Re: Accurately measuring clock drift
Rob Janssen <chrony-users-Cohe+2zZ/ftmR6Xm/[email protected]>
| Newsgroups | gmane.comp.time.chrony.user |
|---|---|
| Organization | PE1CHL |
| Message-ID | <[email protected]> |
On 6/21/23 13:31, Kevin P. Fleming wrote: > On Wed, Jun 21, 2023, at 07:17, Rob Janssen wrote: >> Miroslav, >> >> Are you aware of any affordable PCIe card that supports PPS input on a >> generic PC server? >> >> Ideally would be like this: >> >> - TTL input of a PPS signal with minimal pulse length all the way down to 100ns >> - some hardware time stamping method >> - supported interface to chrony > That's going to be the tricky bit... Chrony does not talk to hardware directly, it relies on interfaces provided by the kernel, so any interface of this type would need to conform to some existing kernel interface, or the supplier would need to provide a driver for the user to compile/install. Well, what I mean is that it provides a kernel pps module, software that maintains SHM, a socket interface or whatever other interface that can be configured in chrony. Of course it will always require some software. > > At this point it seems the industry has moved towards doing this over Ethernet, using PTP, so in this 'ideal world' your GPSDO would have an Ethernet connection and provide a PTP source. > Unfortunately in our case we mainly work with surplus equipment like the Trimble Thunderbolt or even Datum 9390, or we use locally available 1PPS/10MHz signals from distribution amplifiers that also feed other local transmitter equipment. Rob -- To unsubscribe email chrony-users-request-kWFZVVI9zxvPqho9SqqRMmD2FQJk+8+b@public.gmane.org with "unsubscribe" in the subject. For help email chrony-users-request-kWFZVVI9zxvPqho9SqqRMmD2FQJk+8+b@public.gmane.org with "help" in the subject. Trouble? Email [email protected]