Constructional Notes for G8ACE MKII OCXO – April 2004

 

The OCXO is under continuous development and improvement and these notes detailed here supersede those notes found on my website.  The Web notes should still be referred to for details not covered on this sheet.  Go to: http://www.microwaves.dsl.pipex.com/mk2/pdf_files1.htm 

                                http://www.microwaves.dsl.pipex.com/mk2/pdf_files2.htm

 

When fitting components to the PCB start with the smallest parts first.  The only exception being the inductors L3 and L4  which can obstruct the positioning and soldering of the trimmer.  Some trimmers supplied have an extended metal tab which needs carefully shortening before fitting to the board  The threaded hole on the heater plate need a small piece of tape applying before epoxy is applied to prevent epoxy flowing into the hole and blocking the threads for securing the TIP31A heater transistor.

 

Initial Oscillator testing can be performed without the heater plate by following the instruction in the web notes.  Ensure you use an epoxy with suitable high working temperature when attaching the heater plate to the PCB.  The TIP31A and insulator, the thermistor and the crystal should all be smeared with heat sink paste to ensure good thermal conductivity. 

Inductors L3, L5, L7 are wound using the supplied 0.5mm wire on the supplied cocktail stick, nominal 2mm diameter, using the table of turns for the particular crystal frequency in use.  Once the inductors L5 and L7 have been squeeze tuned for maximum output all three inductors can be doped with Polystyrene Cement in preference to the Instant glue originally suggested.

 

A suggested layout arrangement for the socket and connections through the tin plate box again can be found on the web page URL above.  When drilling the box ensure a wooden or similar block is used to support the box side to prevent distortion.  The feed through capacitors should be threaded thro the small earth tags supplied to help support the feed through if the hole is slightly oversize.  The earth tag also serves for the ground connection for the service on that particular point.  The sma may need the insulation and inner conductor shortening before fitting to the box.  The OCXO module should be supported in polystyrene or similar insulation inside the box.  If the module is used in an outside installation then an extra external layer of insulation is beneficial.  The fine frequency control potentiometer can be routed back to the shack if required in this case but two core and screened cable must be used and the ground connection be made only at the OCXO and be fully floating elsewhere.  Otherwise heater current flowing in the braid will modulate the tuning voltage.  The OCXO can be used with the CT1DMK Reflock PLL and functions well with the Reflock  PCB mounted at one end of the OCXO box.  A prototype version of the integrated unit can be seen at:

http://www.microwaves.dsl.pipex.com/pll/mk2.htm

 

The OCXO will operate down to below 11.5v dc supply voltage to be trouble free when operating portable from 12v battery supplies.  In normal operation the dc supply should be stable.

The oven will tolerate a few hundred milli volts change without moving the frequency to any noticeable degree.  However larger fluctuations must be avoided otherwise the oven will need to re-stabilise.  A simple external regulator such as a 1 Amp 78L12 should be used for unstable dc supplies.  When the PCB is used with a home constructed heater plate the thermistor is best embedded in the plate however surface attachment works but if the thermal conduction or insulation is poor the oven gain set R23 may need adjustment.  This is very unlikely but if the problem occurs it will show as a cyclic movement of frequency over several seconds.  Typically 100 – 200 Hertz  at 10GHz.

 

If you employ a crystal supplied by  Quartslab in the UK or Klove in Holland then its possible it might show signs of LF jitter.  This really becomes noticeable for 24 GHz and higher bands.  It is a matter of luck as to whether this is a problem or not and seems determined by the quality of the crystal blanks used.  Round numbered crystal frequencies for 144 MHz IFs  tend to be better than oddball frequencies used for 145 MHz IFs.  Thermal cycling of the crystal between 80C and placing in the freezer can reduce any LF jitter.  An hour at each temperature for a total of at least ten cycles should see an improvement.  Jitter is often bad soon after soldering a crystal so do not jump to a rapid conclusion about the crystal quality or whether subsequent treatment has had any effect.  Always try another crystal if you can before assuming an OCXO fault when results are not to expectations.  

See: http://www.microwaves.mcmail.com/jit/jitter_reduction.htm

 

MKII OCXO pdf file web notes:   http://www.microwaves.dsl.pipex.com/mk2/mk2.pdf

 

G8ACE April 2004