[time-nuts] Ublox timepulse

[Bob kb8tq]

Hi

I suspect that even with setting the timepulse to 60 seconds and aligning it, you still will
be routinely parsing the serial just to be sure “all is well”. There are always random things
that come along and mess this or that up …..

Bob

> On Feb 4, 2020, at 10:11 AM, Tim Shoppa <tshoppa at gmail.com> wrote:
> 
> freqPeriod is in microseconds up to 2^32 so you ought to be able to set it
> to 60,000,000 and get one pulse a minute.
> 
> If you set alignToTow will align it to the top of second but not top of
> minute so you'd have to parse the UTC timestamp and use it to "start" the
> 60 second cycle at the correct second. I don't see an exact spec for the
> phasing detail but you might try starting it when you see UTC timestamp at
> 59 seconds or 00 seconds and see if either of those guarantees minute
> phasing.
> 
> A more conventional way of dong this with any GPS module with PPS, is to
> parse the UTC timestamp looking for 59 seconds, and use this to enable a
> gate that lets the next PPS pulse through.
> 
> Note that freqPeriod=60,000,000 will require re-alignment after each leap
> second.
> 
> And even looking for 59 seconds to enable a PPS gate will have problems in
> the minute immediately after leap seconds. And a whole missing minute pulse
> if there is ever a negative leap second. This is a leapsecond list !
> Enabling the one-after-next at 58 seconds will get you the right number of
> minute pulses per day should there ever be a negative leapsecond.
> 
> Tim N3QE
> 
> On Tue, Feb 4, 2020 at 5:32 AM Christophe Huygens <
> christophe.huygens at kuleuven.be> wrote:
> 
>> Hi,
>> 
>> 
>> I need a sync pulse every minute at start of minute, aligned with UTC time
>> (let's not discuss what UTC time really is :-)). Any quick assessment
>> whether uBlox TIMEPULSE out can do this?
>> 
>> 
>> It does not seem to be a problem to generate a pulse every minute based on
>> UBX-CFG-TP5, but how can this be aligned with a minute. Is the only
>> solution to query time eg every top of second to see if minute has started?
>> 
>> 
>> I can parse this out of the GPS data even by polling but this seems ...
>> not elegant.
>> 
>> 
>> A few 10s of millsecond after exact start of minute would be fine.
>> Thoughts? Target CPU is very limited, say ATmega328. Or none at all if it
>> can be avoided.
>> 
>> 
>> br
>> 
>> Christophe
>> 
>> 
>> 
>> 
>> 
>> ________________________________
>> http://distrinet.cs.kuleuven.be/people/showMember.do?memberID=u0013167
>> 
>> prof. dr. ir. Christophe Huygens  IMEC-KULeuven-DistriNet
>> _______________________________________________
>> time-nuts mailing list -- time-nuts at lists.febo.com
>> To unsubscribe, go to
>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>> and follow the instructions there.
>> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.