[ntp:questions] Shared PPS source/Multiple PPS sources

William Unruh unruh at invalid.ca
Sat Feb 7 09:01:30 UTC 2015


On 2015-02-07, Rob <nomail at example.com> wrote:
> walter.preuninger at gmail.com <walter.preuninger at gmail.com> wrote:
>> Ok, so these questions might be off the wall.
>>
>> Is there any reason why I could not share the PPS output of say, my u-blox 7 GPS module on multiple computers? Would it be good or bad to peer these 2 systems with each other?
>
> We do that in our project to synchronize several systems to the same
> GPSDO accurately.   We use surplus "video distribution amplifiers" because
> they neatly fit into this (BNC In/Out, correct levels).  There also exist
> dedicated devices for this.

Actually the biggest problem with this is that it takes time to service
a gps interrupt. Thus the first pps to come in, or be serviced delays
the servicing of the second interrupt. Thus the accuracy of the two GPS
is very ( 10-20usec) different. (I call that "very" because a gps pps is
capable of 1usec accuracy, so a factor of 10 worse is  alot)

>
>> On the opposite end of the spectrum, would it be reasonable to have a PPS output from said module on a computer, with a second PPS (from another GPS module) on that same computer?
>
> I have that here at home, but mainly for evaluation of the results of PPS.
> (comparing kernel-PPS to usermode-PPS etc)
>
> I'm not sure if that would bring much, aside of redundancy.
>
>> The Chronodot/DS323X chips have 1 Hz sqaure wave output and they have +/- 2ppm. It seems that the ATOM clock driver could be used, but would it be better to write a driver specifically for this chip?
>
> A special driver is not required for PPS, but it may be that you need to
> write software to read the serial data when it is not supported by any of
> the drivers.   I wrote one it a Datum GPSDO, and I just made an external
> daemon that writes the time info into SHM to be read by ntpd using the SHM
> driver, so I don't need to try to get my software merged with ntpd.
>
>> And a side question: Is it the GPS module that calculates when the PPS goes active? Is this signal compensated for the time it takes the signal from the sats in the module, or on the SV?
>
> Yes, the module calculates the position and time fix from the signals of
> at least 4 SVs, and then outputs the PPS related to the calculated time.
>
> This is automatically compensated for the delay between the SV and the
> calculated position.



More information about the questions mailing list