Arcing in Tank Circuits of HF PA
This test circuit is easy to duplicate, and demonstrates the extreme voltages that appear in an amplifier when the load is improper for the amount of drive power applied.
Years ago, almost every amplifier had a output transformer. The transformer was similar to tanks circuits in RF PA's, in that it transformed or matched impedances. Even though Q was very low, the transformer could still store and release energy.
Experienced people knew if the volume was turned up too high in a transformer-coupled output stage, and if the load was absent, voltage across the transformer would soar to many times the normal operating voltage. No load or grossly mismatched loads often resulted in damaged transformers, blown output devices, or other output circuit component failures.
Failures induced by load or matching faults would occur in conservative amplifier designs, where components would last years in continuous proper operation. An audiophile would NEVER think of operating his expensive tube-type amplifier at anywhere near full volume into an open load, let alone a load where grossly improper impedances are selected in amplifier output-transformer taps.
Blame was never placed on stability. Everyone knew and understood conservatively designed well-constructed amplifiers with energy storage systems of any type, even very low Q systems, would still produce over-voltage failures when grossly overdriven or improperly terminated.
TV manufacturers used this well-known effect to advantage in horizontal output sections, where a flyback transformer with moderately low Q would produce many times the actual turns ratio in peak voltage, because of energy storage. Even modern switching supplies and our automobiles depend on energy storage to produce entirely new voltages, far above supply voltage, without requiring parasitics or high Q.
RF systems are certainly no more immune to mismatch than audio amplifiers, they very often are much worse. RF circuits are generally single-ended, and tube-type amplifiers have moderately high-Q (efficient) energy storage tank systems. Single-ended amplifiers with conduction angles under 360-degrees almost always contain intentionally designed “fly-back” systems, where tank circuit Q re-creates the missing portion of a sine wave from the half-cycle (or less) tug of the single-ended output device.
Somehow we have forgotten all this, and allowed ourselves to be misled into believing it takes a circuit or design flaw producing an oscillation to cause an arc or component failure.
This article demonstrates how easy it is to produce very high voltages from normal perfectly stable PA's with normal tank systems.
It isn’t safe to poke around in a high-power vacuum tube amplifier while looking at voltages, but a simple demonstration circuit can be constructed.
C1 is driven with a signal generator, L1 is selected to match the FET output to a 50-ohm load. The FET is operated at low current and has a series fast-switching diode, to simulate a one-way conducting vacuum tube.
This system was matched at 1.8MHz using the return-loss function of a network analyzer.
0dBm drive was applied with a signal generator, and the resulting waveforms appeared:
The upper trace is the output at D1 anode (point “B”). The black-dashed line was set at zero-volts. The scale is 10v/div.
The lower trace is drive voltage (point “A”). The scale is .1volt per division.
From this we see peak drain voltage swings approximately 20 volts, from around 5 volts to around 25 volts. This would be normal operation of a PA stage.
If we increase drive power and overdrive the PA, we get the following voltages:
We now see the peak drain voltage is almost 50-volts from a 15Vdc supply! In a 3000V PA stage, this would be 10,000V peak! It should make sense that amplifiers arc from grossly excessive drive power.
The next question would be what makes a normally driven amplifier arc. Often is when the load is inadvertently disconnected, either through poor relay timing, a bad cable or connection, or perhaps a failure in a component between the antenna and the amplifier.
Here is a scope picture with normal drive, but the load disconnected:
The drive voltage has increased slightly because of feedback through the FET, now that the drain is swinging wildly almost 70 volts, all from a 15-volt supply. This would be the electrical equivalent of 14kV on the anode of a 3-500Z operating from 3000Vdc!
It’s easy to see why perfectly stable HF amplifiers, if overdriven or operated at moderate drive levels under conditions of a load fault, can be damaged by severe arcing.
Virtually all PA arcs (other than those caused by component failures or reduced voltage breakdown from dust or contaminates) occur when the load is interrupted or mismatched, and the PA no longer transfers energy to a proper load.
The vast majority of PA failures are caused by improper operation or defective components, not by "strange events" that are unpredictable and non-measurable.