On the home page is an article about an electrical power sub-metering system for use in apartment buildings or other installations where there is one main electical power meter feeding several tenants.
I given this a brief look and think I have found a few problems. I have some experience in measuring power but do not claim to be an expert in this field.
First, the schematic appears to have the current transformer wired incorrectly, if it were turned 90 degrees I think the connections would be correct but the burden resistor seems to be missing.
Second, the article describes that during their tests that they found the unit to have varying degrees of inaccuracy. They are converting the AC signals to DC signals for the microprocessor and I wonder if they are just measuring the average values of the voltage and amperage, then multiplying them together to determine the wattage consumed. This would only be valid for a non-reactive load such as an incandescent lamp or resistive heating device. Motors and transformers would cause a phase shift which could easily result in an error of 10% or more (research: "power factor" and "apparent power").
Unless the processor were taking many samples during the AC cycle, it would not be able to determine power factor. The article says that to determine power being consumed that they multiply amperage times voltage. Unless you also multiply this by the power factor, the answer could be wrong.
Aside from the possible phase-shift caused by reactive devices, distortion of the AC waveform could also cause errors in the readings.
If my concerns are valid, it would be good if someone could forward them to the authors of the article.
If I am wrong, comments would be appreciated so that I may learn from my errors.
Rat
