[time-nuts] Hydrogen maser frequency jumps
magnus at rubidium.dyndns.org
Sat Dec 5 07:25:06 EST 2015
On 12/05/2015 11:45 AM, Poul-Henning Kamp wrote:
> In message <CALH-g5bQ-muUQP=HpXR6wNciNqBwwOFJ0XOb3q15wYOEwOXmxg at mail.gmail.com>
> , Jim Palfreyman writes:
>> The weird thing is that the jump is ~0.7ns and that 1/0.7E-9 is close-ish
>> to 1420.4MHz which is the hydrogen line. Too much of a coincidence for me.
> Is there a direct digital divider chain to the PLL which steers the
> 5Mhz OCXO ?
> If so the obvious thing to look at is a weak input signal on the input
> side of that divider.
Indeed, as it does cycle-skipping, so it seems you miss detecting one.
You could get similar behavior if you ripple ahead a cycle every once in
I've been fighting cycle-skipping issues before, but not on H-masers.
Check your signal level is a good start.
More information about the time-nuts