[time-nuts] Housing LPRO and Thunderbolt together

[Magnus Danielson]

Bill,

On 04/24/2011 02:50 AM, WB6BNQ wrote:
> Magnus,
>
> I just cannot believe you are disagreeing with me !
>
> However, I still stand behind my statement.  Considering the two mechanisms, the quartz
> blank seldom fail in a proper design.  Whereas the Rubidium is guaranteed to fail as you
> have admitted.  But you do have a point that the Rubidium can be revived with a good
> degree of success.  The quartz blank, if fractured, is no more period.

Well, there is a huge difference between failure as in need total 
replacement and failure as in needing mild service.

Hence, I wanted us to separate the issue into not two but three 
categories. The reason is that for too long it has been seen as a 
(non-restorable) wear mechanism, as if the rubidium was lost. It's not 
lost, it is just relocated inside the assembly in a way that makes the 
lamp having problems. These can be solved.

So, to make my point clearer we have these three categories:

1) Fundamental wear, will fail in approx X hours from leaving factory
    - Caesium clocks has this mechanism in several places, such as
      mass-spectrometer, ion pump, caesium source and just general
      caesium pollution.

2) Relative wear, will fail in approx X hours, but can be revived 
(possibly multiple times) for additional Y hours of operation
    - Rubidium and hydrogen clocks has this mechanism. Most rubidium
      clocks can operate for about 10 years and then be revived.
      Hydrogen clocks require hydrogen refill and replacement of ion
      pump.

3) Slow material shift, will operate for very long, but may drift out of 
useful range
    - Good crystal oscillators can belong to this range. They may
      continue to drift and oscillate, but they may drop out of useful
      range as they no longer can be pulled into the useful range of
      frequencies needed for the application. The crystal blank etc. may
      not fail as such, but it can become useless never the less.

So, I mostly disagree with your grouping of 1&2 into one group and to 
some degree of the long term aspect of group 3. The division is not as 
sharp as you claim it to be, and I could come up with even more groups 
if I where to add lower quality oscillators in it... and start to 
consider electronic design issues...

So I don't think we are in wild disagreement really, I think we only 
need to talks things over to come to a mutual agreement. It's about 
weighing in sufficient aspects and knowledge. We all contribute to that 
here.

> And of course we are ignoring all the standard electronic parts, etc., and only talking
> about the primary mechanisms.

Indeed.

Cheers,
Magnus