Help with system power usage

pepebuho

Distinguished
Jun 8, 2009
8
0
18,510
Hi
I put together the following system
CPU: i5 4570
Mobo: Gigabyte GA-H87M-HD3
memory: 8GB DDR3
PSU: Themaltake Smart SP 750M
Case: Cooler Master K282
Cooler: CM 282+
3 HDD 7200 rpm
1 DVD Writer.
1 USB card reader.

Accordng to several power consumption webpages, this system was supposed to consume about 300 to 400W
After assembling the system, and connecting it to a Kill-A-Watt, the max power drawn is 100W aprox (Prime95).

I feel kind of disapointed, because If this is true, then I could have spent less on the PSU. In fact I had been fretting about possibly having to purchase a 1000VA or better UPS, but if this is how much the system draws, a 650VA (450W) UPS should be more than enough even leaving space for future expansion.

Am I overlooking anything? Is this the way it is? I can see onscreen how the cpu temp increases when Istart prime95. Certainly I was expecting something more than 200W. I am fascinated with 100W (saving nergy), but I wonder if I am arm-binding my system.

Thanks
 
Solution
HI,

On the UPS: I currently use 4 different UPSs in front of a 1/2 dozen computers. The VA rating of a UPS is only moderately useful in choosing a UPS.

The decisions are:
1. Does the UPS have enough power to handle the PC? This is the VA numbers. Note as you get to higher VA numbers you get ugly things like FAN noise. My 1000 VA APC ups is totally silent. The same model APC with a 1500 VA rating is nasty noisy.
2. Does the UPS switch fast enough to keep your system running? I have one UPS that protected a gaming PC that could not switch fast enough under a heavy gaming load. Guess how I found that out. The power dropped, the PC dropped. With power still off the PC came back up nicely on the UPS. YOUR oversize PSU will help...
Hi
You have not wasted your money but have absolutely done the right thing.
The kill-a-watt meter takes samples at intervals and displays an average result,thus ignoring any rapid peaks.
If you had fitted a lower power PSU then it would not have been able to handle these peaks correctly and would have got hot and also would have had a reduced lifespan.
And yes the 650VA UPS will be more than enough for your system.
 
HI,

On the UPS: I currently use 4 different UPSs in front of a 1/2 dozen computers. The VA rating of a UPS is only moderately useful in choosing a UPS.

The decisions are:
1. Does the UPS have enough power to handle the PC? This is the VA numbers. Note as you get to higher VA numbers you get ugly things like FAN noise. My 1000 VA APC ups is totally silent. The same model APC with a 1500 VA rating is nasty noisy.
2. Does the UPS switch fast enough to keep your system running? I have one UPS that protected a gaming PC that could not switch fast enough under a heavy gaming load. Guess how I found that out. The power dropped, the PC dropped. With power still off the PC came back up nicely on the UPS. YOUR oversize PSU will help here. The 'hold up time' where the PSU keeps supplying power after the wall power is lost is specified for full load, you are running a small fraction load so your PSU has much longer than expected hold up time, and will give the UPS more time to switch over and get stable load.
3. How long do you want the UPS to run? A few seconds to bridge over a power glitch? Long enough to shut down? Long enough to work for 20 mins? Finding battery run times takes work, the easiest way I know to compare two UPS runtime is to look at the replacement battery and see how many amp-hours. Most large UPS take a standard size battery (7-9 amp-hr) and then use multiple (one, two, four) batteries inside the case.
4.Do you like the software used to automatically shutdown the PC? I don't use any. Most people who have UPS do. Instead I size the UPS to last 20+mins and assume that the excitement of losing lights, etc will get people to shut down their systems.
5.Do you like LCD displays on the front, etc ?

On your system power consumption. That's the right range for that PC. No problem there. The Max TDP for your CPU is 84 W. The MB etc using 15w is about right to get you to max 100w. At 100W max your kill-a-watt would have read about 120w % 80 PSU efficiency. Even my older power hungry i7-920 X58 chipset based PC with HD7850 video is only pulling 113W at the wall as I type including the 24" monitor (per the display on my UPS which read the same as my kill-a-watt).

On the PSU size, yes companies dramatically overestimate power supply requirements, mostly because cheap PSUs are crap and do not actually hit their ratings, are rated at really low temps, and may start producing dirty power under load. Its safer for NVIDIA and AMD to say 'min 450w psu' for a 75 w video card then to say '220w will do, but don't buy a crap PSU'. Aside: your PSU is not crap: http://www.hardwaresecrets.com/article/Thermaltake-Smart-730-W-Power-Supply-Review/1497/1
 
Solution


Hi, I agree the money spent on a good PSU was not wasted. The cheapest good PSU would only have cost $50 less that the one he got so no big deal. On your other point I've not found that kill-a-watt numbers need to be much lower than PSU number to survive "average result,thus ignoring any rapid peaks". Ex, my gaming PC peaked around 330W (kill-a-watt measured at the wall) on a 460W PSU. No problems I can see.