Are there ways to check wattage?

Starrfoxx

Weaksauce
Joined
Apr 27, 2007
Messages
67
A couple of questions:

1) Where can I find a device that will test my power supply and tell me how many watts it's putting out? I have a 500 watt PSU, but want to make sure it's still working at full compacity.

2) Is there a program that tests how many watts your computer is actually using?
 
A couple of questions:

1) Where can I find a device that will test my power supply and tell me how many watts it's putting out? I have a 500 watt PSU, but want to make sure it's still working at full compacity.

2) Is there a program that tests how many watts your computer is actually using?

1) A power supply only delivers as much power as the components require of it.

2) Not accurately
 
Although I've learned that the Kill-A-Watts can be pretty inaccurate, especially with certain APFC power supplies?

But, yes, you don't really want your power supply putting out 500W unless your components are drawing that much (and even then, running a PSU at 100% isn't likely to be healthy for it). Your system definitely won't have that kind of power draw, though.
 
That will give you an inaccurate AC number but won't give you an accurate DC number which is the power supplies output.

The margin of inaccuracy isn't really all that large though, except for startup draw, which lasts for what, a couple seconds?
 
The margin of inaccuracy isn't really all that large though, except for startup draw, which lasts for what, a couple seconds?

*sigh* People never listen.


100_1218.jpg


229w DC and 220w AC. Yeah go Kill-A-Watt go!

100_2710.jpg


219w DC and 215w AC. Yeah go Kill-A-Go I am making energy here baby!

Or you could look at the correct numbers on the current transducer next to it 2.94A *120v=352.8w or 62% efficient as opposed to the Kill-A-Watt's 102% efficiency.
 
Listen to what? I've never seen a comprehensive review of the product, because frankly I don't care.

25-45% disparity is pretty funny though.
 
Listen to what? I've never seen a comprehensive review of the product, because frankly I don't care.

25-45% disparity is pretty funny though.

The accuracy of a kill-a-watt has been gone over here....a lot and given what I do I may know what I am talking about. ;)
 
The accuracy of a kill-a-watt has been gone over here....a lot and given what I do I may know what I am talking about. ;)

Well, I read this subforum about bi-weekly, and I've yet to see it discussed in depth... But saying people don't listen, without knowing the extent to their understanding of the issue at hand, is just a bit unfair, I think.

And I never said you didn't know what you were talking about. :confused: I KNOW you know what you're talking about, which is why I come here to listen to (usually) you're advice, and your's alone.
 
I often see the Kill-A-Watt kicked to the ground and then kicked some more. But I don't think I have seen many suggestions, if any, on other equipment to consider. I am sure the equipment in the above pictures are probably beyond most average users (perhaps in cost, at least).

Just sayin'.
 
And I never said you didn't know what you were talking about. :confused: I KNOW you know what you're talking about, which is why I come here to listen to (usually) your advice, and your's alone.

While Paul is great, I believe that Jonny, Matt (madmat), and Oklahoma Wolf among others might have an inkling of how things work too. ;)
 
So, I understand that a Kill-a-watt is inacurate now. But the meter pictured is >$150 (if bought from the makers site). Can you recommend a power meter that will tell the outputs that the Kill-a-watt says it will do, with a cost closer to a kill-a-watt? An alternative to the innacurate Kill-a-watt for close to the same cost.
 
On the consumer level, there really isn't any options to the Kill-A-Watt and other devices like it. And I can't see somone paying hundreds of dollars for equipment just to measure power consumption unless the review power supplies or do something for a living that would require that.

Some people say adjust for efficiency (80%) and then consider keeping the load within 50 to 75% of whatever PSU's capability. So for example, 300W from the wall, x 80%, 320 WDC... 500W to 550W PSU. But to take into consideration inaccuracy of the Kill-A-Watt and immeasurable transients, I would just suggest double whatever the Kill-A-Watt is reading as a PSU recommendation. So 300W from the wall, 600W PSU. Just to be "safe". You're still not suggesting 1600W PSU's for people with 8800 GT cards (one side of the spectrum) but your not telling people to buy 380W PSU's that will die in a year's time due to stress (other side of the spectrum.)
 
The accuracy of a kill-a-watt has been gone over here....a lot and given what I do I may know what I am talking about. ;)
Does the killawatt show the delta between idle and load numbers with a higher level of accuracy? or does the changing efficiency between the idle and load numbers effect things too much?
 
At different loads you get different accuracies, so idle to load readings are going to be effected.
 
So, I understand that a Kill-a-watt is inacurate now. But the meter pictured is >$150 (if bought from the makers site). Can you recommend a power meter that will tell the outputs that the Kill-a-watt says it will do, with a cost closer to a kill-a-watt? An alternative to the innacurate Kill-a-watt for close to the same cost.

Like jonnyGURU said you are paying $20 you really aren't going to find anything near that. The Seasonic Power Angel was built using the same IC the Kill-a-watt uses IIRC, the Watts Pro uses a different one I believe but costs ~$125 and I ahve not used it to know how well it does. The one in the picture is quite a bit more than $150 as is the Extech I also use. The other issue that that picture doesn't stell is taht Kill-A-Watt's are very sensitive to having anything go wrong on the AC side of a power supply, but again it is $20.
 
I don't expect a Kill-A-Watt to be nearly as accurate as a good meter, but I don't understand how Kill-A-Watt can make grossly wrong watt measurements when it does consider power factor and doesn't merely report VA (although it does have a function for that).
 
I don't understand how Kill-A-Watt can be considered inaccurate with AC or APFC PSUs because it can measure true watts, meaning it compensates for phase angle, and doesn't measure just VA (but it can measure VA, too).

Well I would certainly consider it innaccurate given it is telling us the AC draw is lower than the DC output, unless of course power supplies really are greater than 100% efficient?

;)
 
this does not jive with other reports of the KAW being pretty accurate at the wattages in question...

http://www.jonnyguru.com/forums/showthread.php?t=2695

http://www.consumerreports.org/cro/...ergy-monitors/overview/energy-monitors-ov.htm



what is that variac transormer doing... could you map out what exactly is going on in this test?... why not use AC directly from the wall... It could be that the AC waveform is not pure, altered by some lab equipment like the transformer, and therefor the KAW will not yield a proper measurement since its designed for perfect sine waveforms from the power company.
 
Last edited:
From someone who also reviews power supplies, these are my thoughts.

The KAW is more or less a toy. It's reasonably accurate enough to give you a ballpark figure most of the time with most PSUs on the market, but it will not give you absolute accuracy. And when I say "ballpark," I mean a big one - I have two of these things, and they don't even agree with each other let alone my Brand meter. I once tested mine with a 60W light bulb. KAW 1 gave me 54W, KAW 2 gave me 61W. The difference between the two gets exponentially worse the higher the power draw. I don't use #1 for anything - it's too far off. I sometimes use #2 to double check the Brand's numbers, but the slow sampling rate on it is frustrating to deal with. These things are way too slow on the uptake to give you any kind of accurate transient readings.

For any kind of accurate measurements, you need to spend some cash... even my Brand 4-1850, which at $150 is a significantly better instrument than a KAW, can still be fooled by some APFC units. What I'd like to do is grab a $500 Extech meter, but to be honest I don't get enough of those oddball meter fooling units to make the investment pay off.
 
There it is.

Lab equipment is expensive. Tight specs are expensive. ATX rails are only speced to 10% (havent checked recently) point is if your $200 power supply can provide 10.8 to 13.2 V on the 12V and still be in spec it seems a bit much to expect a $25 meter to be more accurate.

A current shunt and good DVM that handles true RMS is likely the least expensive fairly accurate way to input measure current and caluclate wattage. Output from the power supply is another ballgame due to the precision high power loads required. (not news to the pros in this thread who have such equipment).

A google of "ac current measurement techniques" provides a world of information.
 
ATX rails are only speced to 10% (havent checked recently) point is if your $200 power supply can provide 10.8 to 13.2 V on the 12V and still be in spec it seems a bit much to expect a $25 meter to be more accurate.

It is supposed to be 5%, though some people like to use a creative reading and give a peak value of 10% at a peak load value.
 
this does not jive with other reports of the KAW being pretty accurate at the wattages in question...

http://www.jonnyguru.com/forums/showthread.php?t=2695

http://www.consumerreports.org/cro/...ergy-monitors/overview/energy-monitors-ov.htm



what is that variac transormer doing... could you map out what exactly is going on in this test?... why not use AC directly from the wall... It could be that the AC waveform is not pure, altered by some lab equipment like the transformer, and therefor the KAW will not yield a proper measurement since its designed for perfect sine waveforms from the power company.

The AC waveform through the variac and power conditioner is closer to a perfect sine wave than from the wall.

Hate to tell you but your power coming from the wall is only very rarely a perfect sine wave. And consumer reports tested a grand total of 1 sample. 1 sample. I have seen literally hundreds of SMPS so I would venture to guess that my sample size and sampling is a bit more thorough. They also don't tell what "accurate" is. Is 6 watts out of 100 accurate?

Like Oklahoma_wolf said they also sample ridiciulously slowly. The entire transient test I do is missed by it. That is over 100W that it misses during that test.
 
Back
Top