content/hp5334a.rst

Fri, 02 Mar 2018 01:12:57 +0100

author
David Douard <david.douard@logilab.fr>
date
Fri, 02 Mar 2018 01:12:57 +0100
changeset 91
d2632122f838
parent 44
96b26fb52f0b
child 115
6b6e13653348
permissions
-rw-r--r--

add src to hgignore

=================================================
 Quick overview of the HP5334A Universal Counter
=================================================

:author: David Douard
:Category: Electronics
:Tags: test equipment, 10MHz, Rubidium, hp5334a, counter


The HP5334A is a decent universal counter. It has a two 100MHz
channels. Some models come with an optional third channel (1.3GHz), an
optinal DVM, and an optional oven controlled
oscillator. Unfortunately, my meter has no option at all.

.. image:: {filename}images/hp5334a/front.jpg
   :alt: My HP5334A Universal Counter
      
There is a nice quick instruction set printed on the top cover:

.. image:: {filename}images/hp5334a/top.jpg
   :alt: HP5334A Quick Instruction Set


On the rear panel are 2 BNC (ref in/out and gate), the power socket
and the GPIB connector. One thing is a bit annoying on this device,
it's the fact that there is only one BNC for both the external 10MHz
reference input and the internal 10MHz reference output, with a small
Ext/Int switch to select the mode. So I have to reach the rear of the
device to switch this later when I want to use my `Efratom 10MHz
reference standard`_ (since I don't intent to keep it running every
time I want to user the 5334).

When removing the covers, one can see:

.. image:: {filename}/images/hp5334a/pcb.jpg
   :alt: The top view of the PCB.
	   
with a very nice star-shaped ground lattice. Among other interesting
things are these funny little variable caps:

.. image:: {filename}/images/hp5334a/c89.jpg
   :alt: C89, a nice little variable cap.

And as one may expect in such a device, no fan.
	 
The bottom side of the PCB looks like:
	 
.. image:: {filename}/images/hp5334a/pcb_bottom.jpg
   :alt: The bottom view of the PCB.
	 
Adjustment
==========

My 5334 does not have the high stability timebase option (Option 010),
so my counter's stability specs are:

- aging rate < 0.3 ppm per month
- tempco < 5 ppm (0°c to 50°C)
- line voltage < 0.1 ppm for 10% change

And it was not so out of specs:

.. image:: {filename}/images/hp5334a/uncal.jpg
   :alt: Measuring a 1kHz reference

In this picture, the signal is a 1kHz sine wave generated by a
HP8904A_ using my `Efratom 10MHz reference standard`_ as 10MHz
reference, so it should be pretty close to a solid 1.000000kHz. The
gate time delay is the default 0.3s.

On this device, the crital oscillator adjustement variable capacitor
is accessible directly from the rear panel. I guess that without the
high stability option, the 10MHz time base was not part of the
calibration process and had to be adjusted quite often.

So I've used the rubidium reference standard to adjust the internal
10MHz oscillator and I have, for now, a mush better 10MHz internal
clock (here we measure directly the 10MHz signal, and the gate time
delay has been set to 1s, as described in the service manual when
adjusting the oscillator):

.. image:: {filename}/images/hp5334a/cal.jpg
   :alt: After adjustement against the rubidium reference standard.

The adjustment capacitor is not fine enough to reach a spot 10MHz
value, the best I could achieve is a 0.2Hz away from the expected
value, which is quite excellent, actually.

But let's see within the next few days if it drifts a lot or not (as I
am writing this, it has moved a tiny bit to 10.0000005 MHz... 0.5Hz
off, something like 0.05 ppm).


.. _`Efratom 10MHz reference standard`: {filename}/10MHz_ref.rst
.. _HP8904A: {filename}/hp8904a.rst

mercurial