[ntp:questions] X2 assert and clear events (PPS) on LInux with sure GPS Board

Edward T. Mischanko etm.pjm1 at frontier.com
Sun Jan 13 00:01:56 UTC 2013

Does Richard B. Gilbert have nothing better to say, than to post the same 
reply 5 times, and not deal with the question at hand?  And Unruh,  let’s 
grow-up and reply to posts in clear, COMPLETE, thoughts that truly answer 
the question.


"Richard B. Gilbert"  wrote in message news:50F1D262.50103 at comcast.net...

On 1/9/2013 7:09 PM, unruh wrote:
> On 2013-01-09, thelastromantic at live.fr <thelastromantic at live.fr> wrote:
>> Hi,
>> Thanks, it seems to work.
>> So, i just need nanoseconds precision with this 1pps on my localhost,
>> don't need NTP or other.
> You will NOT get nanosecond precision. If that is your goal, give up
> now. No computer can reach that. a)the time sources are not good enough
> without corrections after the fact, and b) your computer cannot process
> the pps signal fast enough to give you that precision.
>> I've have compiled succefully compiled my program with timepps.h,
>> and i want to create a nanoseconds chronometer between each 1PPS.
>> What function should i use from timepps.h code ?
> You don't.
>> Another solution : i've modeed ppstest, and i settimeofday to "0" each 
>> PPS,
>> but the accuracy of settimeofday() seems to be bad, and i want just 
>> change
> the seconds to "0" on my local PC without change HH.MM and date.
juste change
> the seconds to "0" for each PPS/seconds
> That takes too long. And if your frequency is out by 1PPM ( and that is
> a very good computer clock if its drift is only 1PPM)  your clock
> will be out by 1000 ns by the time the next pulse occurs.
> Ie, you have set yourself and impossible task.
>> Sorry for my bad english

Your English is alright!  Have you noticed the linguistic
sins of some of the native speakers? ;-) 

More information about the questions mailing list