content/10MHz_ref.rst

==========================================
 Building a bench 10MHz reference: part 1
==========================================

:author: David Douard
:date: 2016-02-21
:Category: Electronics
:Tags: test equipment, 10MHz, Rubidium


Many electronics test equipment need an accurate time base. Most
often, it's build on a 10MHz base.

For quite a while now, I have a used rubidium frequency and time
reference (Efratom Model LPRO-101) waiting for a nice enclosure. I
also have a small Extron Electronics video signal amplifier I thought
I could use to make my freq reference able to provide its signal to
several test equipments at once.

The Rubidium Frequency Reference
================================

The device is a small enclosure with only a 2x5 pins connector:

.. image:: {filename}images/freq_ref/efratom.jpg
   :alt: The Efratom LPRO-101 frequency standard.
      
The exact model is an Efratom Rubidium Frequency Standard Model LPRO-101:

.. image:: {filename}images/freq_ref/efratom_sticker.jpg
   :alt: The Efratom LPRO-101 frequency standard.

Inside the enclosure, there is a single PCB with SMD and trough hole
components. It's quite dense.

.. image:: {filename}images/freq_ref/efratom_inside.jpg
   :alt: The PCB of the Efratom LPRO-101 frequency standard.

The rubidium lamp with the resonant cavity occupy almost 1/4 of the
total area. The cover of the enclosure is compartimented, mainly to
shield the lamp+cavity device from the remaining of the board:

.. image:: {filename}images/freq_ref/efratom_enclosure.jpg
   :alt: The enclosure of the Efratom LPRO-101 frequency standard.


The Rubidium Freq Ref seems to workjust fine. I've powered it a few
times, and it take a few minutes to lock the lamp and stabilize the
output frequency. I've made a quick measurement of the power
consumption; at startup time, the current goes up to 1.5A, but after a
while, it stabilize around 0.5 or 0.6A. According a voltage 24VDC, it
starts with 36W then goes down to 15 or 16W.

Now, I have to embed it in a nice enclosure (preferably an aluminium
one so I can use is as a heatsink, since the frequency reference
produces a bit of heat.)

The Extron Video Amplifier
==========================

The video amplifer on the other hand has a bit of a trouble:

.. image:: {filename}images/freq_ref/extron.jpg
   :alt: the Extron Electronics video amplifier
      
As you can see on the picture above, the inductor is broken, and there
is no visible mark on it. This inductor is used to produce the
symetric voltages required to poser the opamp used to dispatch the
signal.

The switching power modules used on the board is a LT1616_ (marked as
LTNB on the S6 package). So I reversed engineered a bit the schematic
of the power to try to guess what kind of value this double inductor
could be, since the reference schematic in the datasheet does not use
such a dual inductor(/transformer).

Fro the schematic I could extract, it appeared that the design is in
fact very close to the reference schematic proposed by LT in the
datasheet_, in the Bipolar Output DC/DC Converter configuration page
16:

.. image:: {filename}images/freq_ref/lt1616_bipolar_output_DC2DC.svg
   :alt: LT1616 Bipolr Output DC/DC converter.

It produces +/-5V from the 12V input. But still, I did not know what
kind of inductor/xformer I should use. I did not notice at first sight
the two references for the inductor that are proposed in the reference
design. So I first searched using the nice faceted search tool of
digikey (reducing the selection according the informations I have on
the broken device and then looking at the pictures), I finally could
find the reference of the inductor. By the way, it's a Sumida CLS62
Series inductor (as indicated in the datasheet). And according to the
reference design, it should be a 22µH one. So let's make a purchase
for this inductor as well as some 20k resistors for my HP8904A_.

.. _HP8904A: {filename}hp8904a.rst


.. _LT1616: http://www.linear.com/product/LT1616
.. _datasheet: http://cds.linear.com/docs/en/datasheet/lt1616fs.pdf