Low Voltage Spark Gap

Look up transorb. A back-to-back large junction zener which is designed to absorb transient overloads.

Depending on cost and the circuit topology, we replaced the little radioactive gas discharge tubes with both MOV's and transorbs. There is some parasitic capacitance when added to any circuit, but I believe a transorb should meet your requirements. We preferred them over MOV's for improved performance and life. Good luck!

We have used tranzorbs in other applications as well. The only problem is those were for limited or one time use (carbon block). We have no way of monitoring the gap firing at this time, so a lifetime (10+ years) of service is what we are looking for. I don't want to design a separate monitoring circuit just to replace the gas tube.

Of course, the current devices have been in place for over 20 years, but we have no suitable replacement-duh-so they just stay in there, working or not (?). If I can find a direct replacement we may be able to sneak it in with a minimum of overhead and no need to design a monitoring circuit.

Limited use? What? Transient Voltage Suppressors (TVS), the generic name, are zener-diode avalanche devices meant for repeated use without degradation (unlike MOVs). They have sizable built-in chunks of copper to absorb the breakdown energy and are available with quite high power ratings.

mjb1962853 mentioned capacitance. This is an issue, especially for large high-power TVS parts. For a 60V rating, a small 600W part (e.g. P6KE62) has 200pF at the standoff voltage, and over 600pF at 0V; a large 5kW part (e.g. 5KP60) has 500 and 3500pF.

But there's a workaround, a way to have your cake and eat it too: Add a series diode. For the case of dealing with unipolar transients, a single diode with a TVS will do the job. You'll also need a voltage source to pre-bias your TVS (and back-bias the diode) so its high capacitance doesn't have to be charged from your protected signal. For bipolar transients, two diodes with two TVS units are required, as shown in the sketch.

Suppose your chosen diode is a 1N4004, it has less than 4pF of capacitance with 60V reverse voltage, a 150 to 1000-fold improvement over a bare TVS. A more serious 1N5402 diode (peak current 200A for a 8.3ms half-sine, more for shorter events) has less than 15pF. You'd double the parallel capacitance for the bipolar configuration, natch.

Always glad to hear your take on things, Winfield.

We utilize matched pairs of transzorbs in a different circuit (not from my OP) in a communciations system. Yes, they will continue to function as surge suppressors for many years, but degrade the circuit (data receive) to the point of needing replacement. Lightning strikes nearer our receivers tend to break them since that is outside of its ratings (1.5kc, 7.5 CA). Our signal strength is very low by design, so matched pairs are kept on hand and are quickly and easily replaced. We actually still have stock on hand from the 70's. The equipment was designed this way in the late 50's and is still in use today.

The equipment from my OP is the "newer" design from the early 70's (also still in use today, obviously). We do not have direct evidence that the gaps are even still functional. I am trying to get some testing done to find this out. We have no replacements on hand, so I have to be very careful to not let out the magic smoke.

As far as my OP goes, I am still looking for a "direct replacement", but a smaller circuit design may work as well, that's direct enough. The NRC issued our permit with specific instructions to not handle the alpha source, so I will need rock-solid material before I present my case.

The description from the assembly manual: Electron Tube: two-electrode bilateral type, uses a T-9 glass bulb and is encased in a permanent aluminum housing, type 458C. The power dissipation in the tube at normal system current is approximately 5W. Avg cathode current 5mA, surge cathode current 1500A, 1.5 millicoulombs, 60V breakdown, 10V sustaining voltage, breakdown time 5usec at <500V, 0.7usec<4kV/usec.

I have included a schematic with the tube labeled 458C between points E1 and E2.

The spark gaps are designed to divert MICROSECOND or MILLISECOND PULSES . . . typically <10ms - which is 1/2 period of 50Hz.

What is it that you are applying for 500 ms?? A AC - DC voltage in excess of firing voltage with substantial current behind it??

This would be a miss-application for the spark gap.

Capacitance of large MOV or TVS is also large. GDTs are chosen for their small impedance loading. The best would be to replace the GDT with one that does not contain the radioactive gas. I would suggest to search for an equivalent device. I would try if the same manufacturer (if still in business) is offering a non-radioactive alternative. If you need help, post a picture of the part and all the data that you have.