True RMS Calculation

[udhay_cit]

This question is asked before lot of time here & other forums. But still i want to know more details about the True RMS voltage measurement by using a microcontroller. I'm a hardware designer & new to microcontroller.
My requirement is to measure the True RMS value of the AC mains (Three phase) voltage. For that i selected a PIC microcontroller PIC24FJ128GC006. It has built in 12 bit ADC with capable of 10Msaps sampling rate.

My questions are...
1. I'm going to scan the ADC channel with 1Msps sampling rate at the least. Its always advisable to take as much as samples, but is it decent enough sampling speed for a True RMS measurement?
2. I'm taking samples of every alternative cycles or couple of cycles only & utilizing the cycles gap period for calculation and other works. Is this method affect my measurement reading?
3. I want to know the basic calculation structure of the True RMS calculation specially for the microcontroller measurement.

Thanks in advance...

Regards
Udhay

 

[barry]

1) I'll answer your question with a question: How accurate do you NEED to be? 121 Hz might be good enough if you sample a long enough time.
2) If your input is continuous and periodic, i.e., each cycle looks like the cycle before and after it, then this should be fine.
3) RMS=Root-mean-square. Take the square root of the mean of all samples squared. Right?

 

[andre_luis]

This sample rate seems quite exagerated to solely take True RMS value of a 50/60Hz AC mains. Would sound more feasible to employ such magnitude if referred equipment is designed to obtain spectrum of a signal with high crest factor, which is not the case.



+++