Power Meter Calibrator

[Phaedrus]

So we're getting a new power meter to use for our 1200W+ power supplies, since our current one tops out at 1400W AC. Problem is, the new one is less accurate and less time/temperature stable than our current meter. So I want a quick load I can use that is stable and accurate to calibrate with before doing serious measurements.

So I've picked up some goodies:





I got some 500 ohm, 50W, 1%, 20ppm/C power resistors, and a solid copper heatsink out of our scrap room.

The resistors are 1% accurate, but in practice will likely be within +/-0.1%. I'll profile them with my multimeter to get the actual value. 20ppm/C, going from 25C to ~70C, gives me a resistance change of .45 ohms. At 120VAC that's a 0.025W difference per resistor. So this will be plenty stable for my needs.

The heatsink is designed for server CPUs working in a low-airflow environment. It's solid copper and is designed to, without a fan, cool a 95W CPU to less than 75C. Using three resistors, I'll be around 85W, so I should be good. If it turns out they get too hot, I'll pop a fan on each side, which should keep it to below 50C.



I'm going to drill into the copper and screw mount the resistors with some high quality thermal paste (we have a whole box of the stuff in the warehouse). Then attach the resistors in parallel with some stripped 14ga wire, solder it securely, then seal it with heat shrink. Attach the leads to a standard power cord, and plug into the meter for calibration.

To calibrate the power meter I just need to plug it in, let it warm up for 10 minutes or so, then measure the AC voltage and the resistance of the calibration load, calculate the wattage, then plug that into the power meter. I should be able to get within <1% accuracy if I use the meter within the next couple hours, which is good enough for my usage scenario.


Also the resistors are labeled "M1337", so clearly this makes me a pro h4ckzorz

[Jay_Diddy_B]

Hi,
Testing with a resistor load will give you the accuracy of the power meter with a unity power factor load. You can measure the current with an ammeter and the voltage with a voltmeter

 power = I x V

Do you power supplies have a PFC and draw a nice sine wave current from the line, or are they non-PFC designs that draw a current with a modest crest factor?

If the load is not unity power factor

then W is not equal to V x I

PF = W / V x I




Jay_Diddy_B