n1.taur.dk

n1 is a stratum 1 server steered by pair of GPS receivers.
n1 exists because I need a reliable source of time. One where I know it'll either give correct time, or be honest and say it is unsynchronized. I've been burned one time too many.

2019.12.13: Server replacement
I replaced the hardware with something faster, more efficient, and with dual power supplies. At the same time, I forgot to put in the additional serial card for the GPS receivers, so at the moment the server is running off of a GPS receiver in the vault. It is getting a surprisingly good signal through 40cm of steel reinforced concrete and fire barrier.

2019.3.17: Anniversary
17 years of stratum 1 NTP. A bit of a ride. DCF77, single GPS, GPS and OCXO, Dual GPS and OCXO, Common view and OCXO, and then back to dual GPS. In 17 years there has been, according to monitoring, 90 hours of downtime. 99.94%. Pentium 2, Pentium 3, K5, Athlon, Atom, Pentium 4, Pentium D, Xeon, and more Xeon. Six power supplies have died, and this is the fourth chassis. The fifth chassis is on my dining room table, four sockets, redundant power supplies, and Rubidium timebase. I am not going to promise another 17 years; If you have a location, and 2U of space you want to lend to me, then maybe we can come to an unofficial deal.

2018.11.7: Another antenna relocation
The structured cabling used for the GPS2 receiver was needed for other purposes. The new location may be slightly worse, as it has both obstructions north and east of it. On the upside, the obstruction to the north is not as impenetrable, and the environment is quieter. Once I mount it, that is. Right now it has a bundle of 3-phase cables less than 30cm from it.

2018.6.1: Transceiver upgrade and antenna relocation
The transceiver driving PPS over the 133m long cat5e cable between the receiver (GPS1) and the server had died halfway and gotten weak. Since I had to unmount GPS1 anyway, it now lives in a new location. I may be better, but it might not. Time will tell.

2018.5.31: gpsd died. Again.
Now running on github.com/kkpdk/refclock-tools/ .. Eating my own dogfood, and it is yummy.

2018.5.21: new clock wiggling
After the update, the clock is now wiggling a few microseconds every two hours. It has no effect, but it is annoying not knowing what causes it. Something is causing bi-hourly temperature excursions.

2018.5.20: Downtime during upgrade
An update that was supposed to take half an hour ended up taking several and a road trip. There will be another round of updates, but for now, it is stable.




Sky view
This is an intensity-graded plot of signal strength, seen both as the usual sky view, and as a south-centered cylindrical projection. Blue indicates that the satellite is excluded from the solution, red that the receiver has lost lock.