[time-nuts] Line Frequency

[M. Simon]

I have done some calculations on the differential (op amp/comparator) method of zero crossing detection and it turns out to be very good. Amazingly good. 


The delay caused by the RC is roughly equal (actually very close) to the RC time constant of the RC if the RC 3dB point is 10X or more the line frequency. 


At 10X the line frequency (~600 Hz) for the 3 dB point the shift in zero crossing going from 59.5 Hz to 60Hz is about 16nS. Which is about 1 ppm of nominal 60Hz. It is roughly 1/10,000th of the change in frequency.  And of course the less filtering you do the closer to the actual zero crossing you get. And the less variation in delay with changing frequency. Changing IC temperatures is likely to cause more variations. Changing transformer temperature (for transformer isolated versions - which is what I plan on since I'm not phase controlling a triac - just measuring) will likely cause more variations. And I will know within better than 1/50th of a cycle (not counting other delays - like transformer phase shift and computing [comparison] delays among others) if the line is out of spec.  

 
Noise with a 60KHz fundamental is 40dB down! if you set the RC 3dB at 600Hz. 

To Hal Murray who suggested this method bravo!!!! It is truly time Nuts worthy. 


Simon


Engineering is the art of making what you want from what you can get at a profit.