REV3 is coming along well. I have fixed most of the problems, and it is nearly ready for a final test after the new enclosure is completed. 

One of the most annoying things about REV2 was flicker in the display. Most of the time it was just a few counts of the least significant digit, so not consequential for reading accuracy, but it was annoying and looked amateurish. 

It turns out it was mostly a firmware/algorithm improvement that fixes this. Hardware filtering before the ADC definitely matters, but what really made the display look good was using an update customized IIR filter instead of just a windowed running average, and adding some hysteresis between real measurement and what was displayed. I considered adding hysteresis a while back, but I ended up not doing it because I couldn't find any evidence that it is used in commercial meters. After further study, I think it is likely that it is implemented to some extent, but just not documented. It is obviously common practice with single bit  analog to digital (comparators), and I just don't see how commercial meters can keep as steady as they do without it. I only added a little over a count, so a little over 1mV of hysteresis in voltage mode.