This picture shows Wafer C of the band-switch along with the Pi and L input networks, the large filament choke, ALC circuit components and the two valve sockets. The two grid bias resistors are of ceramic-composite type. I also installed bigger (both capacity and max voltage) silver-mica caps in the grid grounding circuit. This will lower the impedance to ground on the lower bands and should result in more output. The original caps can't handle a lot of current and often fail. The length of the leads must be kept to the minimum to avoid extra inductance. A metal-film low-inductance power resistor between the input circuit's rotary switch and the tube's cathode circuit is installed to limit the possibility for over-driving the amp. Before installing the input circuits, I made sure that all of the ferrite cores in the inductors are moving so I can align them later. Extra caution is needed since these cores sometimes get stuck or are glued inside the coils with wax and it is very easy to break the ferrite core when rotated.
Top view of the amp with the RF deck on the right and the HV power supply and power transformer on the left. Both variable capacitors were inspected for signs of arcing, cleaned and lubed. Any signs of arcing and sharp points on the blades must be smoothed with very fine sand paper and cleaned well from the dust afterwards.
Picture of the RF deck. The two original Centron 572B valves seen in the picture were replaced by a matched pair of RF Parts, Inc. 572B Taylor valves. The 10m/15m silver plated tank coil was cleaned with Deoxit D5 solution. Wafers A , B and C of the band-switch were inspected and carefully cleaned with extremely fine (0000 size) steel wool and Deoxit D5 and then treated with Deoxit ProGold. Using steel wool to clean an RF switch requires *very good* washing with alcohol afterwards, as well as using strong magnet and compressed air to get rid of ALL metal particles. The switch should absolutely clean from metal particles and contaminants before installation. After the cleaning, the switch looked almost as new. The output multi-band tank coil was encased in Kapton foil and all leads were inserted in teflon tubing.
The new Low-Q parasitic suppressors designed by Rick Measures, AG6K. The new design includes two sets of suppressors in series (per valve) for an improved performance and amp stability. Resistive Nichrom wire is used to make the inductors - the kit includes silver solder and special (very corrosive!!!) flux for soldering Nichrom. Metal-film resistors used in the suppressor can handle power dissipation in the excess of the rated 3W. An extra 1 Ohm resistor in the anode circuit of each valve serves as fuse in case of a catastrophic valve failure. 572B is a relatively unstable tube when it comes to parasitic VHF oscillations so these suppressors should help a lot.
Another view of the RF deck. I had to repair the plate choke with a new winding. The old one had badly burned wire enamel. The new choke is covered with kapton/silicone adhesive tape. It is barely visible in the top-center of the image. Right next to it is the new blue high-voltage plate filtering capacitor. The door-knob capacitors in the output circuit were inspected and one of them had to be replaced (it was showing signs of arcing). I used a layer of Kapton tape around this capacitor (the red one in the middle of the two variable capacitors) as extra precaution.
Update: I did also some work on correcting the filament voltage. It is posted here.