Triggered spark gap design

Imagine two metal electrodes separated by gas. Gas doesn’t conduct electricity so there’s no way for current to flow between the electrodes. That’s a spark gap in it’s nonconducting state. There’s a strong electric field between the electrodes when high voltage is applied and this electric field will accelerate any charges between the electrodes due to the Coulomb force. An errant electron will feel the pull of the electric field and accelerate until it hits a gas molecule. If the electron has enough kinetic energy it will knock another electron out of the molecule and leave behind an ion in a process known as impact ionization. Both the initial electron and the electron from impact ionization will be accelerated by the Coulomb force and may cause further impact ionization if they gain enough kinetic energy before their next impacts. This avalanche or Townsend discharge, where a single electron begets many, leads to an exponentially increasing electron density the result of which is conductive plasma bridging the gap between the electrodes.1

Recombination of electrons with ions fights exponential growth of electron density and prevents Townsend discharge at low electric field strength. When the recombination rate is higher than the electron growth rate, electron density doesn’t increase and so there’s no chain reaction and no plasma. Recombination doesn’t care much about the surrounding electric field, but the electron growth rate increases with electric field. That’s because as the electric field increases, an electron can gain more kinetic energy before it runs into a gas molecule and so is more likely to cause impact ionization. Impact ionization beats out recombination at a critical value of electric field strength described by Paschen’s law.

Paschen’s law is the boundary between increasing and decreasing electron density in a spark gap. This great divide with avalanche on one side and plain old gas on the other is a straight line in the pressure-voltage plane and is called the breakdown voltage. I guess you could also call it the breakdown pressure, but people will look at you weird. Paschen’s law is a phase diagram demarcating the gas phase from plasma and since the relationship between voltage and pressure is linear2 it’s easy to predict where a phase transition will occur: double the pressure and the breakdown voltage also doubles.

Crossing Paschen’s line from gas to plasma takes either dropping the pressure or turning up the voltage. Either path will take you from teraohms to milliohms in nanoseconds. One path is easy and reliable but not good for synchronization. The other takes some effort but gives you sweet milliohms exactly when you want them.

Of course, there’s a third way and I’ll cover that first. A two-electrode spark gap can be used to speed up the output of a pulse generator. Spark gap pressure is held constant and you just wait for the pulse generator output to exceed the breakdown voltage predicted by Paschen’s law. This technique is commonly used to speed up the rise time of a Marx generator and a two-electrode spark gap used in this way is called a self-breaking or overvoltage spark gap. The term peaking switch is also applicable because the spark gap quickly connects the Marx generator output to the load after the Marx generator output has slowly risen. This means that the load sees a much sharper voltage peak than if the Marx generator were connected to the load directly.3 A two-electrode spark gap might also be found in the company of a Blumlein pulse forming line charged by a Marx generator where it is placed between the load and the Blumlein to suppress the prepulse that occurs as the Blumlein is charging.4

A peaking switch is a voltage comparator: when the voltage that the spark gap sees is less than the breakdown voltage set by the pressure, the spark gap blocks current, but once the voltage exceeds this threshold it turns on immediately. Thinking of the spark gap as a comparator and the pressure as the reference value it’s clear that high stability pressure control is needed for consistent output. That’s because if the pressure increases a bit then the spark gap will turn on later and at a higher voltage.

Pressure triggering is easy and reliable because you just let some gas out of the spark gap with a valve. Opening the valve drops the pressure and we cross the line from gas to plasma. Pressure triggering is especially convenient because the valve used for triggering is galvanically isolated from the spark gap: the only path between valve and spark gap is plastic tubing. Galvanic isolation makes pressure triggering ideal for high side switching. But, we pay for simplicity and isolation with high jitter. It takes milliseconds for pressure to drop after opening a valve so pressure triggering is limited to millisecond synchronization. This limits pressure triggering to applications like lifetime testing components where timing isn’t critical.

Pressure triggering and overvoltage operation both have their niches, but it’s darn convenient to tell a spark gap to jump and have it reply with “How soon?” That brings us to three-electrode spark gaps which have an extra electrode just for triggering. Even though two-electrode spark gaps can be triggered, the extra trigger electrode means that three-electrode spark gaps are usually called triggered spark gaps. There are many possible geometries of trigger electrodes. The two most common are the midplane and trigatron.

A midplane triggered spark gap has a trigger electrode in the form of disk halfway between the two main electrodes. Two resistors are used as a voltage divider between the main electrodes with the trigger disk tied to the middle of the divider. This puts the disk on an equipotential surface of the electric field that would exist even if the disk weren’t present, so the disk does not distort the electric field between the main electrodes. In other words, adding the disk doesn’t do much as long as it’s tied to the middle of a voltage divider.

Changing the voltage of the trigger disk to any value other than halfway between that of the main electrodes distorts the electric field and causes Townsend discharge. Hence the midplane spark gap’s secret alias: field-distortion triggered spark. Let’s say we drive the trigger disk positive. Then breakdown will occur between the trigger disk and the negative main electrode because the field strength in that region increases. Once the negative main electrode and trigger disk are connected by conducting plasma the trigger disk will find itself at the same voltage as the negative main electrode. This is like halving the distance between the main electrodes and doubles the field strength between the trigger disk and the positive main electrode.5 Breakdown is sure to occur when the electric field strength is doubled  and so the entire switch fills with plasma. Driving the trigger disk negative has the same effect even though the intermediate steps happen in reverse order.

Trigger polarity is much more important when talking about trigatrons. That’s because a trigatron doesn’t have the same geometric symmetry as a midplane triggered spark gap. There are three electrodes, but the trigger electrode is a pin inside one of the main electrodes called the adjacent electrode. Initially the trigger pin is at the same voltage as the adjacent electrode so the setup looks basically like a two-electrode spark gap. Driving the trigger pin with the same polarity as the adjacent electrode increases the field strength between the trigger pin and the opposite electrode. Man who comes up with these names? High field means breakdown, so in one fell swoop an arc crosses the entire distance between the main electrodes. This is the heteropolar option and, aside from the fact that it takes twice as much trigger voltage, performance is comparable to that of the midplane spark gap.

Here’s where we run into the importance of polarity to trigatrons. If we instead take the homopolar route where the trigger pin is driven with polarity opposite that of the main electrode it turns out that the trigatron becomes an unreliable friend. Homopolar triggering looks attractive at first because the short distance between the trigger pin and the adjacent electrode means that very little trigger voltage is needed to initiate breakdown between those two points. But, this doesn’t do much to the field between the adjacent and opposite electrodes so Paschen would say that there's no reason for plasma to fill that space. It turns out that some secondary factors related to UV, heat, and the plasma between the trigger pin and adjacent electrode can in fact cause the main spark gap to breakdown if the voltage across it is very close to the breakdown voltage, but triggering is unreliable with high jitter. Homopolar trigatron: just don’t.6

And that concludes our bus tour of spark gap design. As we pulled out of the parking lot we saw that Townsend discharge explains the process that transforms gas into conducting plasma. On the right was the surprisingly straight wall named Paschen’s law. We watched through binoculars as gas hopped over the wall to become plasma after having its pressure reduced. Finally stepping off the bus, we found ourselves at the gates of a zoo filled with galloping midplane spark gaps and trigatrons resting in the shade. I’m glad you came along. You can find me at the lemonade stand.