I have a cluster of cRIO-9068 systems running LabVIEW Real-TIme 14.0.0 and NI-TimeSync Plugin for Software-Based 1588 14.5.0. The cRIO systems will normally synchronize to our local network time master to within tens of microseconds after a number of minutes elapse from system power-up. That's all good. Additionally, I've observed that if a cRIO system clock is grossly out of synch with the network master on power-up, the underlying PTP slave will abruptly slew the clock so as to quickly synchronize with the master. However, if upon power-up the cRIO clock is fairly close to that of the network master, then PTP will gradually steer the clock into synch with the master. The problem is that in this scenario it can take as long as 20 minutes to attain sufficient synchronization for us to perform our mission, an unacceptably long period of time.
So my question is, can anyone suggest a reliable way of forcing the PTP slaves to always slew their clocks on power-up rather than steering them? Looking here I don't see any configuration parameters in ptp.conf that would allow me to accomplish this. Nor is there a PTP command equivalent to the ever-handy ntpd -gq.
TIA
Hugh
