Experimental Physics and
Industrial Control System

"Konrad, Martin" <[email protected]> · Mon, 9 Dec 2013 15:33:15 +0000

Hi Franck,
it's good to hear that ITER is prefering TAI over UTC as well :-)

> Actually, we are suppressing POSIX UTC from timing boards' drivers
> only & convert to UTC for applications (inc. EPICS records).
What exactly does converting from TAI to UTC for an EPICS record mean? I 
see two ways to do that:

1. Convert from TAI to UTC. This requires leap second information on all 
EPICS IOCs.
2. Just copy the TAI timestamp into an EPICS timestamp and let the 
clients do the conversion. Since TAI has no leap seconds this should 
work smoothly. But strictly speaking this is against the definition of 
EPICS timestamps (which are defined to be in UTC - see epicsTime.h).

If you chose 1.: Do you periodically download the leap second 
information from a central server into all your IOCs? How do you deal 
with the fact that EPICS timestamps do not have a representation for 
leap seconds?

If you chose 2.: How do you make sure all your clients are converting 
time in the right way (including leap seconds)? Do you have any 
procedures in place for that?

Thanks a lot,

Martin

-- 
Martin Konrad
Control System Engineer
Facility for Rare Isotope Beams
Michigan State University
640 South Shaw Lane
East Lansing, MI 48824-1321, USA
Tel. 517-908-7253
Email: [email protected]

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System