I am making a controller cum display card for a UPS to display parameters and do essential housekeeping. I had developed such a product 4 years back, and want to build a better product now, in terms of RMS aacuracy, CT linearisation and parameter measurement. 3 phase input V, I , frequency, apparent power, active power, reactive power and PF, same for output 3 phases and battery voltage. Output current won't be sinusoidal, and input will have distortions even with APF.

1. One reason that I am considering an energy meter chip (ADE 7753) is that, I can measure non-sinusoidal RMS current with better accuracy and I can get my Powerfactor easily. Plus, I get the temperature without a sensor. Has anybody had rms current accuracy issue with non sinusoidal waveforms for ADE7753? Can I safely expect 1 percent accuracy?

2. If I use an energy meter, what is the best method for CT linearisation other than look-up tables? ( Generally I use look-up table and it works well, but is a bit of a mess setting up tables for different CTs for different loads!!!)

3. Energy meter is meant to measure parameters single phase or 3 phase inputs. If I want to measure both inputs and outputs of a UPS, I will have to use an analog switch and sample inputs and outputs for certain period - 5 or 10 cycles- and resetting energy values between swtiching. Does this method work with energy meters? (Of course, with ADC instead of energy meter, this works well). Is there anything that I am overlooking? Can I acquire current /voltage waveforms from energy meter using this sampling method?

4. Do you think I should abort the energy meter chip idea and go for normal DSP / controller method? Generally, people tinker with the filter after the rectifier and measure the RMS with 3 to 5 percent accuracy. (I did the same four years back, in my last design based on 89C52. Since the load consists of computers mostly, with the right resistors and capacitors we can get decent accuracy by calibrating the meter with the actual load.) But then, with technology advanced so much, Shouldn't I be computing the actual RMS? Are there any cost effective alternative chips with RMS measurement functions?

5. Could anybody throw some light on the counter price of ADE 7753/7754 in India ? MAX Q3180? Development board price?

6. Is MAXQ3180 a better choice than 7753? I find the protocol bit complicated. Could anybody guide on this?