Author Topic: HP5334A replacement oscillator  (Read 2074 times)

0 Members and 1 Guest are viewing this topic.

Offline babysitterTopic starter

  • Frequent Contributor
  • **
  • Posts: 900
  • Country: de
  • pushing silicon at work
HP5334A replacement oscillator
« on: December 16, 2014, 09:21:43 pm »
Hello board,

for a while I own one of these fine HP Universal counters I bought off ebay. The packaging was crappy, with the front BNCs pinched thru the single-layer cardboard box, but there was no transport damage after all.

Damage, however, was present as burned resistors in the 50 Ohm termination of Channel A. It was not the first instance of blown resistors there, as someone repaired the resistors in the PCB simply by stacking new resistors on top of the bad ones! Also, another resistor in the ARM Path was blown also. The missing cover of the voltage selector was made out of a piece of PVC angle profile, cut and folded to shape with hot air support.

The challenge

One disdvantage of this otherwise nice counter is the local reference oscillator which is not as nice as the rest, so I decided to put something better in. It was really frustrating seeing the last digits going all over the place from a reasonable stable source.

The original OPTION 010 was not available at that time on everybodys favourite T&M auction site.
So I went for a OPTION 009.999999999

The first solution

Help appeared as a STP2390C from ebay. After a first sanity check of the OCXO and figuring out how to use a oscillator at the OPTION 010 port, it went in... connected by a terrible wire salad. With a 7812 and his free-floating capacitor friends  screwed to the chassis. Something to proudly hide. But giving good results on the readings.

The improvement and the drawbacks

As a christmas project I decided to beautify this hack.
A PCB which fits into the option socket J8 shall give a home to the OCXO and the voltage regulator. I decided to do it double-sided trace, double-side THT mounted board to keep it small and reduce the stress on the connector. Besides the immediately necessary parts, I prepared it to house additional stuff like a choke for supply voltage and a series-resistor that might go between the Potentiometer and tuning voltage input.

Being a true Über-green german (No I am not, but worried about my electric power bill) I hooked it up to the +15V rail that is switched with the mains switch, so I have some warm-up time every time. The original OPTION 010 is powered from a always-on +30V rail.

Having children running/crawling around your legs while quickly making a PCB introduces a lot of errors, so it came that I switched two pins. Oh well, I realized hours after sending the file off to the factory!

But the end result is somewhat nice, even after I had to use the drill press, scalpel and enameled cooper wire to fix my botch.

The penalty for working too quick and distracted

The backside picture shows most of the workarounds.  I swapped the tuning voltage input and ground. The ground hole was killed with a 4 mm drill, its fed by the enameled cooper wire link now. The hole wrongly dedicated to the tuning voltage got a solder bridge to ground. Also, the oscillator didn't have a reference output on the pin, so the tuning pot is connected to supply voltage now.

picshures, picshures

The two pictures show the assembly before adding a thin foam rubber tape around the OCXO module to keep away air currents.  The potentiometer, linked with a short run of 3-wire cable, is mounted in a hole in the backside which is adequately labeled as OPT 010 OSC ADJ.

Adjustment and Outcome

A Rubidium source was powered up to get a wonderful 10 MHz signal to compare to. First step was adjusting the OSC ADJ Potentiometer to read a perfect 10 on frequency counter mode. That was easy.

Then, the  time interval counter mode was selected, Internal reference (available on back-side BNC) to START Channel, Rubidium Standard to STOP channel and tweak the potentiometer for slowest rate of change of the reading, which means that the phase difference drifts slowly. This is comparable to the well-known method of adjusting the phase difference drift with a O-Scope. (Hello W2AEW!)

I'm happy with how it became, just a little bit sad that the christmas project was finalized shortly after PCB arrival, way before christmas!

Long term
Why do I have such a lack of HP feet? Any hints for cheap sources welcome!

Shameless advertising

I have a few leftover PCBs... want some ? :) The Minimum Panel size resulted in 11 pcs.


« Last Edit: December 16, 2014, 10:31:07 pm by babysitter »
I'm not a feature, I'm a bug! ARC DG3HDA
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf

 

-->