[time-nuts] GPSDO for my rubber duckie
lists at rtty.us
Mon Aug 1 19:41:09 EDT 2011
....or drop the GPS and just run NTP. Simpler still run SNTP on something small. You won't hit 1ms in either case. You will be plenty good by wall clock standards.
On Aug 1, 2011, at 6:27 PM, Chris Albertson <albertson.chris at gmail.com> wrote:
> The below could work but even it is overly complex. Try this..
> 1) Get any computer and install NTP. Use any decent timing GPS as a
> reference. Now you have a computer with system time accurate to UTC
> and the sub millisecond level. This is a 100% conventional setup, no
> rocket science.
> 2) Next in software you run an infinite loop, not interrupts, no
> precision timing, just a plain out "user mode" loop. The loop does
> the following. read the system clock, convert to siderial time,
> display the time, go to start of loop. Yes, there is a delay during
> the conversion but it is hundreds of time less than the delay between
> your eyes and your brain so you will never notice. Plus NTP is the
> ability to back out a fixed bias, you can use that too.
> The above does not require any advanced level of programming, no
> interrupts or real time software. NTP takes care of almost all the
> details and will easily work at the 100 micro second level. Google
> will turn up the UTC to Sidereal conversion code. You need to add the
> loop and display code.
> On Mon, Aug 1, 2011 at 2:56 PM, Bob Camp <lists at rtty.us> wrote:
>> Without knowing the full objective it's always hard to offer up suggestions. I'm going to *assume* that you want a PPS that is good to < 1 ms on your custom time scale. I'm also assuming that sidereal time is a reasonable analog to what you are trying to do. I think your approach is overly complex for implementing this.
>> An alternative approach:
>> 1) Take the PPS out of the GPS and route it to an interrupt on a cheap processor.
>> 2) Use the interrupt to start a timer
>> 3) Take your custom PPS off of the output of the timer
> Chris Albertson
> Redondo Beach, California
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
More information about the time-nuts