My 5065A came to me after suffering a short in the
oven heater winding. The heater is a bifilar wound solonoid
coil built to cancel any magnetic field that the heater current
would produce (which would steer the Rb filters). The winding was
made of enamel coated nichrome wire (about #36 gauge). Well,
The enamel probably had a small scrape between two of the windings,
and shorted the oven out so that all of the current was concentrated
in the small section of the coil that surrounded the lamp and its
circuit board. The oven got so hot that the solder on the lamp board
melted and flowed away from the joints. The PC board was a nice
chocolate brown in color. Interestingly enough, it didn't blow any
thermal fuses (there weren't any!).
I scraped away the excess "carmelization" from the board, and filled in
the fibers with epoxy, and resoldered the empty solder joints, and the
lamp came right up. I rewound the heater with a coaxial heating element
and everything was ducky.
Until some part in the main chassis blew. I haven't had time to go back
to it.
-Chuck
[EMAIL PROTECTED] wrote:
Hi Robert,
I guess you're talking about real "failures",
rather than predictable wearout mechanisms (lamp death).
My lamp failure is similar to tires wearing out on
a car. One doesn't consider the car to have "failed"
just because the tires have worn out. Anyway, we
have certainly seen the common failures you mention.
My first job out of school was on a project for HP to
design a mini-rubidium source. We used a transistor
to heat the oven mass and I remember cringing at that
thought since I had just learned that heat was a
transistor's worst enemy! I watched somebody else go
through the design and verification of a power oscillator
used to drive the lamp ... a few failures there, too. :-)
Tnx,
Jim
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Behalf Of Robert Lutwak
Sent: Saturday, August 06, 2005 5:11 AM
To: Discussion of precise time and frequency measurement
Subject: Re: [time-nuts] Newbie with new questions: Rubidium
The lamp runs quite hot (typically over 100C). Failure is
generally the
heater, which is typically a transistor or FET bolted to the
lamp body.
Occasionally, the RF transistor which drive the lamp discharge
(something in
the range of 60-150 MHz) will fail. Again, it's the
electronic components
failing, not the physics.
In a very old Rb, you may see some reduction in signal level due to
darkening of the lamp glass. This may lead to diminished stability
performance and contribute to long-term frequency drift, but
is unlikely to
cause lock failure.
-RL
------------------------------------------------------------
Robert Lutwak, Senior Scientist
Symmetricom - Technology Realization Center
34 Tozer Rd.
Beverly, MA 01915
(978) 232-1461 Voice [EMAIL PROTECTED] (Business)
(978) 927-4099 FAX [EMAIL PROTECTED] (Personal)
(339) 927-7896 Mobile
----- Original Message -----
From: "Magnus Danielson" <[EMAIL PROTECTED]>
To: <time-nuts@febo.com>; <[EMAIL PROTECTED]>
Sent: Saturday, August 06, 2005 7:40 AM
Subject: Re: [time-nuts] Newbie with new questions: Rubidium
From: <[EMAIL PROTECTED]>
Subject: RE: [time-nuts] Newbie with new questions: Rubidium
Date: Fri, 5 Aug 2005 16:45:17 -0700
Message-ID:
<[EMAIL PROTECTED]>
Our experience has been that the most common
failure in Rb standards is the Rb lamp. There
is a known wearout mechanism in the lamp and
a somewhat predictable lifespan. The filter
and absorption cells have no wearout mechanism.
Some Rubidiums have a "hot-swap" of the Rubidium lamp, such
as my R&S
Rubidium.
Unfortunatly I don't have the replacement part.
With hot-swap means that all electronics, heating etc is all
running as
normally, but during the operation there is a natural loss
of tracking.
Cheers,
Magnus
Have a nice weekend,
Jim
Agilent Laboratories
Palo Alto, CA
_______________________________________________
time-nuts mailing list
time-nuts@febo.com
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
_______________________________________________
time-nuts mailing list
time-nuts@febo.com
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
