[ntp:questions] Windows time question.

unruh unruh at wormhole.physics.ubc.ca
Wed May 4 14:49:46 UTC 2011

On 2011-05-04, David J Taylor <david-taylor at blueyonder.co.uk.invalid> wrote:
> "unruh" <unruh at wormhole.physics.ubc.ca> wrote in message 
> news:slrnis0nra.4on.unruh at wormhole.physics.ubc.ca...
> []
>> Unfortunately , AFAIK, usb is terrible for delivering a PPS-- ie no
>> interrupt lines-- the problem with having only two data lines (one beign
>> signal ground)
> USB is not as good as a direct connection, but it is not "terrible" - see 
> my tests here:
>   http://www.satsignal.eu/ntp/NTP-on-Windows-serial-port.html#usb
> PPS delivered over a serial-USB convertor which carries the DCD line shows 
> significantly better performance than syncing over a LAN connection to a 
> stratum-1 server.  Jitter was around 45 microseconds with the PPS over 
> USB, and about 160 microseconds when LAN synced.  It could be that the 
> faster polling with the PPS source also helped reduce the jitter.  PPS 
> over USB is likely to be significantly better than a WAN sync as well.

Jitter and accuracy are not the same thing. It would be interesting to
attach the PPS to a real interrupt (ie one which can be serviced on the
1usec level) and see what the offset of the usb pps actually is. 
>From what I read, the usb interrupts have a minimum latency of 125us but
more likely much larger ( but less than 100ms). 

> Although I have a GPS on a USB stick, it's a logging device and I don't 
> think that it supplies a NMEA feed.  Perhaps I could test this later this 
> year.
>   http://www.ventusdesign.com/products/g730-ventus-gps-logger/
> Cheers,
> David 

More information about the questions mailing list