MOV/TVS spark gap

[LaserTazerPhaser]

How would a MOV or a TVS diode perform implace of spark gaps in marx generators or tesla coils?

They can be stacked series to increase their breakover voltage. Gas discharge tubes have very limited ammount of cycles they can operate then electrodes wear out.

[MosherIV]

MOVs have a limited life as well.

[Gyro]

In addition to the [Edit: severe] life limitation, I suspect that the turn-on characteristics of a MOV would be too soft. A TVS will act as a Zener rather than something that breaks down to a very low voltage drop.

An air discharge gap with chunky electrodes should have a more or less infinite life, longer than the capacitors anyway.


EDIT: I do remember one application where a Thyristor was used in Avalanche mode (floating gate) to replace a GDT in a repetitive capacitor discharge application.

[Etesla]

Another thing to consider is that a spark gap flashes over at voltage X, and continues conduction down to a small fraction of voltage X (it conducts for an overall long period of time). A TVS diode or MOV will stop conduction very soon after they start conducting if put in the configurations you mentioned (they conduct for an overall very short period of time). Because of this, I would guess that you would get very poor performance from a TVS diode or MOV in the two circuits you mentioned.

[T3sl4co1l]

They don't.

Except possibly under extremely intense conditions (starting with a spark gap anyway, to generate extreme dV/dt, then putting the MOVs or TVSs in a distributed transmission line structure), they have positive incremental resistance, so they just burn the power applied, no sharpening, no triggering.

A SIDAC or any other 4-layer semiconductor can be used instead, but these are typically slower, and relatively fragile in that they are made of semiconductor; the current flow may not distribute evenly within fractions of a microsecond.  (They're usually rated for surge, getting similar ratings to other diodes and SCRs, good enough to clear a fuse in many cases, but still nowhere near the energy absorption of pure plasma, give or take how much electrode damage you consider acceptable of course.)

Tim

[james_s]

A friend of mine used to use TVS diodes to trigger a triac to make HID lamp igniters, apparently he found them to be more robust than diacs.

Don't MOVs work by punching a small hole through an insulating oxide each time they are used? I remember playing with one years ago and finding it only worked for a few cycles before shorting permanently.

[T3sl4co1l]

MOVs work by zener/avalanche between semiconducting grains; as far as I know, you get the same excess-local-heating-damage effects, but there may be something more specific to the chemistry used.  The statistics are like a population of small TVSs with varied breakdown voltages; a few junctions tend to carry the most current at low voltages, and are probably the first to be damaged at high voltages.

Lifetime curves are typically provided and indicate the maximum energy rating should be basically a one-time event, but smaller surges and lower currents can be withstood for thousands or N-->infty events.

Tim

[MagicSmoker]

One other thing is that both a MOV and TVS will exhibit a much higher voltage drop while conducting - see the clamping voltage spec in the datasheet - and this will obviously waste a considerable amount of energy.

However, avalanche diodes and light-activated SCRs (LASCRs) can sometimes be used instead of spark gaps.