Hi Even a small PIC has room for a fairly large table. Rather than dealing with leap years outside the table, just make it 4 years long rather than one year. If you are trying to deal with 15 minutes, the table could get to 0.1 second with two bytes per entry. It would fit in < 3K bytes. It would be half that size if you used some sort of “sum the bytes” compression approach. You would do 1,500 adds to get the number for today. I’d bet that can be done in < 24 hours. You still need to deal with the century stuff and ummm …. errrr …. (dare I bring it up …) leap seconds ….
Bob On Jan 19, 2014, at 7:25 AM, Tom Van Baak <t...@leapsecond.com> wrote: >> I wonder if you really need a special clock? Can't you adjust a normal >> spring driven clock to run fast (or is it slow?) by about 1/3 of a percent >> (one day per year)? This should be within the range of adjustment. > > Chris, > > When you mention 1/3 percent, you're thinking sidereal time, which is a > completely different concept, and much easier to implement than equation of > time. Sidereal time is simply a calendar-day independent, fixed (2730 ppm) > frequency offset. I already have PIC chips that do this; see PD28 under > www.leapsecond.com/pic/picdiv.htm or read the comments in the source code at: > http://www.leapsecond.com/pic/src/pd28.asm > > Solar time, on the other hand, is continuously variable in rate (and phase) > throughout the whole year. A microprocessor implementation of solar time also > needs to know calendar date, time, and longitude. A 4800 baud GPS NMEA stream > input would be a convenient way to obtain this information. Without using > floating point or trig functions, a tiny PIC implementation would probably > use a 365 entry lookup table to adjust the output tick rate on a per-day > basis. A more capable Arduino or RPi might allow one to accurate calculate > EOT directly from planetary motion equations, avoiding hard-coded tables > altogether. > > /tvb > > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. _______________________________________________ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
