Although DCF77 seems to be the ‘default’ time reference for the European continent, I want to have a backup because I run a stratum 0 NTP server.
MSF (60 kHz England, UK) is a good candidate. Compared to DCF receivers, MSF receivers are difficult to obtain and expensive on the continent.
So, I modified a Conrad DCF77 module for the reception of MSF, using a 60 kHz watch crystal as filter.
I tuned the antenna/ferrite coil from 77.5 to 60 kHz.
Click on the picture to enlarge v0.1 of the modified radio module.
The coil is from an old transistor radio but later I retuned the original ferrite coil as supplied with the Conrad radio module.
Most radio clock IC’s want to see a resonated loaded impedance between 50 – 100 kΩ. I simply multiplied the capacitance of my DCF77 ferrite antenna with 1.67 ( [77.5/60]²), i.e. for my Conrad radioclock module I added 4n7 parallel to the original 6n8 capacitor.
Using ntpd with the SHM refclock enabled and radioclkd2 as ‘atomic radio clock parser’ the MSF signal from Anthorn is decoded here:
pulse start: at 1206318120.024987
MSF : |0 |5 |10 |15 |20 |25 |30 |35 |40 |45 |50 |55
MSF time: 2008-03-24 (day 1) 00:22 GMT
clock: radio time 1206318120.000000, average pctime 1206318120.024964, error +-0.004383
pulse end: length 0.494541 – 0: 5
pulse start: at 1206318121.039543
pulse end: length 0.089901 – 1: 1
pulse start: at 1206318122.025892
Update 1: it seems that there is some (propagation?) interference in the night which causes ntpd to loose lock sometimes. Perhaps I need a better antenna as the ERP of MSF was decreased after the relocation (April 1st 2007) from Rugby (ca. 50 kW ERP) to Anthorn (ca. 15 kW ERP).
Update 2: Today (18 april 2008) I received a Meinberg analogous DCF77 antenna module from a collegue. I tuned it to 60 kHz. First impression is that it works significantly better than the original antenna.
(click on the image to enlarge)