During the ARRL 160 meter contest my Drake L7 kilowatt amplifier failed. It happened while I was checking email during an off period. There was an almighty bang and the amp went dark. Out of the corner of my eye I saw a flash of light in the darkness under the operating desk where the power supply is located. I didn't leap too far out of my seat but it was startling.
The problem was easy to diagnose. The amplifier is ~40 years old and has the original filter capacitors in the high voltage power supply. Electrolytic capacitors have a finite lifetime, especially high voltage ones of an earlier generation. I made a note to replace them at some point. Of course I didn't.
A temporary repair to route around the failed capacitor was attempted so that I could run the amplifier at its lower B+ setting. It didn't work properly so I continued the contest without the amplifier.
There are two filter/rectifier boards in the power supply, one for each side of the full wave rectifier. The large cylindrical parts are 220 μF 450 VDC electrolytic capacitors wired in series to give 55 μF at 1800 VDC. They see half the 2800 VDC of the no-load plate voltage (B+).
You should have no difficulty identifying the failed part in the picture above. I cleaned the power supply and surface it was on of the solid mass and liquid electrolyte that escaped from the ruptured capacitor. The material is not dangerous to clean up if you are careful to wash your hands afterward.
Pricing of the individual parts is not high but inconvenient to order and would not easily fit on the original PCBs. For a modest premium I ordered the Harbach Electronics PM400 kit that includes all the parts and one PCB. Modern electrolytic capacitors of the same rating are much smaller so it all fits on one PCB.
The kit is excellent. I had heard good report of Harbach's amplifier kits and I was not disappointed. The parts and PCB look excellent and, perhaps most important, there are detailed instructions for installing the new board in vintage equipment like my L7. The instructions proved accurate, right down to the length and colour of wires in the power supply.
After I assembled the new filter/rectifier board it sat on my work bench for a couple of weeks. It was the holidays and my free time was spent on a variety of other projects. Then the ice storm hit. I was so relieved after repairing the 80 meter array that I dove in that very evening to install the board in the power supply.
Installation was quick and went smoothly. I carried it upstairs to the shack and plugged it in. It worked perfectly. My first QSO using it was ZC4UW on 160 meters, with whom I had been unable to complete a QSO running 200 watts.
Voltage and other operating parameters are the same as before. The power output is limited by the plate transformer not the power supply filter and rectifier so there is no increase in power output.
The pair of 3-500 tubes is capable of more than the L7 delivers with the relatively low anode voltage and current capacity. I have no intention of replacing the transformer.
I am now ready for the CQ WW 160 contest later this month. Eventually I will have a second and more modern amplifier added to my station for high power SO2R and multi-op contesting. It made good sense to start with an inexpensive vintage amplifier. Despite this understandable failure the L7 has not disappointed.