diff -r db7bbef8e263 -r 6ede33beeeed content/hp5334a.rst
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/content/hp5334a.rst	Sat Mar 12 01:36:48 2016 +0100
@@ -0,0 +1,96 @@
+=================================================
+ Quick overview of the HP5334A Universal Counter
+=================================================
+
+:author: David Douard
+:date: 2016-03-12
+: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