Slashdot Mirror
Google Calls For Power Supply Design Changes
Raindance writes "The New York Times reports that Google is calling 'for a shift from multivoltage power supplies to a single 12-volt standard. Although voltage conversion would still take place on the PC motherboard, the simpler design of the new power supply would make it easier to achieve higher overall efficiencies ... The Google white paper argues that the opportunity for power savings is immense — by deploying the new power supplies in 100 million desktop PC's running eight hours a day, it will be possible to save 40 billion kilowatt-hours over three years, or more than $5 billion at California's energy rates.' This may have something to do with the electricity bill for Google's estimated 450,000 servers."
google hires experts on Electrical Engineering to figure out how to reduce the power bill on those 450,000 servers. Hell, I'm all for it. Less power means less heat means quieter fans (w/o spending an arm/leg on an Antec Sonata or whathaveyou).
Hi! I make Firefox Plug-ins. Check 'em out @ https://addons.mozilla.org/en-US/firefox/addon/youtube-mp3-podcaster/
In the old days, disk drive motors and fans. But many of these now run on 5V, hence the cheap USB-powered drive cases out there. Chips at CMOS power levels run at 3.3v, TTL is 5v, but hardly anything runs at 12v anymore. It seems to me that if they'd just pick their hardware carfully, they could run their entire server rack off of 5v+- rails.
SJW: a person who perceives an injustice, and while correcting it, commits a greater injustice.
I assume Google is employing some smart electrical engineers, which are more than qualified to make this kind of recommendations I would think...
-- the cake is a lie
If a caveman were to tell you "fire hot", would the fact that he is not educated invalidate his findings?
The spleen? You don't need it.
Its a nice idea and one that is probably a long time coming, but phasing something like that into place will take an incredibly long time. Look at the struggles of PCI express, its still not in 50% of the newer motherboards and systems though its benefits are more than apparent. Its just been in the past couple years that we have seen a shift to full usb and most machines still come with ps2, serial and parallel ports anyway. Dramatic changes to the PC standards are very difficult, there are millions of existing machines that still need support. Perhaps if it was tied to a new socket standard in the future it could slowly be phased in through upgrades, but I see the chances as very very slim.
I would bet a lot of the employees at Google have Electrical Engineering degrees. Don't underestimate the brain power Google has in its employee base. But the power supply issue they're trying to address isn't a technical challenge, but a political challenge.
Combine it with a 12 V battery and eliminate the cost and inefficiency of a UPS. Why isn't the battery inside the computer?
I'm surprised Google isn't running off 48V DC power supplies already, which, from what the documentation has shown me, already exhibit some of these savings...
On the other hand, it might have to do with Google's policy to use as much off-the-shelf equipment as possible, which 48V is not(iirc), so unless the "off-the-shelf-standard" changes, google might be in a position that they have to either break their own rules, or pay for following them.
Considering the number of servers google has, I'm surprised they haven't made 48V mandatory, just for the durability factor though. Converting to DC outside the pc also reduces noise, vibration, and allows bigger, heftier, more efficient conversion IMO, it also leaves more room inside for what the server does best, process data.
At the recent linuxworld in sf, I noticed some dc power supplies being pushed in the rack pc sector. I guess the lack of conversion from AC to DC saves a bit of juice, which makes a difference in large colocation centers. Combined with dc conversion on the motherboard, it would just be a matter of hooking up 12V DC direct to the board, which would be much nicer on the equipment and save a bit of power.
You can say goodbye to USB powered devices. An example would be the canned drink cooler.
Thanks,
Jim
Well they do have a lot of servers Im not sure they should be a guiding force in this but I think they can put the sugestion out there.
$DO || ! $DO ; try(); > try: command not found
5 volts harkens back to the days of 7400 series TTL.
Time to give it a rest if it's not a necessary voltage.
A feeling of having made the same mistake before: Deja Foobar
Any links to the Google white paper detailing their reasons for this system architecture?
No, but your analogy assumes that Google's proposal is an obviously good engineering solution, which hasn't yet been proven.
The ability to have all my machines powered by a heavy cable carrying 12VDC would be pretty useful for several reasons.
*sigh*
Check out my sci-fi/humor trilogy at PatriotsBooks.
I've done my share of Google-bashing (mainly due to their inability to move their newer products beyond "beta"), but here's an accomplishment I have to admire: 100,000 servers designed and installed, pretty much glitch-free, in just 2 years! By contrast, my old web presence provider, a reputable and successful outfit, botched a simple expansion involving just a few computers, forcing a lot of customers (including me) to eat their contracts and move on.
There is no reason to be annoyed by people trying to do good things!
Reality is nothing but a collective hunch.
Just think! DC Power Bus... Standard connectors... Good-bye wall-warts.... Mmmmmmmm... Donuts!
The motherboard itself is an outdated concept. It's no longer really necessary if you've dealt with small form factor boards you can easily see that the boards are just a substrate to stick the chips on and for that a flat board-like surface doesn't make sense. What you really need is a cubic cartridge like device that gives you access to more surface area for interfaces close to the memory and CPUs and other chips in a smaller area. It would also facilitate cooling reducing power requirements at the system level.
n/m found it.
t e_paper.pdf
http://services.google.com/blog_resources/PSU_whi
With all due respect, who cares about power consumption? Have you seen the rumored specs of the latest video cards? Don't innovate, just brute force the problem. A rather sad state of affairs. Give me twice the performance with half the transistors and I am SOLD.
/rant (karma burning in effect)
Seriously though, props to Google for sayin what many have probably already said. Its just that Google isn't Joe Sixpack. Maybe people will listen to them.
Microsoft will specify a Vista-compatible power supply that not only uses a single voltage but can radiate power like Star Trek's newest Enterprise does for nearby devices.
Seems like there was some talk a couple of years ago about doing the AC-DC conversion on a massive level, then running individual servers off a server-room wide source. If you create +-12v output on a block of plugs, and +-5v outputs on another set of plugs, you could achieve much better efficiencies. You'd also probably cut your costs significantly. If you put a massive AC-DC transfomer in another area, you could isolate the cooling systems, etc. One large cooling system for a single, large power supply woiuld be more efficient thatn 50,000 individual power supply cooling systems. You might also be able to isolate any out-of phase conditions all those servers would create.
You could also use off the shelf motherboards, and just run longer cables to the wall. THere's likely not enough power running through the wires to account for a significantdrop over a relatively short distance.
However, some of the issue would be a single point of faiturel in the power supply, and one bad motherboard could take down an entire set of plugs nore easily than with AC.
running Ultra-compacitors with only 5-minute charge could last 500 years
I have been saying this for years. We lose 10-20 % of energy charging a battery in a UPS with 117V, we lose another 20-30% in the inverter to get it back to 117V, and then we lose another 10% getting the 117V back to usable voltages for the PC.
It does not take an expert in electrical engineering, just common sense.
Can I sue google for stealing my idea?
A lot of telco equipment is designed to run on -48 volts DC and PC and server power supplies are readily available at this voltage.
The advantage of -48 over 12 volts is that there will be less loss through resistance and smaller conductors can be used. Of course, there is a greater risk of electric shock, but I would think -48 would be pretty safe.
48 volts is also the standard for Power over Ethernet (IEEE 802.3af). This may not be compatible, though, since telcos run -48, not +48, though some equipment can operate with either (though some cannot).
This post is encrypted twice with ROT-13. Documenting or attempting to crack this encryption is illegal.
You can buy a ps that takes 12 volts and turns it into the other voltages. It is tiny and isn't much larger than the power connector on the motherboard. It isn't even that expensive. People commonly use them to make computers that are small enough to fit in whisky bottles or other interesting enclosures. :-) I've been using such a ps for a while and it works well.m l
http://www.silentpcreview.com/article601-page1.ht
Google manufacturered those 450,000 servers themselves. That gives them some expertise. If they had given that many people medical care, I'd listen to their medical advice too.
1) Losses are I^2R. This means that you have more power loss if you transfer power at low voltages through the same wires, connectors etc. You need switchmode power supplies anyway, so may as well switch down from a from a higher voltage.5V means more current than 12V, meaning thicker wires, higher current connectors etc and less headroom in the system for voltage loss.
Engineering is the art of compromise.
Bruce
Bruce Perens.
I really don't think that Google is a bunch of shoeless code jockeys playing with database results and fiddling with power connectors to fry capacitors.
I'm pretty sure that there are intelligent, educated people that get a Google payroll paycheck.
I'm also willing to bet that if some of the intelligent, educated people that are experts in this area that aren't on Google's payroll did receive a Google paycheck for their opinion and dissertation on the matter.
Once could conclude that with a server farm of 450,000, that they probably employ some knowlegeable, educated people that probably know more than your average lay person about power distribution in the enterprise.
Hell, the lead maintenance man of any skycraper could teach you gobs of information on how to cool a room.
At any rate, what would make Google's opinion worth any less merit than anything from Microsoft or Apple with regards to power distribution on a motherboard?
if you steal from one source, that is plagiarism, if you steal from many, well, that's just research.
Statistics. "The Google white paper argues that the opportunity for power savings is immense by deploying the new power supplies in 100 million desktop PCs running eight hours a day, it will be possible to save 40 billion kilowatt-hours over three years, or more than $5 billion at Californias energy rates." ....
Someone must have been aiming for a $5 billion goal to raise eyebrows with this sentence. Either that or google studies only use units like California-powerbill-equivalent-kilowatt-hours/tri ade.
dood stop fukkin my eye
Google's whitepaper is interesting but the fact is that DC in the Datacenter is already happening, and it's not gaining much momentum for multiple reasons.
Google's perspective is rather unique, they use super-cheap desktop systems that individually do not use a lot of power and thus running them off 12v DC might make sense. But in any other, more conventional datacenter, servers have multiple power supplies that can EACH pull 800w of power. Now when you're running 110v AC that means you're pulling ~7 amps through a single cable. You need datacenter grade power cables for this, but it's still sane. Now you can get datacenter equipment that runs 48v DC, but those cables end up running ~15 amps through them, so now you need substantially stronger cable - cable so thick that running it becomes a seriously difficult task due to the guage of the wire!
More likely the direction people are going (and have been for some time) is to 208v AC or 3 phase 220v AC. Now you've just halved the current draw, meaning that your PDUs don't need to be as hefty, your wire doesn't have to be as thick, your coils don't get as hot, etc.
Running 12v DC in any real data center would be ludicrous - the amount of current you'd have to draw through your cables would be way beyond a safe level.
Also AC/DC conversions are cheap these days. And remember, DC can kill you just as easily as AC when your DC Voltage is that low.
"I want to get more into theory, because everything works in theory." -John Cash
Most of the postings so far have it all wrong. Google is not proposing 12VDC into a desktop PC or 12VDC distribution within the data center. What they're proposing is that the only DC voltage distributed around a computer case should be 12VDC. Any other voltages needed would be converted on the board that needed it.
This is called "point of load conversion", and involves small switching regulators near each load. Here's a tutorial on point of load power conversion.
It's been a long time since CPUs ran directly from the +5 supply. There's already point of load conversion on the motherboard near the CPU. Google just wants to make that work off the +12 supply, and get rid of the current +5/-5/+12/-12 output set.
I wouldn't look to MS or even Apple for expert opinions on power distribution.
S
ave the conversion process all together and tell motherboard makers to rectify 120 V directly on the motherboards. It worked for old TVs, it could work again. Maybe.
Take the cheese to sickbay, the doctor should see it as soon as possible - B'Elanna Torres, "Learning Curve"
So now Google thinks it's an expert in Electrical Engineering.
Naaaaaaaaaah! They've just been reading my posts.
KFG
They're not talking about reducing the voltage the PS uses, they're talking about not having the PS produce things like +5 and -5 as well as +12, INSIDE the computer.
Many recent motherboard use 12V to control voltage regulators' MOSFETs gates because the higher the voltage, the lower the internal resistance, so the higher the efficiency. 5V is generally too low to achieve good efficiency, but 12V is fine.
From 12V, the MB can produce 3.3V and 1.xxx Volt for the CPU. It's easy to also provide 5V on the MB.
There's already a small-scale example of this: the PicoPSU. You use an external 12V power brick, and then internally replace your entire computer PSU with something about the size of a matchbox. However, it is only 120W, and a bit short on connectors.
Quattuor res in hoc mundo sanctae sunt: libri, liberi, libertas et liberalitas.
My mountaintop observatory is entirely 12 Volt DC-powered. Unfortunately a lot of otherwise-useful devices run off AC adapters with weird and wonderful voltages, such as 7.5 or 9 Volt.
;-)
It'd be great to know that I can hook anything up without having to kludge mods to avoid frying them. At least I didn't go to 24 or 48 Volts...
Sorry, but assembling servers out of COTS components doesn't make you an expert on computer design any more than working on a car assembly line makes you an automotive engineer. I'll bet the "manufacturing" effort at Google was carried out by relatively low paid workers, not Google "puzzle" solvers.
Though getting rid of all the voltage levels will take more than the motherboard work... you'll also need to do something about all the disks and other components that are currently getting a mixed feed.
The Mac Mini will boot and run on a 12-volt supply. It only takes 17v so that it can provide Firewire power.
-jcr
The only title of honor that a tyrant can grant is "Enemy of the State."
I am sick and tired of all the different kinds of power adapters. I must have about 20 different power adapters e.g. Mobile phone, Digital Camera, Camcorder, Portable Video player, MP3 player etc etc and none of them are interchangeable. I have to stick labels on them to remind me which is which.
Why can't we just have one universal power adapter. You plug it into your device and the device communicates its voltage and power requirements back to the power adapter. The Adapter them supplies to correct voltage etc. Surely with modern electronics and cheap embedded processors, this sort of thing is pretty easy to do.
Not quite on topic but..
It amazes me that so few people realize that a "nominal" 24V is the norm for all manufacturing. Just about EVERY manufacturing plant has 24V throughout the facility, they may (or may not) also have 120/240, but they WILL have 24V - amps and amps of the stuff.
This means there's a full range of 24V equipment, millions of devices. 24V PC's, 24V hubs/switch and all the other infrastructure as well as specialized industrial controllers, etc. etc..
There's some logic behind this that doesn't related to power saving, mainly that you really have to work at doing yourself serious injury with 24V, but you can still pull enough power to run things (like PCs).
Having this low voltage standard is very useful, but before we consider adding another, how about just considering using the one that already exists.
Sco's lawyers might be out of a job soon. I'm sure they'd love to help!
Reality is nothing but a collective hunch.
You compare PCI Express and AGP with USB and PS2, and you're comparing apples and oranges. PS/2 and USB can exist harmoniously; have you ever seen a motherboard with a PCI Express slot and AGP? I've never seen one, though they might exist, they're not common, and certainly not cost-effective. But, what really brings in the adoption of new hardware? Ah, it's the need for it, and the money to buy it.
its a webpage that someone hit print to PDF on and decided to call it a paper.
regardless of the lame format, its point is correct.
Those of us who've been around a while may remember the S-100/IEEE-696 bus, which used this principle. I've still got a few three-terminal regulators left over from the days when I made my own S-100 boards.
Google is proposing 12V between the server's internal power supply and the motherboard. Everything outside the server would still be 208V.
I agree that a standardized 12VDC connector on all electronic devices would be nice, like every other poster here has pointed out, but I don't think that is what google is talking about. You can already get power supplies that take 12VDC in, or even dual 48VDC (telecom standard), and I would be surprised if google isn't using something like that already.
What they are recommending is that the power supply only have 12V out, and all other DC-DC conversions take place on the mother board. Unfortunately, the article didn't go into any detail as to how this would save power, and I don't see how it would make much difference. To me it just seems like you are moving components off the PS and onto the motherboard. Perhaps there is an EE around who could explain it to me.
The logic here is that Apple actually designs and builds computer hardware and Microsoft writes software that powers computer hardware.
Both have a hand in engineering hardware to accomodate their software. The big picture that they miss is the power part as it isn't likely that either camp has the number of servers that Google has. Thousands, maybe yes. Close to half a million servers is a mind blowing amount.
Their[Apple and MS] standpoint is how to make it consumer friendly.
Google's standpoint is how to make it efficient. Then we can make it consumer friendly later.
Can you imagine dealing with passwords on half a million servers? (joking)
if you steal from one source, that is plagiarism, if you steal from many, well, that's just research.
Saving $5e9 is great over three years, but 1e8 machines are going to cost at least $5e10 ($500/machine) - it could be as low as $100/machine, but for most people, probably at least $500/machine, or more. 3% return/year isn't going to wow anyone.
The energy savings are nice, but unless electricity spikes again or electric cars become commonplace (driving electricity costs up again), the upfront cost of replacing that many machines is a big pill to swallow just for the electricity savings.
Their answer is the Petabox. It's a server setup designed to be "shipping-contained friendly", meaning they can build out a container stuffed with these racks, and have it operational on site with connections for power, cooling, and bandwidth. With this design, they can deploy a mirror of the Internet Archive anywhere that's willing to host it, without having to build a machine room or individual racks on site.
Capricorn Tech of San Francisco builds these machines and their site has more info.
--Pat
Oh. So, we have lots of switching power supplies and tantalum capacitors (because we have to supply lots of current at low-voltage) on the MB. Thus moving work from a cheap part of the computer to an expensive part. Not sure I want more power-supply electronics on the MB than is already there.
Bruce Perens.
This might actually be good in emergency cases, as 12v car batteries are readily available as a power source. Most Amateur Radio equipment runs on 12-14v (battery or alternater).
OT, but reading some of the other comments, it would be nice to have an eventual migration to 3-phase for residential distribution. 3-phase induction motors are energy savers. Mexico has a 127/220 system. My line voltage typically runs 127...so why not?
If more equipment was on the market for 120/208...that would be a nice option too. If your too scared to be Europe with the possibility of a 240 to ground short...we can still keep our 120...but add 3-phase as an option.
Isn't -48 still 48Volts of differential, so why not just +48V? I'm no EE, so I'm obviously missing something here.
Finally, a link to the actual freakin' paper. Thanks alot; I wish I had mod points for you.
100,000,000 computers running 8760 (3*365*8) hours over three years saving 40,000,000,000 Kilowatt/hours means that each computer will burn 21.9 kilowatts less, right? Is this realistic?
Rocky Mountain Institute did a charrette (design intensive) on energy efficient data centers a few years ago. Some of the points are a little out of date, but there's still some very good stuff there. They outline a set of techniques to reduce the energy consumption by nearly a factor of 10 while increasing reliability and without impacting performance.
"Research is what I am doing when I don't know what I am doing." -- Wernher von Braun
Is this percentage loss when filling/using a battery applicable only to backup power (0.1% of the time hopefully), or also to normal operation power? This makes a huge difference, and you need to get an electrical engineer to look at a particular UPS to answer this.
Anyone have any links to their patent / white paper?
It's amazing how many people need that pointed out to them. You should print up bumper stickers.
I don't think it has to be that bad. Just have the converter in a power strip along with regular 110 outlets. That way everything doesn't have to convert all at once.
Yeah, this is more like "Hey Motherboard designers, we wanna buy this...you know, us guys who except to buy hundreds of thousands of machines in the next few years".
The cake is a pie
Stnadardising on 12V would make it easier and more efficient to use alternate power sources such as solar as no inverter is required. UPS technology ccould also be simplified (and made more efficient) for the same reason. Of course you would need a 12V in connector on the back of the case.
Rack mounted systems could be made even smaller by removing the individual AC power supplies and having reliable/efficient 12V sources in the rack/building -telecommunications gear has been built like this forever (though it is more likely to be powered at 48V or 24V).
--Murray Barton
that's their answer - a 48v dc feed for telco-quality (high uptime; think NEBS).
12v is too low; I think its a lot easier to lower voltage (if you have to) than to up-convert with dc-dc converters.
higher voltage also means that for the same power, you can have lower current. higher voltage travels better, too.
12v is great for cars, but stoopid for the MAIN power for pc's. google, in all their wisdom, should know better!
--
"It is now safe to switch off your computer."
Uh, who gives a crap who physically built it, an expert designed it. You're right, working on the assembly line doesn't make you an automotive engineer, but those workers are the tools for the engineers that build the car spec. Google gets cred because they've got probably the worldest biggest server deployment. That's no small feat, and they didn't just buy a bunch of dells and a few routers to do it.
gets you one such product at 60W. They have similar products up to 200W. I have 'em. They work great. They have flavors up to 200W.
Help stamp out iliturcy.
The point is that there is precious little left on a modern mainboard which uses one of the PSU voltages directly. Even the CPU is now powered by the 12V rail through a voltage converter right next to the socket. Most of the converters required for the 12V-only PSU are already on the mainboard. The remaining few consumers could have their voltage supplied by one low power converter.
With point of load converters, each of the converters only needs to be designed for its specific purpose. A general purpose PSU typically runs under much less load than it was designed for because the uneven current draw from the various voltage rails means you can't buy exactly the PSU you need. The efficiency of the PSU suffers due to the suboptimal working point.
The average switch-mode PSU is moderately efficient at providing the various different voltages required for the average motherboard, I can't see why moving these electronics to the motherboard would make any great change. In a rack based server installation, I would have thought that the existing 24 or 48 volt power would be a better choice. I should read the paper, but why RTFM?
There was an unknown error in the submission.
Yes, but Google didn't design the engineering spec for their servers either.
Considering the thermal hysteresis of power regulating ICs, I'd like these things as far away from any expensive components as possible i.e. not on a motherboard.
I appreciate that if each 'component' did it's own regulating that the effects are going to be less, but I've lost more equipment due to faulty inbuilt PSUs than idiots wiring in the external PSU the wrong way round.
I also don't want the quality of the PSU circuitry tied to the quality of my chosen components, I like my decent quality PSUs, I like cheap memory, GFX cards and NICs and don't want to have to spend more so they don't blow up.
For the rack maybe, but for my desktop I want as much of the regulation done by external components as possible.
Asking motherboad designers for a lower-power board is one thing, telling them how they should do it is quite another.
Serial remains one of the most manageable approaches to console management. Video is, obviously, not loggable, not automatically monitorable, not greppable, and not amenable to low throughput, high latency remote access.. Serial devices and consequently drivers for them are so simple and straightforward, and the behavior so deterministic, that it is far preferable to something more complex (ethernet and usb) for a console. Ethernet certainly in questionable circumstances may suggest a driver unload/reload as a step to problem resolution, which is safer if not using as a console (though many times I have used ssh and chained the commands using semicolons). For example in that case, if your path contains an nfs mount, and you forget about it as you yank the network out, your chained command will hang as the shell tries to stat the nfs mount for the path. Part of the problem with relying solely upon the ethernet for console is the ethernet has more than one job to do, so it takes a fair amount more competent engineering to get to work right. Many newer systems offer to redirect textual serial traffic over IPMI, and that is admittedly decent *if* the vendor architects it robustly, which is difficult to ensure beyond hands-on experience with a brand and trusting in their consistancy. For example, e326 servers from IBM I wouldn't trust the net console, but an IBM x3455 I would be more confident in. USB, again, has similar complexity issues (it's multiplexed for keyboard/mouse/mass/storage/printing/scanning/etc etc). If you theoretically had bi-directional text console over some usb device, it's more difficult for a low level, simple piece of software to set up the usb controller and all requisite activities, then traverse the bus, identify the console devices, and then use it. Just like with an ethernet device where you may have cause to unload and reload a driver, a usb controller out to lunch with respect to a mass storage device would cause a similar issue. Enterprise distribution kernels tend to compile in the serial console and leave the usb controller modular, specifically with serial consoles in mind.
Serial console servers, in answer to your question, provide a scalable way for systems to access via the network serial consoles. By being dedicated, moderately simple systems with 40+ serial cables, they can provide access (via telnet generally) to a rack's worth of 1U servers, automatically log the content, or at the very least provide an administrator with remote console access at will to any given system.
Serial console is not obsolete in the least bit, just because it can't run your '31337' aero interface, or whatever nice and shiny interface that makes poser administrators and PHBs drool, doesn't mean good, serious systems administrators don't consider the technology to be a vital part of a robust management strategy.
XML is like violence. If it doesn't solve the problem, use more.
They're right that this would save them money.
But for those of us who use a lot less power, the cost of putting switching converts at each point of use would not be cost-effective. We'd pay a lot more for motherboards, and never make it because we never used enough power to make up the cost in savings.
Additional note, this is was S-100 systems did back in the day. It ran into the same problems. Cards cost a lot because each one had to have its own regulator on it.
You also spread the heat generation around, instead of putting it in one spot where a single fan can hit it.
I'm not saying Google is wrong, but I think for regular users it doesn't make sense. Perhaps rack-mounted equipment should go to this standard.
Anyway, Google buys a lot of equipment, they could go to Foxconn and ask for 100,000 of a motherboard that works this way, and they'd barely pay any premium.
So why don't they move ahead and tell us how it works out? Maybe the world will follow.
http://lkml.org/lkml/2005/8/20/95
What google is proposing is supplying only 12 VDC to the mother board instead of the multiple power rails that go to todays power rails. This says nothing about how the supply of 12 VDC gets there or where it comes from. Frankly the idea is a good one though I have to wonder if 12 VDC is the optimal voltage for this.
As has already been mentioned Apple has computers that run fine off of 12 VDC and there are others that also work well with a single power supply rail. This is not a big deal and in fact is good modern engineering. Often it is referred to a point of load regulation and is happening on PC mother boards today! In fact many point of load regulators for the CPU on motherboards are running off 12 VDC. Many other odd voltage loads run off other supply rails. All google is really saying is to get rid of the multiple supply rails and standardize on 12 VDC and allow point of load regulation from there.
I hope every one grasps this now. The discussion of external facilities supplies don't fit in to this discussion at all.
Thanks
Dave
Powered by Google!
"and most machines still come with ps2, serial and parallel ports anyway"..and I'm glad, too, because USB support is not universal, not consistent, and if your machine has the least bit of a hiccup with it you aren't inputting *anything*. No mouse, no keyboard, no nothing, so what can you do then without the old standbys? You have an expensive electronic box that just sits there. I'll hang onto my PS2 and serial and parallel as long as possible. Why? They work, and work well. My one dollar PS2 keyboards work, grab them by the handful from thrift shops. I use my serial modem whenever my ethernet connection goes down, and it does. And my USB printer is a doorstop, it's only worked a couple of times for a few days then there's some ludicrous CUPS or whatever update and it doesn't. If there was a parallel port on the thing I bet I could get it to work, but being "USB only" and not a network printer now it just sits there. I am going back out to look for an old cast iron workhorse parallel connected printer and be done with it because I'm sick of USB, the universal sucky bus.
Go to any forum for any distro and see the horror stories with USB "support" and various peripherals people try. You never have any idea if your new USB gadget is going to work or not. It is beyond a las vegas crap shoot, and when it doesn't work for you you are screwed. Plug and play is not there yet, they claim it is, but in the real world it is not.
Just an old big truck battery or a trolling motor battery. You can get a female plug adapter for them that have 1-4 outlets (5-20 dollas), then you'll be able to plug some stuff in when your grid juice goes south, like in a storm. I keep two truck batteries here in the house just for that purpose, use them all the time in storm season. Power goes out, 60 seconds later I have a 12 volt TV and flourescent light going, a little more than that and an old laptop is up and on the net with the car adapter. You can always add solar later but you can get going with 12 volt stuff now with less than a hundred dollars, that's one good storage batt and the female plugs.
What ever happened to using 12V outlets in server rooms?
By replacing desktops with notebooks --
Of course, to get the maximum benefit, they would need to have cube walls with integrated DC power rails, with a power supply for every half-dozen or so cubicles. An idea for an enterprising cubicle manufacturer.
used linear regulators, which are quite different. Using a linear regulator to go from 12 to 5 volts is extremely inefficient (40% or less). Point of load designs use high efficiency switchers, which waste much less power as heat, but are also more expensive.
It's a bit frustrating to look for exactly what Google is proposing - the article cites a Google document - "High-Efficiency Power Supplies for Home Computers and Servers," which isn't found with a Google search!
"National Security is the chief cause of national insecurity." - Celine's First Law
I'd like less power consumption, too, but...
In an AC power supply, you can generate all of the voltages on one transformer (multiple secondaries) and design a single power supply that generates all necessary voltages and is 90% efficient for the system. If Google is asking for the main power supply to generate 12 volts, and then leave the power conversion for the motherboard, they are actually making the problem worse. You could assume that each power supply on the motherboard is 90% efficient so...
You'd have the efficiency hit at the main 12 volt power supply (90%) and another efficiency hit at the point of load (another 90%), leaving you with a total system efficiency of 81%. How exactly is that better?
Most power supply designers DO actually pay attention to efficiency, and they should continue to push the envelope there. Intel is finally realizing that they need to make the processors more efficient, too. If you've got a 140 watt Pentium 4, you need 155 watts to power it with a 90% efficient power supply. I'd rather attack the big issue (the 140 watts) than worry about the extra 15 watts of power supply inefficiency.
Google's scenario may be OK for a data center, where they have banks of backup batteries that are likely 12 volts, but for the millions of the rest of us it just doesn't help. Google should shut up and stick to what they know - software.
I have wondered about that too, but, like the fact that after 20 years the monitor doesn't shut off with the pc, there has to be some other reason or agenda involved. It's just too obvious to have been an oversight. As for the dc power supplies, why not go with a form-factor equivalent retrofit: http://www.powerstream.com/DC_PC.htm?
I'd like less power consumption, too, but...
In an AC power supply, you can generate all of the voltages on one transformer (multiple secondaries) and design a single power supply that generates all necessary voltages and is 90% efficient for the system. If Google is asking for the main power supply to generate 12 volts, and then leave the power conversion for the motherboard, they are actually making the problem worse. You could assume that each power supply on the motherboard is 90% efficient so...
You'd have the efficiency hit at the main 12 volt power supply (90%) and another efficiency hit at the point of load (another 90%), leaving you with a total system efficiency of 81%. How exactly is that better?
Most power supply designers DO actually pay attention to efficiency, and they should continue to push the envelope there. Intel is finally realizing that they need to make the processors more efficient, too. If you've got a 140 watt Pentium 4, you need 155 watts to power it with a 90% efficient power supply. I'd rather attack the big issue (the 140 watts) than worry about the extra 15 watts of power supply inefficiency.
Google's scenario may be OK for a data center, where they have banks of backup batteries that are likely 12 volts, but for the millions of the rest of us it just doesn't help. Google should shut up and stick to what they know - software.
Is that how UPSs do it? I've read somewhere that some UPSs shunt wall power when available.
omgomgomg
I am supprised that they are not running all their servers off powersupplys that run off 48 vdc. While there is still a power supply involved, its DC to DC. I remember awhile back, there were several PC power supplys on the used market that ran of 48 volts. I was curious and asked. Seems that almost all phone systems have one large battery backup system, in fact, all their equipment it tied into it, where it runs off it all the time. Yes there is a charger that maintains the batteries, and of course it must handle the entire load too, but there is no switch over time on power failure, and this reduces the electronics required too.
I was thinking the same thing. IANAEE, but in my experience I've had power supplies fail far more often than motherboards (even the good antec PSUs). If we offload switching to the mobo won't that lead to higher rates of failure for a FAR more difficult to replace component? I have enought trouble with blown caps and bad (cheap) elecrolytic mixtures as it is.
I wouldn't count on it. When you're buying in quantities as large as Google is there is no small chance that they designed thier own motherboard and case specifically for their purposes. There's no point getting the floppy controller, USB, and any of the other stuff that you normally get (even on server machines) that's totally useless to Google. They probably paid a lot of attention to the power consumption, not only to install the smallest reliable power supply possible, but also to figure out how many they can cram into a rack and how much power/cooling they'll need.
I read the internet for the articles.
Telcos for years have been running equipment off -48v DC power. It makes a UPS unnecessary, have an A and B battery bank, feed it with a set of expandable rectifiers, and hook up the equipment through some breakers. You can even get -48v servers today from several vendors. It's high enough voltage the transmission loss isn't bad, but low enough and DC that it's easy to make efficient DC-DC converters.
Why is reinventing the wheel at 12v better?
12v DC power is pretty much standard in automotive industry, as well as marine. As all of our gadgets try to hang on us as long as possible, why not have a 12 volt standard? There would need to be some sort of box that keeps the power feed to the computer/appliance 'clean' and regulated, but if you're starting at 12 volts, you're halfway there, no?
Probably not worth replying, but what the heck...
For most consumer appliances (phones, dvd players, etc) the same task is being performed - moving electricity. It *is* all a hammer's job. In many cases the only thing stopping you from using one adapter in another setting is the size/shape of the connector, not the quality of the power moving through it. That's why universal adapters can work so well. If manufacturers made different types of nails that could only be hammered with specific types of hammers solely/primarily for the purpose of selling more hammers, wouldn't that "grind your gears"?
creation science book
What's it got to do with Google?
God Be Gone
Google, impose your will on us AFTER you've secured control of our planet.
Check out the cave on the east side of lake Hylia. Strange and wonderful things live in it.
if you have to re-convert anyway, 5V as intermediate voltage is not optimal. When converting to 5V, the voltage drop in the power diodes and in the wires to the mainboard eats a much higher proportion of the power than with 12V as intermediate voltage. 24V or even 48V would be even better.
Telephony has been running on redundant -48V DC supplies to the racks (typically from rooms full of floating storage batteries) since the early relay days. Much modern networking equipment also conforms to this standard, so it can be used in such racks with no local power supply (except the per-card isolation diodes and downconverters).
Power conversion modules running from 48V are in volume production.
Why does Google want to reinvent this wheel?
(However, if they do insist on using 12v, I hope they make it able to work from 11.75v to about 15V, with glitches, and shut off at stable levels below 11.75v. That way such boards could be used directly with 12v renewable energy systems, plugged directly into an automobile "cigarette lighter" power outlet, or easily wired into a vehicle or travel trailer as an appliance.)
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
1) Go straight to the source - make everything run off of 120V AC, like God intended. No conversion inefficiencies!
2) Run everything off direct Solar Power, man (no specifics given - dude, don't bogart that...!)
3) Make everything run off of 1.5v DC - that seems to be the most common battery I have!
* ok I am making it up.
This issue is a bit more complicated than you think.
Google wants 12V motherboards, that take a single input voltage and convert it down, onboard, to the lower voltages they need. What kind of Data Center could we build if those 12V board were out there?
Our input power is 120V or 208V. We convert it down to 48V (the new automotive standard) and we use 48V UPS banks to keep it clean and stable. We distribute 48V power to individual racks. Each rack has a 48V to 12V converter, to feed the motherboard and the rest of the hardware in the rack - the fans, the drives, etc... If we can't find 12V drives, and instead rely on conventional dual-voltage 5V/12V disk drives, we can add a second converter (48V->5V) to each rack.
So, what have we gained? We don't have to distribute any AC power around the room; this reduces EM interference. We can't trust conventional AC power without running it through a live UPS; using DC power distribution guarantees clean power with much less power transformation stages.
DC->DC power conversion is remarkably efficient, especially when stepping down from higher voltages. You might wonder why we don't use 12V for power distribution, and instead go with 48W? Even though this means we need an extra 48V->12V converter in every rack, it reduces the size of the power distribution wiring and reduces power losses.
For a small data center, or even a small server room, going straight 120V->12V makes a lot of sense. We don't need to use an intermediate voltage like 48V, because given the size of the installation, power losses aren't going to matter as much. Let's say that I have only a half-rack of servers, with say, 8 machines in it. I could have a single 120V->12V UPS/Converter unit mounted in my rack that would power all those machines.
At first, the machines would each have a small 12V->5V converter to power drives, and a 12V feed to the motherboard, but as 12V technology evolves, your average 1U server could be built entirely out of 12V components, with each component being responsible for its own power conversion. Keeping the lower voltage conversion as close to the final user of that power reduces transmission losses, resulting in a much more efficient system.
Well, based on my quick analysis of what 12V motherboards would allow, I say that Google has had a *WONDERFUL IDEA*. I, for one, welcome our 12V overlords. All hail Google!
On the other hand, while things don't get adopted quickly, in computers they always get adopted eventually. Mice went from a serial port to a PS/2 port to a USB port. We had ATI then PCI then AGP and now PCI-E. Motherboard designs change with every new chip generation, as does the power supply plug that supports it.
Changing the standards on all of the installed computer userbase computers would be difficult. But if Intel decreed that the reference motherboard for the Pentium 6 took only a 12v rail, the change would be guaranteed.
The ______ Agenda
You plug in a 12v or 18v dc power supply and the motherboard does the rest...
This person is a New York Times writer, yet he doesn't know how to use an apostrophe. Look at the title and throughout the body of the article. I am glad this is brought up by such a worldly newspaper and news hub, but it gives rise to what was being discussed on Rush Limbaugh's program about how the Times has gone downhill over the past 5 years, getting rid of about 200 writers and hiring an official perfume critic, etc... i loled.
In a power supply unit transfomer you can generate other _AC_ output voltages on one transformer (using multiple secondaries), but to get DC outputs you still need those AC/DC rectifiers and regulators to get the normal _DC_ voltages going out of the typical power supply unit. The selection of the ratings on the rectifiers and regulators limit the current/power individually for each voltage coming out of the PSU by the designer/manufacturer.
In contrast, if all the power comes out of the PSU at 12VDC, then each device can draw what they want using a DC/DC converter without worrying about the quantization of X power available @ 5V and Y power available @ 12V. If you aren't drawing on one of the other rails, the power allocated to control that other rail is basically wasted.
Others have address the conversion efficiency elsewhere, suffice to say distributed regulation starting at a higher voltage (e.g. 12VDC to 48VDC) is going to be more efficient if you need tight regulation to low voltages (e.g. 1VDC highly stable) anyhow. This is why automobiles are going to higher voltage batteries. The _DC_ regulation available in todays PSU doesn't really cut it since it's really too far away from the device and generally the wrong voltage anyhow.
You already have one. Take a look at a dual Opteron motherboard. See those toroids and capacitors and big heatsinks near the CPU socket? You're looking at a DC-DC converter, rated for about 80 amps at 1.4 volts. In fact, two of them, one for each CPU chip. They're running off the +12V supply.
PCI slots need +5V or +3.3V. But a 1U server usually doesn't have PCI slots, at least not with anything in them. Despite this, 1U servers and their motherboards normally come with a power supply with outputs for +5V DC, +12V DC, -5V DC, -12V DC, and +3.3V DC. The +12VDC supply is doing almost all the work, powering the CPU and the disk drives. The other outputs are mostly idling. So one can see why Google, which is basically a big collection of 1U servers, is annoyed about having all those useless power supplies in their server farm.
As Animats pointed out, modern mobo's already have power supply components. Modern CPU's are running at ~1.2V and maybe 60 to 80 amps or more - there is no way in hell that you're going to be able to supply that beast with something other than a switcher dedicated to the CPU - and it really doesn't make much difference if that switcher is being fed with 5V or 12V (as long as it is being fed the correct voltage). At 12V, you can get by with 6X less copper than at 5V for the power traces (remember (5/12)**2 ).
A Shadeless room is a brighter room.
Where's the savings really? If you just shift the hardware to down grade 12 to 5v or even 3v to the motherboard, guess what? The computer as a whole is still doing some voltage conversion. What's the difference between doing it in the PSU versus on a mobo??!
Au contraire. The road to hell is paved with good intentions.
If you like, I can name a few thousand examples from history of "people trying to do good things" that ended horribly for all involved.
This isn't one of them, of course, but just because they're trying doesn't mean it's necessarily benign.
Bruce
Bruce Perens.
There are over 6500 Walmart stores with an average size of 120000 square feet. Every 500 sqaure feet they have a 4-tube fluorescent light fixture, drawing 4x40 or 160 watts. Multiplying out, the total square footage is ~6500*120000=780,000,000 square feet. Divide by 500 square feet to get the total number of fixtures, 1,560,000, and multiply that by 160 watts to get the total watts, 249,600,000. Probably 75% of those Walmart stores are 24 hour, while the rest are 12 hour: (.75*24)+(.25*12) = 21 average hours
.068 = $130,096,512 per year in electricity.
Total Watts x Avg hours x 365 days per year = Wh per year
249,600,000W x 21h x 365d = 1,913,184,000,000 Wh per year
Wh/1000 (kWh) x the going rate (approximately 6.8 cents nationwide)
1,913,184,000 x
If they took out one tube per fixture, they would save $32,524,128 per year.
*This doesn't include the parking lots, which have a similar consumption.
So, what's the point? There are other, easier ways to save a lot of power. I'm glad Google wants to change the computer world, but what about replacing 10% of the incandescent bulbs with fluorescents and save 50W x 10,000,000,000? Or just TURN OFF your computer when you aren't using it! Retooling the entire industry would cost MORE than it would save in power. That's not to say I don't agree that we need to start making a lot of little changes and this is as good a place as any. But the benefits are very far in the future, when we run out of oil. Not now.
Cool! Amazing Toys.
There are different kinds of UPSes that do this in different ways. The two major types used for PCs are called "line interactive offline" and "dual conversion online". The first just passes the AC power straight through to the output. If the AC power coming in goes out of range, then it flips a switch internally (relay, contactor, thyristor, etc) to supply the power from an inverter driven by the battery. The second converts the AC coming in to DC all the time, and converts that DC back to AC for output. It then does the switching in DC, or parallels the DC with the battery directly. These variations are classified as "topology" by many manufacturers.
Both of these kinds can have inverters that produce square waves, pseudo-sine waves, or very nice clean sine waves. The dual conversion type can also isolate a poor power factor (the deviation of the current wavefrom from being a sine wave in sync with the voltage) of the PC power supply from the power source. A poor power factor means the product of the average current times the average voltage (apparent power) exceeds the actual real power (average of all the products of the voltage and current and each point in time) being used, which results in reduced efficiency and other problems.
now we need to go OSS in diesel cars
Basically after reading the thread it seems that the electrically cheaper solution would be to have one line of 24 or 48VDC. Knowing that we would need MANY voltage converters in the MB and on the devices would the electrical savings overcome the converters price increase ?
If you must ask the HDD manufacturers to put in their devices a "mini power station" to get the 5V and 3.3V needed from the 24 or 48 what would the price increase be ?
After all google's idea is good but it could be a bad good idea
Another way to get more efficiency is to operate the Switched-mode power supply at the higher voltage it supports, usually 220 to 250 volts. In most of the world this is already done. In North America computers are typically run on 120 volts (in Japan this is 100 volts). In general, these power supplies are more efficient by about 3% or so, on the higher voltage. Of course, be sure to flip the voltage switch if it has one, or otherwise verify that it does support the higher voltage.
For a single computer, it would not be worth adding the extra circuit to get 240 volts. But if you run several, it could be worth doing so, especially if you have so many that it exceeds the capacity of one 120 volt 15 or 20 amp circuit (you could have twice as many on the same amperage if operating at 240 volts). If you already have a circuit dedicated to the computers, that circuit could be converted from 120 volts to 240 volts by changing out the circuit breaker from a one pole to a two pole type, marking the white neutral wire with red or black tape to comply with electrical code identification requirements, attaching these wires to that new breaker (not to the neutral bus), and installing a 240 volt style outlet (NEMA 6-15R or 6-20R). These are the steps that would be used to install an outlet for a big window air conditioner (which you might need anyway with so many computers). Then you can use this power cord.
now we need to go OSS in diesel cars
Ach, shoot, we spent it on USB ports. Never mind.
With a nominal 12v supply that is on-board regulated down to the necesary rails it would be almost trivial to add a battery backup. You only have to support one rail, not two or three. Of course, with a desktop pc that still wouldn't stop the screen going blank during a power outage but for unattended operation no-one's going to be there looking at the screen anyway.
Wal-Mart closes on Christmas (and a few other days/half-das, now, I think). P.S. Do you work for Wal-Mart? I was thinking the Supercenters were more like 500,000 sq ft.
For simplicity, say you have a 120W power supply. Supplied with 110AC works out to be ~1A. Supplied with 12VDC, works out to be ~10A. We'll assume you'll use the same size wire and same length run between the circuit breaker and server in each case.
Using our old friend P=I^2R, where R is essentially equal in both cases, you can seethe transit power loss is ~100 times greater using 12VDC. And it only gets worse when scaled up.
I'm not saying switching to 12VDC is a bad idea, only that a "common sense" is not always valid substitute for engineering analysis.
I am billdar, and I approve this message.
Sue Google?! What are you trying to do, collapse the fabric of space time?
Burn the land and boil the sea, you can't take the sky from me
I'm still right in relation the the GP, but this post explains I was a bit off-topic to TFA.
Sorry :).
I am billdar, and I approve this message.
You're right, working on the assembly line doesn't make you an automotive engineer,... Google gets cred because they've got probably the worldest biggest server deployment.
Automotive line workers also pump out a lot of cars.
Power conversion circuits cost a lot of money.
1 f4340fd47ef50724f0d9820b6abe5efcfcb.e3eTaxmKbNaNe3 8LaNuNa3yLchn0n6jAmljGr5XDqQLvpAe?sc=8&category=13 &it=A&id=417)
$5bn divided by the 2005 estimate of population in California is $138 per person.
$138 which, over 3 years, may mean 2 PC's per person which cost $75 more because of the new power supply (however efficient) type and extra motherboard components and space required to downconvert 12V to 3.3V and 5V as well as all the other non-standard voltages motherboards require these days.
Is anyone gonna save money?
Global warming advocates might jump for joy at the lower energy production requirements but I object to the "it will save money" proposition.
It may even simplify motherboard design IN TIME, but a 12V->3.3/5V conversion for example for ITX boards is about $50 and measures about 5cm square at the best case. It would need a further $20 12V AC->DC converter brick to supply it with the 12V. These converters suck, they waste power. PicoPSU (http://www.mini-box.com/s.nl;jsessionid=ac112b1b
claims 96% efficiency for the conversion, but the brick is far from that.
Making a switch-mode ATX PSU that only throws out 12V would be a brick that doesn't waste power so much, but costs about the same; with a 12V line out. The motherboard then needs all the components from the PicoPSU (still around 5cm x 5cm of board space and $30 of components). Who's saving money here?
but when dozens of servers are only a few feet apart the loss isn't important anymore... but having a separate 300W switching power supply for every system IS a waste of resources. ATX made sense when mobos were $300 a pop... you wanted to break out the fickle power circuitry as it's usually the first to go, now they're $50...the power circuitry is a trivial cost per board. (and like Google says, nobody uses the volts provided anyway so all that circuitry is duplicated in the PS and MOBO) You'll notice laptops all take about 14.4V directly... so it's possible to do this, just nobody wants to rock the market. Google wants that supply external so they can manage the low power directly... at the scale Google purchases custom racks and UPS equipment is much easier to handle.
Go back to class and learn something.
Another benefit is battery backup. UPSes in office locations (my office has about 12, and its a very small office) would benefit greatly if they didn't have to charge a battery from 110VAC and then change that battery power back up to 110VAC to power devices when power goes out.
Again, RV, Marine, and Automotive industries will only help out a 12VDC standard. I'm sure your hyundai isn't equipped to run an SLI gaming computer + wifi gear + hugemungus screen in the backseat, but the base of the technology is already there. Solar is a great addition as well...
Google wants a standard 12V _output_ from the power supply. You're talking about the input power. If Google got its way and motherboards started requiring only 12V, you'd also have an easier time using -48VDC input; your "power supply" would just be one hefty -48V->12V DC/DC converter, which are already available.
I have to wonder why companies like Google whose servers don't have to sit where the headquarters is don't relocate them to a state or region where power is cheap. Supposedly if you are within a certain distance of a nuke plant, power is very cheap. Why not build a data farm there?
I remember a while back reading the the primary google metric for servers was BTUs/ft^2 ! I suppose that a server the expells less heat likely uses less electricity, but it's a relational rather than functional. My guess is every time a new processor comes out its specs gets run through the equations and they figure out what the optimum number of chips to swap out are.
Apocalypse Cancelled, Sorry, No Ticket Refunds
The whole idea of switching to a single 12v system is so that there is less power loss (and less heat generated) converting the power inside the power supply. Running everything at 5v would be worse than things are already because there's about twice the power loss on a 5v rail as a 12v one. Modern motherboards (especially high end overclock boards) are more demanding on the 12v rails than anything else even though the cpus usually demand 100 watts @ under 2 volts. I'd have to say that what google is suggesting is already being implemented for reliability.
"And we have seen and do testify that the Father sent the Son to be the Savior of the World"
1 John 4:14
You'd lose your bet.
Selling ads is quite different from designing hardware. Google folks are not the first to entertain this idea. 12V -> 5V converters on the motherboard will either be massive or will generate quite a bit of heat. Also, in case one of PSU components breaks, you'll have to replace the motherboard. If you want to standardize, make it TTL 5V, or CMOS 3.3V standard and power your 12V-hungry videocards separately. Now, google, please quit coming up with idiotic ideas and get back to sucking $$ from ads.
Then we could use any 12 volt supply and for go there costly supply.
However, if you want to maximize efficiency, you are going to pick a motherboard input voltage of 42-48VDC, with large power converters at the rack (or 3-rack) level.
I fail to understand the logic of 12V for Google (although it would be nice for cars); their typical racks look like they run at about 6kW, so centralizing the power converter would result in a 500A DC bus at 12V. (250A if it is mounted at the center of the rack like their ethernet hubs).
The industry might move closer to a high-voltage DC (as in 380-480VDC or even 900VDC) over the next few years, and looking towards a 12V standard won't help matters.
Ideal voltage is all a function of what the maximum load is...
A 12V standard doesn't hurt high-voltage DC; you're not going to have that on the motherboard anyway. High voltage DC just means you remove the line side rectifier from your switching power supply; if that power supply now needs to put out +12V (only) instead of many different voltages, so much the better.
I'm not sure why not 48V to the motherboard. Probably because pretty much nothing in the PC world uses 48V now. Even computers at central offices convert from 48V down.
It's actually a direct correlation. Essentially 100% of the electricity which a computer uses gets converted to heat. A small amount goes into noise, a little into light (LEDs and fiber connections), a little into electricity (electrical network connections), some into RF. But nearly all goes into heat. And even most of the fraction which goes into other things ends up as heat within the server room.
I should have known better. I read the whole thing and it said nothing about what kind of employees actualy built the servers and how much they were paid.
My concern is that if you go with 12V boards, you will end up needing a 380/580VDC:12VDC or 48:12V power converter for each computer, which complicates adoption of a DC standard.
If you go for the extra conversion step, you increase your losses by a percentage point at least, and there are reliability concerns (from what I am told) with the "HVDC" to "LVDC" equipment.
The best solution is to pick a voltage that can be used directly by the majority of the loads. Since 5 or 3.5 or 1.8V isn't practical, you have to evaluate the benefits of a higher voltage at a level beyond the individual computer.
I know Google's objective is to impact residential computers more than data center equipment, but history says that the best solution is one that can take advantage of both areas.
At least one engineer who thinks changing standards and calling power supplies inefficient is absurd: http://www.edn.com/blog/1700000170/post/1660004766 .html