Hacker News Reader

Comment by carreau | Hacker News Reader

"Monitor electricity 3600 times a second" in the video seem wrong, as AC is usually 50 to 60 HZ that would be measuring the consumption 60 to 72 times per period which does not really make sens. It's likely 3600/hour or 1/seconds. Immediately make me doubt the quality of the research.

Edit: DC -> AC. was tired.

View on HN

3600 times a second is reasonable if you want to calculate true RMS values at 50/60Hz. This is almost required as well as most inverter and supply waveforms are far from sinusoids.

If you have distorted currents and voltages, you need to do this. 3.6kHz sampling captures harmonics up to 1.8kHz, which is the 36th harmonic. Power is 1/T \int_0^T U(t)*I(t)dt for periodic signals with period T, and if current and voltage cannot be assumed to be sinusoidal, faster sampling is needed.

Standard advice for digital scopes is that you need a sample rate at least 10x the signal you're trying to look at. The main signal may be 60hz but for power quality you're interested in looking for harmonics, deformations to the waveform, current/voltage phase mismatch etc.

*AC is 50 to 60Hz (DC does not alternate). Also I can see how it's possible that electrical fluctuations could happen at a higher rate than the utility frequency. 3600 Hz does seem like a lot though.

Oops. Fixed AC/DC. I know better. 3600Hz seem also quite a weird number while 3600 time an hour ends up being exactly once a second.

> DC is usually 50 to 60 HZ

Immediately make me doubt the quality of the comment

Sorry meant AC, got things mixed up in my head with solar beeing DC.

Reply on news.ycombinator.com