Kill-A-Watt accuracy test

[Kvar]

There are users raising questions about the accuracy of the Kill-A-Watt meter above 700W. I don't see much of a problem, even near 1500W.

Let's see what I got:
Test method:
I compared the values between reference and the Kill-A-Watt using load configuration that is representative of modern power supplies. The meter are daisy chained together so they both read values at the sametime. A lot of high power PSUs have an active PFC with PF >0.98 and behaves very close to pure resistive load, so resistive loads were used to simulate a loaded active PFC PSU.

Reference:
Valhalla Scientific 2101 Digital Power Analyzer (which isn't perfect, but for this comparison, we'll assume it is. )

A. 173W 270VA, non PFC loads only(non PFC PC + LCD monitor)
B. 648W 676VA, non PFC + resistive mix (that PC + resistive dummy load)
C. 1065W add more resistive load
D. 1420W add more resistive load

KAW:
A. 173W 269VA +0.0%
B. 656W 686VA +1.2%
C. 1080-1100W jumpy +1.4 to 3.2%
D. 1435-1460W jumpy +1.0 to 2.8%
There will be variation between units, but I find +/- 10% adequate for PC power supply test and reviews. If you are loading up your PSU so much that +/- 10% makes a life and death difference, you're not giving it enough safety margin. It's only $20 and the measurements are far better than you can guesstimate.

It has a problem with locking up when inductive loads are switched on and off, but it can be revived by power cycling.


How power meters work:
Both the reference meter and the KAW have a front end like this:

(image from Valhalla Scientific page)

The current simply goes through a strip of metal with a very small, but a constant resistance that generates xxmV/A. Each strip is slightly different and at factory, they simply apply 10.00A through the strip, then program the proper multiplication value (calibration factor), so it reads 10.00A. Ideally, the mV/A is linear throughout the range, but actual material isn't. Since the shunt works on principle of resistance, it generates heat with current and the heat generated quadruples when the current doubles. This explains why it reads slightly higher at higher load.

Shunt should be heatsinked for better accuracy, but Kill-A-Watt is a cheapo, so it doesn't have one. You could add one to improve the accuracy under heavy load though.

That arch thing on uppper left is the shunt in KAW.

(photo from http://cafeelectric.com/killawatt/ )

Voltage channel is calibrated the same way. Kill-A-Watt uses 110V. Apply 110.0V, push buttons and it will adjust the multiplication factor to read 110.0v If you have a programmable current/voltage source, you can try to calibrate it, but if you fry it, i'm not responsible.
http://www.pupman.com/listarchives/2.../msg00034.html


The Valhalla meter uses a bunch of analog components and an analog integrator to get a value.

Kill-A-Watt measures both the amp and volts 2048 many many times a second and mathematically integrate over a period of one second or so(which is why it updates about every second)

[Spectre]

Alright, I still pose the same question since you never answered last time but beat around the bush instead. Why does the Kill-A-Watt read a lower AC input than the DC output of some power supplies? I am not talking about a few watts here....I am talking about being off by 100 to hundreds of watts? IE I know the DC load is 900 watts on a PSU that is ~82% efficient yet it is givng an AC draw of 880w. By all means it should be reading ~1100w AC so it is off by ~220w.

[Kvar]

Originally Posted by Spectre

Alright, I still pose the same question since you never answered last time but beat around the bush instead. Why does the Kill-A-Watt read a lower AC input than the DC output of some power supplies? I am not talking about a few watts here....I am talking about being off by 100 to hundreds of watts? IE I know the DC load is 900 watts on a PSU that is ~82% efficient yet it is givng an AC draw of 880w. By all means it should be reading ~1100w AC so it is off by ~220w.

I wasn't quite sure what your question was. You stated that power factor was greater than unity, then you started to talk about efficiency, so I wasn't sure which you were talking about to be specific.

The type of PSU, is this active PFC?
What are you measuring the output load with?

[Spectre]

Originally Posted by Kvar

I wasn't quite sure what your question was. You stated that power factor was greater than unity, then you started to talk about efficiency, so I wasn't sure which you were talking about to be specific.

The type of PSU, is this active PFC?
What are you measuring the output load with?

No I addressed both last time as I have had "free energy" appear (more than 100% efficiency and better than 1 PF.

Like I said last time I am measuring the output with the load tester (SM-8800) so I know exactly what the DC load is. Almost every PSU i get through these days is APFC.

[Kvar]

Originally Posted by Spectre

No I addressed both last time as I have had "free energy" appear (more than 100% efficiency and better than 1 PF.

Like I said last time I am measuring the output with the load tester (SM-8800) so I know exactly what the DC load is. Almost every PSU i get through these days is APFC.

What do you get with a regular ammeter when you measure the input current of that power supply and multiply by the actual input voltage?

[Spectre]

Originally Posted by Kvar

What do you get with a regular ammeter when you measure the input current of that power supply and multiply by the actual input voltage?

When I use the Extech slip on current transducer I get ~1100w.

[Kvar]

That's odd. Time for waveshape check on o'scope

[Spectre]

Originally Posted by Kvar

That's odd. Time for waveshape check on o'scope

???

[signmeuptoo]

Kvar, are you saying there is some kind of timing or phase issue here? How would that be?

[Kvar]

Originally Posted by signmeuptoo

Kvar, are you saying there is some kind of timing or phase issue here? How would that be?

Are you talking about the first post, or my response to Spectre?