[ntp:questions] NMEA ref.clock better than my ISP's timeserver?

Unruh unruh-spam at physics.ubc.ca
Fri Jun 12 17:17:07 UTC 2009

David Lord <snews at lordynet.org> writes:

>David J Taylor wrote:
>> David Lord wrote:
>> []
>>> First I don't think there's any problem getting the nmea down
>>> twisted pair at 4800 baud.
>> .. agreed, but .....
>>> To play safe with the ttl level pps I've setup to use a simple
>>> emitter follower that can drive into 75r coax.
>>> I didn't want to use line drivers as I don't have any handy that
>>> work off single 5V supply.
>>> I need a converter back to ttl before going to pc and don't have
>>> any spare gates in the box with compulsory flashing led and
>>> ticker, so will use another couple of transistors wired as
>>> Schmitt trigger.
>>> David
>> I think you're needlessly overcomplicating the PPS signal.  I have the 
>> GPS 18LVC feeding /two/ parallel RS-232 interfaces, /and/ a low-current 
>> LED, so for a 30m length I think you would get away with a simple wire 
>> connection.  However, have fun with the soldering iron!

>Are you getting better than 10us offsets though? :-)

I get an average of 2usec from pps driving the parallel interrupt ( with
"myown" interupt service daemon). I think much of that is a) ntp's
handling of rate fluctuations, and b)natural variation in the interrupt
handling due to computer issues. My gps 18 has about a 20m run to the
computer, and is simply a "bar wire ( well one of the Cat5e twisted pair
wires) connection.

>I've no idea how much signal degradation there would be using
>utp for pps and will both put it on spice and keep the spare
>pair in cable free to give it a try.

>However I'm making up various pcbs so an extra couple of tiny
>jobs won't add much to the fun and is playing safe in case I
>end up mounting aerials on roof or in garden, although that
>looks very much less likely to be needed now.

>> If you're seeing 4us offset now, that may be as good as it gets.  You 
>> may need a more constant temperature to see better.

>That's another project, putting the servers in a cupboard
>with temperature control. Alternative of locating the crystals
>and fitting heater is also a possibility. Cupboard idea has
>advantage of cutting down noise and will help keep the pcs
>from overheating.

Putting in a thermistor on top of the clock crystal and read the
temperature and use it to predict the rate variation. That will probably
give far better results than using a temperature controll. Also using
chrony ( which now has an shm refclock input at least in beta). will
give better results than ntp. 


More information about the questions mailing list