[time-nuts] Rubidium Oscillators and Racal 1992

[iovane at inwind.it]

I purchased on ebay four rubidium LPRO and I'm quite satisfied with them.
The lamp-Volt parameter (which is an indication of the lamp's health) was 
respectively 8.4, 7.95, 6.8 and 5.1 Volt.
When an LPRO leaves the factory its lamp-V is around 12V. A drop of up to 3 V 
would be expected in the first year of operation, but after that the voltage 
decay rate slows down dramatically.
According to the manufacturer, when the lamp-volt drops down to 3 V (this is 
expected in not less than 10 years) the unit needs maintenance.
I've used the worst unit (5.1V) to make some tests in order to get more 
familiar with them. Adjusting the capacitive trimmer solid with the lamp assembly 
the lamp-V ranged from 1.5 to 6.5 V. It remained locked even when the lamp-V was 1.5V.
All of the units showed a VCXO steering voltage in the middle of the usable 
range.
The unit with 7.95 lamp-V is running continuously since 40 days. I'm monitoring
 and recording every 15 seconds the lamp-V, the VCXO V and the temperature of 
the heatsink at the baseplate (you should provide a heatsink). After a few days 
of stabilization (in wich the lamp-V showed a tendence to increase, may be following 
a long period of no operation) the lamp-V oscillates within 10 milliVolt basicly 
according to ambient temperature, and I've not yet noticed any decay in the averaged
 data. This is a simptom of good health from which one could expect some 10 more 
years of continuous operation.
I'm measuring and recording at the same time its frequency as read by a 
Racal 1992 with option 04E (OCXO). Of course the observable drifts would came 
from the Racal oscillator.
Well, I would spend a few words on the Racal too.
I'm making the measurement at a resolution of 1 part in 10e-10 (gate time 10 seconds,
 good ADEV for both units) and often I' can't appreciate any drift at this level on 
a daily basis. In averaged data, the drift is something like 2*10e-11 per day 
(Racal specs less than 5*10e-10 per day).
By the way, speaking of the Racal 1992, I have to say that mine is the military 
version 02M. I have two of them, and a friend of mine another one. None of the 
three units showed the problem with pushbuttons. May be the military version used
 different materials? Anybody knows? The unconvenience with the 02M version is that
 it uses a military GPIB language (MATE/CIIL), and I had to get familiar with it.
Antonio I8IOV

> On Mon, Mar 10, 2008 at 10:12:46PM -0500, Brian Kirby wrote:
> > For maximum stability, the rubidium should be run continuous.
> > 
> > I lost a lamp in a FRK-L last summer.  The unit had been running for 9 
> > years continuous - and I was the second owner.  It was used at a LORAN 
> > site previous, I am sure they probably ran it continuous, for the same 
> > reasons.
> 
> 	According to various spec docs I have seen, many of the telecom
> oriented rubidiums (such as the LPRO and FE series) are designed for
> between 15-20 years lamp life continuously on... so running one that way
> is not likely to use up available life quickly unless the unit already
> has lots and lots and lots of hours on it.   And they aren't absurdly
> expensive on Ebay these days - bet a new bulb (if even available) is
> almost as much as a  used but working rb off Ebay.