Shaun Merrigan's Information Pages


Racal 9478: Widespread Electrolytic Capacitor Failure


Posted on May 25th, by Shaun Merrigan in Electronics, Restoration, Troubleshooting. Comments Off on Racal 9478: Widespread Electrolytic Capacitor Failure

The Racal Dana Model 9478 Frequency Distribution System has an internal precision, oven-controlled crystal oscillator, or it can be phase locked to an external signal derived from a GPSDO or other master oscillator.  The system will automatically lock to an external signal above 100 mV, and with a standard frequency of 1, 5 or 10 MHz.  “In Lock” indication is provided.

I recently repaired two Racal 9478 units. What I found was widespread electrolytic capacitor failure such as I had never encountered before. A summary of the issues found in both units follows:

  • Virtually all electrolytic capacitors in both units (nearly 90 in all) had failed, showing either very low capacitance and/or very high ESR. Interestingly enough none showed short circuit failure.
  • All the failed electrolytic capacitors were made by Philips in late 1980’s
  • The -15v regulator was originally bolted to the main board which was used as a heatsink. This resulted in visible heat damage to the main board.
  • The internal OCXO was off frequency and the PLL would not lock to either it or to an external source.
Racal 9478 Front Panel with modified monitor output
Racal 9478 Main Board Before Replacement. Note the heat damage to the main board around the -15v regulator.
Racal 9478 Electrolytic Capacitors as received before replacement
Racal 9478 Electrolytic Capacitors as received before replacement
Racal 9478 Electrolytic Capacitors as received before replacement
Failed Capacitors from two 9478 units
Failed Capacitors from two 9478 units
Racal 9478 some Electrolytic Capacitors replaced
Racal 9478 Main Board after Electrolytic Capacitor Replacement
Racal 9478 Main Board after Electrolytic Capacitor Replacement

I completed the following repairs to both units:

  • All electrolytic capacitors replaced (nearly 90)
  • -15v regulator replace and heatsinked off the PCB
  • Q13 transistor heatsink added
  • PS voltages and ripple checked
  • Internal frequency standard checked and adjusted
  • 10MHz Monitor Output BNC repurposed give the output of the 5MHZ internal standard at the front panel

These units are particularly useful as they can be phased locked to an external standard. I lock the units to either a GPS disciplined quartz or GPS disciplined Rubidium oscillator. This lets me have a reliable, accurate “house” frequency standard which can be distributed to various instruments.

I will certainly be on the lookout for 1980’s Philips electrolytic capacitors and check them carefully.

Thank you for reading.





Comments are closed.