> -----Original Message----- > From: Cliff Spradlin [mailto:csprad...@waymo.com] > Sent: Monday, April 23, 2018 1:26 PM > To: Keller, Jacob E <jacob.e.kel...@intel.com> > Subject: Re: [Linuxptp-users] ptp4l and SyncE >
> Good point. Like you said, it should only need one correction. My > frequency source is GPS so even as that frequency wanders, the master and > slave will wander along with it. > I am pretty sure that it will behave this way, (Richard knows more about SyncE, but I believe he's posted examples where SyncE does behave this way just fine). > > What behavior are you actually seeing? Is it possible that your PHC isn't > actually getting the frequency from SyncE? > > This is for a to-be-built architecture so I don't actually know the > behavior yet, I'm just planning ahead. I have some eval equipment coming > from Intel, but it may be difficult to test everything until we get our > real boards built out. I'd try ptp4l and see if it works for you. It *should* determine that the frequency offset is exactly 0, (since the two would be exactly in step at each measure minus the phase offset, which should get corrected prior to any frequency adjustments). > > Maybe I should just not run ptp4l at all and simply use a single incoming > sync packet to program the PHC. There are network switches between the > master and slave, but they all act as transparent clocks, so the > DelayReq/DelayResp is likely unnecessary (or would only account for a > constant amount of delay that could be pre-measured). > If ptp4l doesn't work you could go this route. It should be quite simple. However, I suspect ptp4l will work fine. Thanks, Jake ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ Linuxptp-devel mailing list Linuxptp-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linuxptp-devel
