said, they're cheap. They're readily available (Mouser Electronics or Digikey). They're easily replaced if you're handy with a soldering iron. And, while periodic replacement of them will preserve the surge protection capability of the unit, it will also void any warranty or insurance you might have had before with your potentially-broken device.
(And, no, those little LEDs marked "Protected" on your surge protector are not functional. It's just a resistor and an LED across the power line, so it's absolutely nothing more than a power light. By design.)
No magic, here.
Kid-proof tablet..