Post by andrew on Jan 7, 2016 14:22:03 GMT
Hi all, this is my first posting here. I have been setting up a video system following a talk from William S at the Chester astronomical Soc. With his help and that of Bill W I have got going but with a lack of help from the weather have not captured any meteors yet.
So to pass the time (so to speak) I have been looking into how to get my PC time as accurately as I could on my small LAN. I know the Nemetode network uses Dimension4 but as I have a Meinberg USB MSF clock setting the time on my main telescope PC I thought I ought to be able to exploit it. Reading up on NTP I found that the Meinberg site provides all the software needed to do this for free and simply packaged up.
I set up the original Observatory PC with the USB clock as a Stratum 1 reference and then installed the NTP software and associated monitor on the other PCs I have. I get them to poll the Stratum 1 reference every 2^5 seconds but have included a back up by polling the NPL servers ntp1.npl.co.uk and ntp2.npl.co.uk every 2^12 secs as a back up. These servers are a very good time source which seem very stable and responsive.
Looking at various logs I found the internet time often had a 10-20 ms offset compared to the MSF signal. The meteor PC which is on the wired LAN and is kept about +/- 2ms of the MSF signal on the reference PC. The attached image show a log taken while UFOCapture was running. The spikes are when the PC was rebooting during a set of windows updates.
I would conclude that using the basic NTP tools (which don't need a radio clock as it can just use internet time servers) can give good time keeping and are a potential alternative to packages like Dimension4 if needs be for any reason.
Realistically I would estimate a PC with just internet time will be within +/- 30ms of UCT and, say,+/- 10ms UCT if say an external radio or GPS clock is used. It might well be better than this but I don't have an independent way of assessing the accuracy of the Stratum 1 clock at the ms level.
Regards Andrew
So to pass the time (so to speak) I have been looking into how to get my PC time as accurately as I could on my small LAN. I know the Nemetode network uses Dimension4 but as I have a Meinberg USB MSF clock setting the time on my main telescope PC I thought I ought to be able to exploit it. Reading up on NTP I found that the Meinberg site provides all the software needed to do this for free and simply packaged up.
I set up the original Observatory PC with the USB clock as a Stratum 1 reference and then installed the NTP software and associated monitor on the other PCs I have. I get them to poll the Stratum 1 reference every 2^5 seconds but have included a back up by polling the NPL servers ntp1.npl.co.uk and ntp2.npl.co.uk every 2^12 secs as a back up. These servers are a very good time source which seem very stable and responsive.
Looking at various logs I found the internet time often had a 10-20 ms offset compared to the MSF signal. The meteor PC which is on the wired LAN and is kept about +/- 2ms of the MSF signal on the reference PC. The attached image show a log taken while UFOCapture was running. The spikes are when the PC was rebooting during a set of windows updates.
I would conclude that using the basic NTP tools (which don't need a radio clock as it can just use internet time servers) can give good time keeping and are a potential alternative to packages like Dimension4 if needs be for any reason.
Realistically I would estimate a PC with just internet time will be within +/- 30ms of UCT and, say,+/- 10ms UCT if say an external radio or GPS clock is used. It might well be better than this but I don't have an independent way of assessing the accuracy of the Stratum 1 clock at the ms level.
Regards Andrew
