1572Re: Timing Question
- Nov 17, 2009G'Day Dave,
I like your web site. I have a similar goal in mind - all of which came to a halt when I realised I was getting useless observations because of the problem of delivering stable time data to Faros. I hope you don't mind if I have a look at the Java on your pages.
What are the problems caused by the ISP?
For my journey through "time", I used the the usual ntpd NTP Time Server software and drivers on both Ubuntu Linux 9.04 Desktop and 8.10 Server, installing from the standard "package management" using Synaptic. On Windows XP, I used a port of the same software from Meinberg - http://www.meinberg.de/english/sw/ntp.htm - and Meinberg's NTP Time Server Monitor to observe and plot the statistics from all the ntpd servers.
I used many, and varied Internet time servers, and also connected my Garmin eTrex Vista GPS via a serial port to the Meinberg software on Windows, using the NMEA data strings only - the eTrex has no PPS output.
It was all a very good learning experience, but with a lot of frustrations - all with "the time", and my Internet connection. The actual software installations were relatively simple and straight forward. One could describe the experience like big game hunting - hours and hours of observation and "taking aim", with the actual result occuring in just a minute or two. It is the sort of thing one has to do, to really understand what is happening with time and how it affects the Faros observations.
I went through the usual careful steps - got the radio-PC connection to work with OmniRig, checked for reasonable signals from all the beacons one could expect, then I simply fired up Faros. Initially, I selected a few time servers from the list and observed what happened over a day or so. The "dots" on the detail page were all over the place, and the UTC Clock accuracy "red line" went all over the place. Next I updated the TimeServer.lst with the usual "pool" of servers for Australia - see www.pool.ntp.org. For the UK they would be things like 0.uk.pool.ntp.org. I got the same result - dots all over the place.
That's when I implemented my own time servers.
That gave me a reasonably straight line for the red "Faros time" in UTC Clock Info, Clock Accuracy. But the Faros observations on the Details panel went on a "roller coaster ride" day after day - the line of "dots" for beacon observations varying by 80-90ms over the day. The coloured blocks on the History page looked wonderful, but the data behind them was rubbish. I sat and watched the Monitor page for hours - time and again I watched Faros totally ignore nice strong beacon signals. Faros ended up following my time servers, which in turn followed the Internet time. Faros followed my servers because they were the fastest in terms of packet delay, and time variation. But, over many hours my time servers followed a large sine curve, with the 15 minute panel in the Clock Accuracy looking very flat.
A last act, which has proven reasonably successful so far, was to decommission my time servers, but trawl through their logs and find the servers that were used for long periods of time as the reference. I then did ping tests, and trace routes on them, to find the closest and fastest 10 servers. I now have reasonable, straight lines of dots on the Details panel, and good SP/LP observations with no "missed" observations by Faros. But ... I really don't expect it to last, and I certainly expect it to vary somewhat. I think it is good enough for casual use, but not enough for propagation studies, or good radio contest planning.
The observations from the Garmin eTrex NEMA strings were useless - most of the Internet servers had less variation in time.
In summary, my conclusions.
My "broadband wireless" Internet connection is far too variable in terms of "packet delay" for good time keeping at the precision needed by Faros. Its nice and fast under normal use, but not good enough for time keeping, at the accuracy we need for Faros. The packet delay is mostly in the "wireless link", but add that variability to the variability of the Internet as a whole, the Internet time server load and subsequent variability, and finally, the use of the connection by other people and devices in my household - and Faros cannot "keep time" well enough.
A NEMA strings based "consumer" GPS is not good enough - much too variable for the accuracy we need for Faros. I will report whether a "good" GPS can do it - see below.
A "wired" ADSL or cable Internet connection may be good enough, provided the connection is stable in packet delay - that is, not overloaded by other uses in your household, or the network between you and the time servers. Only good time server selection, and rigorous observation over time will show whether or not the connection is adequate for serious use with Faros.
My current plan.
I initially dismissed creating a stratum 0 reference time server - the Meinberg PCI card is priced at $3500 in Australia. But ...
Alex VE3NEA alerted me to the following web site http://time.qnan.org I have ordered a Garmin 18x LVS, $100 n Australia, and I will proceed to integrate it into my Linux server. On that site there is a reference to http://www.satsignal.eu/ntp/index.html which is a fantastic source of information relevant to getting Faros a good time signal. I have high hopes of having a fantastic Faros installation for propagation study, site and antenna comparisons.
--- In firstname.lastname@example.org, "Dave Baxter" <dave@...> wrote:
> I'm the guy who created the IC-PCR1000 "helper" app, so it can be used
> with Faros/Omni-Rig. I did test it here before letting Dave M have a
> copy, and AFIK no timing issues, but I stand to be corrected on that if
> any are found directly attributed to that program of mine.
> I do agree with Peter, try things with a rig that is directly supported
> by Omni-Rig first, just in case.
> Question for Peter. What software, program, tool, etc, did you try when
> you tried a local GPS driven NTP server of your own? Did the
> GPS/Server use a 1PPS signal, or just the NMEA data from the GPS?
> Regards to All.
> Dave G0WBX.
> PS: My Faros status page, all working from an old Icom IC-R70, and
> another of my own programs to glue it all toghether.
> Yes, I too have some NTP issues, but caused by my ISP.
- << Previous post in topic Next post in topic >>