[time-nuts] 10 MHz Phase Coherence
martyn at ptsyst.com
Sat May 26 19:20:53 EDT 2007
Thank you everyone for your input on the topic of the 10 MHz sinewave output being in phase with the UTC 1 pps output.
My customers actually want the 10 MHz zero crossing to coincide with the 1 pps rising edge.
That's what I really mean in my first question. That has been answered thanks.
But so far, on all the GPSDO's I've tested, the 10 MHz sinewave drifts in phase wrt the 1 pps pulse, even if the 1 pps pulse is derived from that sinewave!
There are different reasons for this. For example, many GPSDO's have distribution amplifiers with LC filters. The LC filter will cause the phase of the 10 MHz sinewave to drift with temperature.
Another reason is that many GPSDO use the GPS receivers raw 1 pps as their UTC aligned output.
Then they use the same 1 pps signal to discipline the 10 MHz oscillator.
However, the 1 pps is only used to discipline the frequency of the 10 MHz, not the phase.
This is how it is done on some of the lower cost GPSDO units (and my first design fifteen years ago).
The same problem goes with distribution amplifiers. Many distribution amplifiers change the phase of the 10 MHz signal over time and temperature. Most customers don't know this happens and don't care, as they are only interested in an accurate frequency.
If a customer is interested in stable phase in a distribution amplifier, they must make sure the amplifier is phase stable.
More information about the time-nuts