Re:Faros has stopped detecting beacons
- --- In email@example.com, "Bert - W0RSB" <bert.hyman@...>
>Generaly, the server timestamps should all be in the same range, +- a
> I think there might be a time problem ...
few mS, straddling the red line. That in turn often slopes up or down,
depending on the exact CPU clock behaviour of your machine.
Try and find out what addresses your ISP's NTP servers are at, and add
them to Faros's list of known servers.
You might also want to use the local (to you) "NTP Pool" server.
I found that you need to watch that time status display, and see which
servers were consistantly the best at timestamping, then de-select the
others, leaving 2 or three relatively good ones.
You could also delete some/all of the other server addresses that can't
be reached. The file you want is 'TimeServer.lst' in Faro's main
folder. You'll need to stop, unload and restart Faros for any changes
to be seen.
I have to say, I gave up about a year ago with both my ISP's and public
servers, as my ISP/WAN performance is too variable now, and set to and
got a local NTP server working on an old PC, using a Garmin GPS16
receiver (any RX *Must* have the 1pps signal) and the Meinberg port (on
Windows 2000) of the NTPD daemon software.
It works very well. Short latency (single figure ms!) and stable
timestamps. I have a FreeBSD based machine waiting in the wings too.
I've also got all other pc's in the house to use it as their time
See what you can scare up from your ISP if posible, but many are
rationalising things to save costs, and sadly they seem to agregate
several services onto one set of machines. My ISP use their gateway
router machines to run their NTP servers now, result, like you, time
stamps all over the place, and variable distribution as their network
load changes over a 24 hour period. Not good.
If you cant find a good couple of public servers, let us know, and I'll
point you at the places I found the info about, re setting up your own
- --- In firstname.lastname@example.org, "Dave Baxter" <dave@...> wrote:
>The servers all appear to be consistent with themselves, but are tens
> --- In email@example.com, "Bert - W0RSB" <bert.hyman@>
> > I think there might be a time problem ...
> Generaly, the server timestamps should all be in the same range, +-
> a few mS, straddling the red line. That in turn often slopes up or
> down, depending on the exact CPU clock behaviour of your machine.
or even hundreds of milliseconds apart from one another.
>I'm using my ISP's server (ntp.iphouse.com), my ex-ISP's server
> Try and find out what addresses your ISP's NTP servers are at, and
> add them to Faros's list of known servers.
> You might also want to use the local (to you) "NTP Pool" server.
(ntp.visi.com), and two servers from the U of Mn. Since they're
all here in town, I expected decent results. I looked at the
pool servers, but haven't added one to the list yet.
The servers I'm using all report delays of less than 50ms on the
Faros test, but if I'm interpreting the numbers correctly the
time difference between my own ISP and my ex-ISP is on the order of
300ms. The two servers at the UoMn appear to be about 170ms apart.
The only way I can get Faros to recognize a beacon is to manually
enter a DelayCorr value of about 350ms.
This had been working flawlessly for a very long time; I've made
no intentional changes here, so I'm somewhat at a loss.