[time-nuts] Maximizing Cs Tube Life

[John Miles]

>
> I have a 5061A and 5061B and both have a position on their front
> panel MODE
> switch that is CS OFF.  When in this position the CS OVEN is off, the OSC
> OVEN is on, the ION PUMP is working and there is signal out (10
> MHz, 5 MHz,
> 1 MHz, 100 KHz, as appropriate) as measured by the CIRCUIT CHECK meter.
>
> Is the 5062C different?  Is there something I am missing here?  I
> guess I do
> not understand the need to 'add a switch' unless it is the desire to only
> run the ION PUMP and the OSC OVEN but not the OSC or anything else.

Right; there was no CS OFF switch on the 5062C.  You can open the loop, but
that's it.  I could have run the ion pump supply from the rear panel, but
that would have taken a separate power supply, and would have left the
crystal oven cold.

It sounds like the switch I added does essentially the same thing as the one
you're talking about.  In your case you could maximize tube life by leaving
that switch set to CS OFF except when needed.  That doesn't answer the
question of how long the Cs tube needs in order to meet specs, but
presumably the manual would...?  I'd be curious to hear the answer myself.

> The one thing I have noticed is the incredible stability of the crystal
> oscillators for weeks at a time as compared to my Thunderbolt.
> On the order
> of 0.1 Hz at 10 MHz after a week.

The Thunderbolt should be in that ballpark (being only 1E-8 per week) when
undisciplined, and of course much better when GPS-locked.  A good OCXO, when
not locked to anything else, should stay put to the tune of 1E-10 to 5E-10
per day.

The HP 10543 oscillator in my 5062C had a bad case of near-constant phase
jumps, so I have a Wenzel Timekeeper-class unit in there now, plus an x2
doubler to get 10 MHz out the front panel.  I'm not sure I have the loop
bandwidth optimized for it yet, but at least it locks up nicely and doesn't
jump.

-- john, KE5FX