Why doesn't somebody...

[Anathor21]

Seems there are quite a few rather knowledgable folks here so I'm curious if anyone can answer this one.

Why doesn't somebody create a powersupply/ups combination to replace the current solution plugging your PC into the UPS.

What I mean is this - Currently I have my systems on UPS's. Thus, in the case of a power failure the batteries in the UPS hold the sysem - the UPS converts it to AC, then the powersupply in the system converts it to DC to be used in the system thus looking like this:

Batteries (DC) -> UPS Converts to AC -> Powersupply input (AC) -> Converts to DC -> Consumed by PC components.

Since each AC/DC conversion is not 100% efficient it seems some power is being lost (heat generated). Couldn't a system be designed to skip the conversion from DC->AC->DC? ie. like this:

Batteries (DC) -> Powersupply (convert to correct voltages) -> consumed by PC components.

This seems like such a simple and obvious thing to do, has anyone done this? Is there some electrical reason that this would be a bad idea?

Looking forward to the discussion...

 

[Fushigi]

Most consumer-grade UPSes don't do double-inversion, i.e. convert incoming AC (line power) to DC (battery) back to AC (sockets for your PC). Most have separate charging circuits and do a cutover from line power to battery when an outage occurs. Hence the switching time that may be noted in the specs. Double-inversion has no switching time.

But that does raise a point. Why not have a larger PSU cage that includes a UPS? Say, enough for 3 or 4 minutes runtime to allow a graceful shutdown? Give it a motherboard connector for signals and let the OS control/communicate with the UPS directly. Drive the PC off the battery (no DC/AC conversion required) and use the line power to continuously charge the system.

Sounds nice, but there are issues. They'd include battery failures/added maintenance costs, additional heat in the chassis, higher upfront cost compared to a straight PSU, more space required, and probably a few things I'm not thinking of at the moment. Oh, some folks would want their monitor up; do you size for that (how much?) or say they'll still need an out-board UPS for the monitor and other peripherals?

Maybe an alternative would be a box the size of a current UPS. It would be a UPS but instead of AC out it would offer multiple 12V DC sockets (maybe 5V and other voltages). The PC, monitor, and other peripherals could tap the DC lines directly, saving numerous AC/DC conversions and probably eliminating internal power supplies on a lot of devices. Now, in this case, there should be cost & energy savings. Disadvantages? Dunno. Would the DC line suffer degradation over a few feet? Would it be likely to pick up (or emit) line noise/interference?

 

[P5-133XL]

There have been PS with builtin UPS's before. I don't think any have been financial successes. I think the market is too small. Servers, where price is less of an issue, generally already have a UPS infrastructure. Non-servers, almost always are price sensative and adding a UPS into the PS signifigently increases the price: Thereby the machines don't compete well. Consumers are too concerned with cost, and noise while a UPS is considered a luxurey.

I agree, that there are some efficiencies to be gained. To offest that, is cost, noise, longevity, and less convienance. The PS's would need to get rid of more heat and thereby need bigger louder fans. Batteries, need to be replaced periodicly creating either a shorter lifespan (if you don't replace them) or an inconvienance factor. Adding a UPS capability to a PS definately would signifigently increase the cost of those PS's. It is just more convienant to buy generic machines and add UPS's to those that need it rather than buying custom machines with UPS's builtin.

In general, I don't think that that feature is worth much to most computer buyers. You can always form a company and prove me wrong.

 

[Bozo]

Having the power supply supplied with DC creates it's own problems. To change the voltage from 12V to 5V or 3.3V would require resisters (lots of heat) or an electronic circut. The output from the electronic circut would be choppy at best, or more filters would have to be added. In a normal power supply, the AC voltage is reduced with a transformer and then converted to DC. Much simpler and a smother output.
A quality UPS will convert AC to DC and then back to AC to filter and regulate the output. This produces the cleanest power. (some refer to it as 'computer grade' output) This also gives a 'bumpless' transition when the power goes off.

Bozo :mrgrn:

 

[time]

Like just about every appliance I can think of, computers use a switching mode power supply - there is no transformer. Ignoring the effect of input filters, it doesn't much matter whether the input is AC or DC.

Voltages are dropped in the same way on your motherboard. Increasingly, the onboard regulators just chop the 12V line down to yield the various voltages that components require.

jtr1962 would be able to elaborate.

 

[Mercutio]

The issue that comes to my mind WRT integration of the UPS and PSU is how much the damned thing would weigh. Most tower cases would become extremely top-heavy if they had a big old UPS battery in the upper part of the case.
Maybe not a big drawback but certainly a design consideration.

 

[LiamC]

Then the answer is BTX

:mrgrn:

<ducks>
<dons kevlar suit>

 

[Handruin]

I thought the same thing LiamC. But I knew I'd get crap for it.

 

[jtr1962]

Like just about every appliance I can think of, computers use a switching mode power supply - there is no transformer. Ignoring the effect of input filters, it doesn't much matter whether the input is AC or DC.

Voltages are dropped in the same way on your motherboard. Increasingly, the onboard regulators just chop the 12V line down to yield the various voltages that components require.

jtr1962 would be able to elaborate.

Yep, that's pretty much how it works-you have a MOSFET switch which turns on and off at upwards of 50 KHz (to keep the switching inaudible and also to use smaller size active elements). The resulting output looks like a square wave equal in amplitude to nearly the original raw DC voltage but with a duty cycle less than 100%. This square wave is filtered to provide nearly pure DC via an inductor-capacitor filter.

Note that when using this approach you always need to start out with DC of some sort. The easiest way to get this is to rectify and filter the AC line voltage. When you do, you end up with a raw DC supply of about 165V, give or take, with a 120 VAC input. The reason for this is that 120V is the RMS value of the AC wave. The peaks are about 169V. You usually get losses of a few volts during the rectification stage, so 165V is a good approximation of where you end up. No matter, the exact raw DC voltage isn't critical since the circuit uses feedback to adjust the duty cycle*. In fact, the beauty of this approach is exactly that the power supply can operate comfortable over a wide range of AC input voltages (usually 85 to 240 VAC). In theory the supply could also work just fine with a battery supplying upwards of 100VDC or so. That brings us to the next problem.

SLA batteries are usually 6V or 12V. You probably could make one with 120V fairly cheaply if mass-produced. Problem is since a 120V battery is really 60 individual cells in series a lot of the package space is used for interconnects. That means less power density than a lower voltage SLA battery. The battery is easy enough to charge from the usual 165V raw DC voltage via another step-down regulator so that isn't the problem. A bigger problem is safety. What if some less than knowledgeable tech or even a home handyman takes one of these supplies apart? Yes, you guessed it-120V raw DC. DC is inherently more dangerous than AC from a shock standpoint. Unlike a high-voltage filter capacitor, which can't store much energy but will still give you a nice jolt, a battery will keep on giving you a shock on and on just like the famed Energizer bunny. You'll not only be very dead, but maybe even well done.

You can get around the above problem using a standard 6V or 12V battery but then you need to have another set of circuits to make the voltages the PC needs. A circuit which gives you 12VDC from 165VDC just isn't going to work with a raw DC voltage of only 6V or 12V. Hence, added cost. Obviously you get added weight, lots of it (as if PCs aren't already heavy enough), and in the worst possible spot as Merc deftly pointed out. From a cost standpoint it would probably cost more than a separate power supply and UPS. Like all things that try to be everything at once, such as those printer-scanner-faxes, it will end up doing none of them particularly well. And I think that answers the original question in a round-about sort of way (just like an episode of Connections).

*PFC complicates things a bit. Let's just say that full-wave rectification of an AC line creates nasty current spikes. PFC uses active elements to synchronize the current with the line voltage.

 

[Adcadet]

A bigger problem is safety. What if some less than knowledgeable tech or even a home handyman takes one of these supplies apart? Yes, you guessed it-120V raw DC. DC is inherently more dangerous than AC from a shock standpoint. Unlike a high-voltage filter capacitor, which can't store much energy but will still give you a nice jolt, a battery will keep on giving you a shock on and on just like the famed Energizer bunny. You'll not only be very dead, but maybe even well done.

Not to disput you JTR, but I was under the impression that AC is inherently more dangerous, given that most electrocution fatalities come from cardiac arrhythmias which are more liken in AC electrocutions. I'm not sure about a quick AC vs. a sustained DC (which I think is what you're talking about), but I've got to believe that either has the potential to trigger an arrhythmia if gien at just the wrong time (or cure one, as I saw attempted last week).

Of course, chared flesh is no fun either. But I'd prefer some burnes to v-fib any day.

 

[jtr1962]

Not to disput you JTR, but I was under the impression that AC is inherently more dangerous, given that most electrocution fatalities come from cardiac arrhythmias which are more liken in AC electrocutions.

At normal household AC frequencies of 50 or 60 Hz I believe this is correct but as frequency increases the current can't penetrate the body internally as well. This so-called "skin effect" is relevant even to high-frequency transformer design. At a few MHz the current basically flows only on the outer surface of the conductor, thus giving an apparent increase in resistance.

I'm not sure about a quick AC vs. a sustained DC (which I think is what you're talking about), but I've got to believe that either has the potential to trigger an arrhythmia if gien at just the wrong time (or cure one, as I saw attempted last week).

The reason I heard for DC being more dangerous is that it can freeze a person's muscles since it's a steady voltage. This means they may not let go of the source, resulting in continuous shock, whereas AC by definition has intervals where the person is not receiving a shock, and might loosen their grip. In general, the longer you're in contact with a voltage source, the more current will flow since the insulating oils will burn away. On the plus side, DC is more likely to just cause your heart to contract rather than fibrillate, resulting in a greater chance of revival. I felt the battery idea was dangerous because although the filter caps for the raw supply voltage in power supplies can give you a nasty 165V DC shock, they don't hold much charge, so the shock will be definition be short duration. Of course, it can still kill you. However, the battery has the potential to shock you for many minutes. Incidentally, it's really current that kills, not voltage, although you do need a fairly high voltage to induce a dangerous current flow through the body. That's why anything up to about 24V is safe although people with wet hands have been killed with 12V.

BTW, I actually did get shocked once from an 800 volt, 60 Hz AC source with no ill effects. There's a good chance I might not be here if it had been DC.

Of course, chared flesh is no fun either.

I've gotten that from shocks, soldering irons, and wires shorted on NiCads.

 

[P5-133XL]

Now, I've had the pleasent experiance of removing a 12V battery with a wrench that was in contact with my watch and having my watch spot-weld itself to the car frame. Now the voltage/current did nothing, but that spot-weld really did a good burn job to my wrist.

 

[Tannin]

I'm pretty sure I've seen combined UPS and PSU in a single unit. Long time ago though, way back in Baby AT days before ATX was around. If pushed to put a date on it, I'd say around late '386 or early 486 days. I don't think I ever saw one in the flesh, but they were a catelogue item from one or another of my suppliers, possibly the same people I still buy cables and spare PSUs and flash cards and stuff from.

They weren't around for long, I don't think, and I imagine they were one of those products that seemed like a good idea at the time but flopped in the market. They were quite expensive, maybe close to the price of a PSU and a UPS, but I don't really remember now. I only remember looking at the item, thinking "that sounds like a good idea", looking a bit harder at it and deciding it didn't seem like a goer. This would have been in paper catelogue days. Who knows? I might still have it around somewhere, the way I squirrel stuff away, but I doubt it after all this time.

One assumes that it was unsuccesful in the market, and imagines that any faint chance it had of selling well went down the gurgler when the 486DX/2 and VESA video cards and high-performance 5400 RPM hard drives and those newfangled CD-ROM things came along and upped the minimum power consumption of an average PC quite a bit.

Am I dreaming? Fantasies of old age? I don't think so. Maybe, just maybe, it was just a surge protection thingie, but that doesn't seem right: I wouldn't have remembered that, nor would I have even momentarily considered stocking one when stand-alone anti-surge devices were so cheap. I remember thinking seriously about this product, but only for a short while before I decided that it wouldn't fly. The more I try to remember, the more sure I become that it was a UPS, because one of the things in the "against" column was the standby time, which was too short. (No wonder!)

See, even then I was a cautious old bugger.

 

[Anathor21]

My original thought was actually not an integrated unit inside the computer, rather something that would live outside the computer - i.e. you would replace your normal PS with this "shell" device that would likely also break out the power cables for various devices. I know it takes a lot to "push" DC so I imagine you would not want the cables to be too long, but...

Now I have read the very technical (and somewhat over my head I'm afraid) explanation of how the current PS's work and all. From this the conclusion seems to be "this would be really big, heavy, and impractical". Ok, then how is it done in a notebook computer? I know power requirements are lower in notebooks, but still, they have managed to get them to work rather well and long with relatively light batteries. Thoughts anyone?

 

[Bozo]

Both AC and DC are equally deadly. With AC you have a chance to break free as the current passes throgh zero twice a second. (60Hz). But this plays hell with the heart.
DC is constant. It's hard to break free. While connected to a large DC potential, the current is burning the flesh, usually from the inside out. Just like an overloaded wire.
Maybe some of you have seen the hotdog with a nail in each end, and voltage applied to the nails?? Cooks the hotdog from the inside out. ( DO NOT try this at home!!!) Works best with DC.

Bozo :mrgrn:

 

[mubs]

Problems with DC are legendary. In the NJ/NC area, when Edison started out, he was a proponent of DC. Cables were buried a foot or two below the top soil. When it rained/snowed, numerous people and horses were electrocuted. This became an acute problem. In the meantime, Tesla, Westinghouse, Gaulard and other were championing AC for it's safety and manageability. I've seen long PBS documentaries on this; see Wikipedia here. There was much at stake, and Edison resisted AC till it became overwhelmingly clear that for distribution, AC was so much safer and less lossy, and he had to abandon DC. I may not have the details right, but the gist is correct.

When I was in college, a friend who was studying electrial engineering, would always tell me to use the back of the fingers (fingernail side) in a tapping motion to test for leakage (as in JTR's capacitors). That way, if there was a leak or a charge, the fingers would curl inward towards the palm, away from the source of the leak. Too many people have died from not being able to let go even when they wanted to.