Any of the problems below can result in destruction of a bandswitch or other amplifier components.
Antenna tuners are a common problem. Many tuners are over-advertised, and will not handle the power claimed. Some models also have poor or unreliable connections in inductors. Many tuners have poorly designed power meters with peak meters that don't indicate true peak power. If you look at the antenna tuner review in the February 2003 issue of QST, you will see power rating problems exist with many tuners.
User adjustments also greatly control tuner power ratings. People adjusting "T" network tuners often forget to use the maximum possible capacitance required to match the load. Using less than maximum capacitance will reduce tuner power handling capability.
A very brief arc or open circuit in a tuner can cause a momentary very large change in antenna tuner input impedance. This large impedance change reflects back into the amplifier, and if the tank is momentarily unloaded a catastrophic bandswitch or capacitor arc will occur.
The worse possible change is one that results in a condition of high impedance at the amplifier tank input, in which cases tank voltages can reach many times the normal operating values.
Always be sure your tuner has reliable internal connections, and is tuned correctly (maximum capacitance that allows matching the load). Be sure the tuner can handle the peak transient power from your exciter-amplifier combination.
At one point, I owned an ICOM 775DSP radio. The peak transient power from that radio, even when set to 50 watts, was over 300 watts! While that pulse would not show on a normal power meter, it clearly showed on peak storage meters. My Bird digital storage, a special Coaxial Dynamics peak meter for medical applications, When I drove an 8877 amplifier with that radio, the few milliseconds of peak driving pulse would trip the very fast grid-current-fault protection system in my amplifier.
When I advanced the loading to prevent the grid trip from activating, the transient peak amplifier out was over 5kW. This would cause my antenna tuner, a large old KW Johnson Matchbox, to momentarily flash-over! When the tuner would arc, the amplifier tank would arc.
The entire system would have to be able to handle the exciter's 300-350 watt peak transient (amplified through the PA gain) in order to not have reliability problems!
Antenna and feed line Failures
Antennas can have arcing problems from bad connections or other defects. If an antenna fails to accept power or changes impedance intermittently, even if for a few milliseconds, it can trigger an arc elsewhere in the system. That point may even be in an amplifier tank system or in an antenna relay.
Traps can arc, as can other components like gamma capacitors or matching networks. Be sure antenna wires are not able to brush up against other objects in the wind.
Be sure all connectors and hardware are tight, and connections are dry and well-insulated. Normal UHF connectors, even though rated less, handle much more power than type N connectors. UHF connectors are much less likely to develop intermittent connections, and actually have LESS loss below 50MHz. The only potential drawback for HF use is UHF connectors are less water resistant than properly installed N connectors. Waterproofing can be cured with proper waterproofing.
Antenna problems will show if you use a peak-storage meter, or a meter with a long hang-time. The only amateur radio products I'm aware of that have enough response time and hang time are the Ameritron meters used in the AWM-30 meter and ATR-30 and 20 tuners. Any momentary faults in my antennas immediately show on the reflected power of those meters when they are in the peak mode.