[time-nuts] June 30 2015 leap second

d0ct0r time at patoka.org
Wed Jan 7 10:53:24 EST 2015


Thanks for the comments ! In my design I am using NMEA as an option to 
set initial time on the clock. Its just faster and more convenient than 
doing it manually. And in the other (rare) occasions when clock logic 
could decide to sync. RTC module time with time received from GPS 
module. Basically I decide not to relay on GPS NMEA too much. My clock 
using 1PPS signal coming from internal GPS module or from external 
source connected to the device. That 1PPS and 32kHz signal from RTC 
module, connected to the MCU timers. So, I know for sure if RTC module 
generate its 32k signal slower/faster than it should be. If the 
difference between of those two signals become too big, clock will do 
"autocorrection" for the RTC oscillator to trim the value for Aging 
Register. RTC module has accuracy 2ppm. I think it suppose to keep the 
time well.
Initially I was thinking that GPS will receive "hh:59:60" NMEA message 
and I could use it as it is. But now I think
I'll add some more code to handle such wonderful thing as a "leap second 
event". I am going to create subroutines which will allow me to enter 
and keep "leaps second" event in battery backed SRAM and apply it as 
that even occurred. Unfortunately it will need the human interaction to 
set up leap second events. But if I'll leave the clock logic as it is 
now, I'll need to correct time on the clock any way. Since 1PPS just 
keep RTC oscillator in tact. But time on the clock will be 1 second 



More information about the time-nuts mailing list