[time-nuts] Confused about Rubidium oscillators

[Magnus Danielson]

On 11/14/2012 10:26 PM, cdelect at juno.com wrote:
> Bob,
>
> Modern Rubidium units like the PRS10 Have a lamp life of over 20 years
> (of continuous operation).
>
> Vintage HP 5065A units likewise had a very long lamp lifetime.
>
> All the other common failures are just in the electronics.
>
> The main interest for time nutters is how long will the lamp last in my
> 2nd hand rubidium!
>
> Considering the low cost of modular rubidiums these days just get another
> when the original dies!

The failure mechanism typical of rubidium lamp isn't that hard to 
handle, and doesn't take that long time to "repair". The issue is just 
that the deposit will create a filter which reduces the needed line 
emissions. This shift in effective intensity will also cause a frequency 
shift over time, a drift component. Modern rubidiums can stablilize the 
intensity to reduce this effect.

One of my rubidiums actually has a built-in lamp-cleanup procedure.

If you have an issue with the lamp, use a heat gun and let the rubidium 
evaporate from the glas, and let the cold spot be at the top of the lamp 
such that the evaporated rubidium collects there... a process taking 
maybe 5 min or so, but removing and replacing the lamp may vary 
depending on the build-up of the rubidium clock itself.

So, well, it's more about how often you need to attend to it. Other 
failure mechanisms come into play, especially considering the life-time 
of 20 or more years. Considering that it seems reasonably simple to 
attend to, I'm not more worried about that then loosing the ability to 
get to semiconductors for instance.

Cheers,
Magnus