[time-nuts] 60Hz zero-crossing

Tom Van Baak tvb at LeapSecond.com
Sun Jul 3 19:37:46 UTC 2011


> OK, I wasn't paying attention as the info passed by. 'Xactly how is this
> huge signal introduced to the PC? I remember something about a voltage
> divider off the hot side of the line, put on an input pin of the PC's
> com port and then somehow timestamped and put on a data file. Howsat
> done again? Sorry to be so lame, but I really wasn't paying attention.
> Don

Hi Don,

What I did in the quick PC experiment was feed the output of a
5 VAC wall-wart transformer through 1k to DCD (pin1) of the
PC's serial port. Probably not wise to voltage divide raw mains
power and send it to a serial port. If using a high-z input of a
microcontroller, the doc2508.pdf app note that you found shows
both sides of mains going through 1M resistors.

There was concern that noise would cause false triggering. There
are a variety of hardware-only or software-only solutions to this
concern. Bill found a robust ZCD circuit if you need a hardware
solution. Time-stamping samples is the basis of software solutions.

Note looking for a sample 990 to 1010 milliseconds from the last
sample is a nice way get 1PPS from 60 Hz. This is more immune
to noise than traditional division-by-60 techniques.

As it turns out, I've gotten clean 1PPS data with no h/w or s/w
filtering so maybe this whole noise concern is overblown.

/tvb




More information about the time-nuts mailing list