• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

How much power do these things use?

S

Sly

Supreme [H]ardness
Joined
Aug 17, 2004
Messages
5,749
I need something to act as my 24/7 home server (FTP/HTTP/RDP/IIS/MYSQL/ETC). My last server was an old netbook and it was pretty adequate, but it got relegated so i'm looking for a replacement.

Anyway, one of the local trading websites has this posted (The guy lives five miles from my house)

intel atom d945gclf (from intel newly replaced)
intel atom 1.6ghz ht technology
1gig ddr2 667
40gig seagate or western digital hdd
zeal ion mini itx case with power supply

I figure i can replace the harddrive with something bigger.

Any idea on how much power it'll be using? From my computations based on how long it took to drain the battery, the netbook was using 12w under load. My current netbook (A dual core) is using 18w under load. How does the nettop i listed compare? This is a 24/7 unit so power does add up.

I'm still on the fence with this, coz a new harddrive will bring the price up to a new one anyway.
 
What case do you have? Is it a m-atx?oh sorry there it is. That thingy won't be even using a 100w
 
Yeah, the thing that amazed me when i calculated the power usage of the netbook is that leaving it running 24/7 is like leaving a lightbulb on all day. It's a far cry from my fully loaded main rig.

I'm curious if a nettop can run with the same efficiency as the netbook (under 18 watts).

D510MO (Dual Core Atom) = 13w
WD 2.5" SATA harddrive (Pulled off a burned out laptop) = 2.5w (converted from 5v---0.55A on the sticker)

That's 15.5w total. Will i expect a generic 500w PSU to pull 15.5w off the wall? Or do i need to compensate for inneficiency? The PSU is waaay overkill, but it's what comes with the cheap casing.
 
Um.......check out the data on the Miller MIG welder joke at PcPer.
 
Spooony said:
Power supplies only start to be efficient running 50 percent of their load. But there's so much power features on them that's last thing you have to worry about. 500w is one huge overkill
Click to expand...

Does that mean i should see if i can find a PSU with a lower power rating? There's another guy selling 250w PSUs. I really don't mind the legwork for this experiment just to see how low i can make the powerdraw and keep the price reasonable (expensive components negates the power savings.)

My biggest concern at this point is the possibility of the PSU drawing 50w to power a 15w device. Which kinda defeats the purpose of building a power efficient PC.
 
Spooony said:
Power supplies only start to be efficient running 50 percent of their load.
Click to expand...
No. They generally reach peak efficiency somewhere from 50-70% load, but they can still be efficient at other load levels. Most 80Plus-rated PSUs have good efficiency at 20% load and decent efficiency at 10% load, relative to their level of 80Plus certification.
 
Zero82z said:
No. They generally reach peak efficiency somewhere from 50-70% load, but they can still be efficient at other load levels. Most 80Plus-rated PSUs have good efficiency at 20% load and decent efficiency at 10% load, relative to their level of 80Plus certification.
Click to expand...

What is decent? How efficient would it be for a 500w PSU to supply 15w? That's about 3% capacity. That would mean it would draw 23w to supply 15w? Or would it use more power?

The 80+ PSU's are on the expensive side, and i'd be hitting deminishing returns if were to use them just to shave 3w. How long does it take to accumulate $100.00 worth of 3w?
 
Good power supplies start at the 400 watt range with the CX430 and the Antec EA380. The best low-powered one is the Seasonic X400, but I really doubt the greater efficiency is worth a $130 increase over the $20-25 Corsair CX430.

Decent is usually at least 70% efficiency. It really comes down to the design, and not many people (translation, hardly anyone) tests efficiency down to that low, since power supplies are usually powering at least 50 watts at idle.

How long it takes to recover is dependent on your electricity rates. The average rate for the US as of January 2011 is 9.34 cents per kilowatt hour. 3 watts is equal to 72 watt-hour per day, assuming it is left on all day. That is equal to 26.280 kilowatt hour per year, which is 245.45 cents, or 2.46 dollars. It would take you roughly 40.7 years to recuperate the cost of a higher-end power supply, by which time your power supply and your components will most likely be long dead, and you would have to buy new components. The highest rate in the US is about 24 cents, which is roughly 2.5 times the national average. At this rate, it will still take you more than 16 years to recuperate costs, which also means your components are likely to be dead by then.
 
Sly said:
What is decent? How efficient would it be for a 500w PSU to supply 15w? That's about 3% capacity. That would mean it would draw 23w to supply 15w? Or would it use more power?

The 80+ PSU's are on the expensive side, and i'd be hitting deminishing returns if were to use them just to shave 3w. How long does it take to accumulate $100.00 worth of 3w?
Click to expand...
Decent meaning above 80%, and probably more depending on the design and the level of 80Plus certification the PSU model is targeted at. An 80Plus Standard 500W PSU would probably be about 70% efficient at a load of 15W, but in that case, the amount of wasted power would only be about 6W, which is so little that it's basically insignificant.

80Plus PSUs can be found for very low prices nowadays. Keep in mind that you need to spend above a certain threshold in order to get a unit of decent quality to begin with. For example, you won't find a good 650W PSU for $10, regardless of whether or not it claims 80Plus certification.

The savings for a 3W reduction in power costs depend on your power rates. If the device runs 24/7, that would equate to about 26 kWh every year. Assuming your rate is somewhere in the area of 10c/kWh, that would be a savings of no more than $3 every year.
 
Yeah, just putting in an atom instantly pushes you against the wall of deminishing returns.

Anyway, i've put together the rig (D510MO, 2Gig DDR2, miniATX/500w case, . First thing i noticed is that it's actually slower than my u230 netbook, and did some benchies to confirm.

d510vsu230.jpg


One thing about the nettop is that it's running for a few hours now and it still feels cold. I can't even say the PSU exhaust is warm.

I wonder if i can even get away with disonnecting the case fans and just rely on the PSU to pull the air out. I'm also checking around if there's a neighbor with a watt meter i can borrow so i can compare it to the netbook.
 
Zero82z said:
No. They generally reach peak efficiency somewhere from 50-70% load, but they can still be efficient at other load levels. Most 80Plus-rated PSUs have good efficiency at 20% load and decent efficiency at 10% load, relative to their level of 80Plus certification.
Click to expand...

if you really want to get on the dot technical then tell me this. When will your pc run at those percentages? It will run at 10 percent load some time at 30 some time at 80 some time at 45. So you can't really use those percentages. And 80 plus don't mean much Cmon lol you know that of all people it gets done at 23 degrees
 
Spooony said:
if you really want to get on the dot technical then tell me this. When will your pc run at those percentages? It will run at 10 percent load some time at 30 some time at 80 some time at 45. So you can't really use those percentages. And 80 plus don't mean much Cmon lol you know that of all people it gets done at 23 degrees
Click to expand...
Your PC will run at those percentages depending on the specific components you're using, the load you put on it, and the rated power output of the PSU you choose. Depending on those factors, it is certainly possible to end up with a situation where a PSU is running at around 50% load most of the time and is therefore achieving peak efficiency.

80Plus conditions are not always representative of real-world conditions, but the fact is that most PSU manufacturers tune their PSUs to perform well in 80Plus tests, and as a result, that performance carries over to the same load levels even at different operating temperatures.
Spooony said:
so miller makes welding machines? and what do you think those welding machines are powered with ?? maybe some AA batteries?
Click to expand...
They're not powered by computer power supplies, and the review is still a joke, which is pretty obvious from the date it was posted (April 1st, 2007). The "Miller XMT 300 PC" is not a real product, and the pictures of the "power distribution module" are taken from this review of the Silverstone Zeus 850W: http://www.pcper.com/reviews/Cases-...ZF-850W-Power-Supply-Review/Closer-Look-Contd
 
Zero82z said:
Your PC will run at those percentages depending on the specific components you're using, the load you put on it, and the rated power output of the PSU you choose. Depending on those factors, it is certainly possible to end up with a situation where a PSU is running at around 50% load most of the time and is therefore achieving peak efficiency.

80Plus conditions are not always representative of real-world conditions, but the fact is that most PSU manufacturers tune their PSUs to perform well in 80Plus tests, and as a result, that performance carries over to the same load levels even at different operating temperatures.

They're not powered by computer power supplies, and the review is still a joke, which is pretty obvious from the date it was posted (April 1st, 2007). The "Miller XMT 300 PC" is not a real product, and the pictures of the "power distribution module" are taken from this review of the Silverstone Zeus 850W: http://www.pcper.com/reviews/Cases-...ZF-850W-Power-Supply-Review/Closer-Look-Contd
Click to expand...

I said psu. Psu means power supply unit. Didn't say pc power supply
 
Spooony said:
I said psu. Psu means power supply unit. Didn't say pc power supply
Click to expand...
I think you're just embarrassed about the fact that you weren't intelligent enough to realize it was an April Fool's joke and not a real review.
 
An Antec 380 AT $40 (or similar at less cost unless you want to spend more on SFF or less efficient.
IF 15 watt = 70% draw is 21.4 or less, 75%=20 watt, 80%= 18.5 watt 85%= 17.6 watt.
Worst case additional cost is 7 watts 24/7. Don't know your rate
Mini-box.com has a pico 90watt 12volt +adapter (atx connector, p4. one 4 pin) 96% efficient [ the 12 volt I believe] for $30 and shipping.
1.6 might be stopper
 
The server won't be online until next week, but i've done a few test runs (i.e. leaving it running with some light loads, remote access through firewall, etc.) and i noticed that while it runs cold, if i turn it off for a few hours, the PSU starts getting warm.

Is this normal?

I never noticed that happen to my main rig. But then again, the main rig is a full sized tower. The Atom rig is a mini ATX with a compact PSU.
 
PSUs do tend to get warm after you run them for a while, yes. Especially less efficient ones.
 
Sly said:
The server won't be online until next week, but i've done a few test runs (i.e. leaving it running with some light loads, remote access through firewall, etc.) and i noticed that while it runs cold, if i turn it off for a few hours, the PSU starts getting warm.

Is this normal?
Click to expand...
Yes, because airflow prevents the temperature from rising much.

You need to look at your power bill and see the billing structure and what billing tier you're on. PG&E charges as much as 40c/kWh for major power hogs to discourage conspicuous consumption.

At the national average of 12c/kWh, a computer that uses 100W around the clock would only rack up $104 annually. If you have many servers running, then it would help to vent the computer exhaust outside and crack a window to reduce load on A/C.
 
My previous rigs get colder after i turn them off. The atom rig is warmer five hours after i turned it off than when it was running. Is the motherboard still drawing power even when turned off? I've never noticed it happen to my main rig, but then again, it does have more mass.

I'm really gonna have to find a watt meter to settle this, i won't have the budget for it 'til the 15th, so i'm just doing guess work right now.
 
The motherboard should not be drawing any significant amount of power while off. If the PSU is still loaded when the PC is off, something is wrong.
 
Finally managed to find a wattmeter.

The desktop when off uses 1w, when hibernating it's still only 1.5w (probably to power WOL).

The Atom uses 3.3w fully powered down.

Not as efficient as the main rig, but not high enough to warrant buying another PSU.

Anyway, since the point of the thread was Nettop vs Netbook. Here're the results after a couple of hours of testing.

Atom D510MO (Accessed via RDP)
Off 3w
idle 27w
shutdown 3w
Downloading 27w
Streaming Videos 27w
Format Factory (CPU 100%) 30w

u230
startup 40w
idle screen off 18w
idle screen up 21w
idle screen up + hub (kb/m) 22w
idle screen up + hub (kb/m/bluetooh) 25-32w
streaming + screen up + hub (kb/m) 26w
streaming + screen down 19w
All Peripherals + streaming + 1080p DXVA + Programming 37-47w
All Peripherals + streaming + 1080p DXVA + Programming + 22"LCD 72w

Something to note tho. The u230 uses a desktop CPU and chipset. I think it was meant for a nettop and AMD shoehorned it into a netbook, even today, this is still the most powerful netbook out there. When i get my hands on the AAO (A real 1.6ghz netbook), i'll measure again.
 
Last edited:
You must log in or register to reply here.
Back
Top