High voltage arc gap design

Are there any signs of pitting, uneven surface area? If it ligitimately arced once and pitted it may become more subceptable to lower voltages. I would expect to see some sort of pitting showing there was in fact an arc, otherwise I would suspect the arc detection circuitry. If it is pitted you might try carefully restoring the surface with fine sandpaper or equivilant.

Air breaks down at about 40kV per inch so you would expect a 3/4" spark gap to flash over at about 30kV, so the 3/4" gap for a 32kV supply would seem too small.

What does the transmitter manual say for setting up the spark gap?

Does this spark gap protect the power supply or is it at the base the antenna?

is the point smooth and rounded like it should be?

Do you have Debris flying around like dust particles

is it possible you have rodents running around getting in there and tripping it off? :))

We deal with over 1 Mvolt units at work and one the problems is sharp edges due to a valid arc at some point. Also, air quality has a lot to do with it.

The round tip designed are more prone to this problem via the ball tipped types. Static build up in the air can cause a build up of electrons on the tip and discharge much like how it does on an antenna which is why they use ball tips to help reduce the static noise.

Just a thought..

Not using a dry gas certainly adds to the variables.

You might check for any surface quality issues on the electrodes - any pitting, surface defects, or tendency to be pointy rather than rounded and smooth will reduce the holdoff voltage. As drawn, it does not look too terrible ("points" not actually pointy).

Hi David,

The other posters have already mentioned cleaning/polishing the surfaces of the gap.

However, because the smaller radius electrode is positively charged, this design will also tend to more easily form positive streamers, and will thus be more prone to flashovers, than a simple sphere gap that uses a pair of larger equally sized electrodes. If your drawing was approximately to scale, the positive electrode diameter appears to be about 1/2 - 5/8" in diameter, and the radius of curvature (ROC) is perhaps only ~0.75 cm. When operated at 30 kVDC, the E-field at the tip of the positive electrode is well above threshold to initiate positive corona, even if the electrode is polished.

I suspect that the design is causing the problem. A sphere gap, using a pair of 2" diameter electrodes will provide more consistent performance. You can use a standard spark gap table to calibrate estimated breakdown voltage such as or IEEE 4-1995 or the table below:

Bert

I'm curious--what's the radius of the tube and of the "points"? Where are you (what elevation above sea level)? What's the temperature of the air around the gaps? At sea-level and 25C, 0.75" is not enough for 30kV and needle-point gap, and could be a problem if the radius on the tips is too small; and of course, at higher elevations and higher temperatures things get worse. Sams' "Reference Data for Radio Engineers" has a section on gaps in the "Miscellaneous Data" chapter.

Cheers, Tom

Are you sure the arc event is happening at the base of the tower? I am assuming this is an AM Station? What make & model transmitter do you have? What type of antenna?

If all else fails, replace the tecnology! Ditch the present spark gap and install a proven type arestor. Check the usual vendors that utilities use, G.E., Westinghouse to name a few. Buy an arrestor rated for the same voltage, 32KV and all is good. No second thoughts of an original miss - design.

Well, i would not worry about it if the transmitter was in the US as the FCC Gestapo has outlawed it; drop dead date coming up fast.

------------------- You have hit it right on the head.

Then you know something no one else does. Tell us when the FCC is shutting down terrestrial radio broadcasting.

List make and model of transmitter. List operating power out. Is this a very old unit or a modern one? what year made? Is it stock or has it been modified?

Step 1: clean transmitter. Blow out dust with compressed air, use dry rag and paintbrush then wipe with distilled water and/or rubbing alcohol. clean all insulators and capacitors. clean transformers and bushings. replace air filters. inspect door interlocks.

Step 2: operate onto dummy load, if no trip then start investigating the antenna system. If fails while on DL start troubleshooting the fault sense circuit.

Observe all safety procedures. If you don't know what these are make will out prior to step 1.

Read manual and contact tech service

Step 3: get power company to monitor the AC for a few days. Inspect TVSS surge suppressor. Inspect main circuit breaker and replace as needed.

Step 4: buy and install new tube

All else failing, you could just measure the arc-over voltage: ramp up high voltage through a resistor and see when and where she snaps.

That'll give you an answer that covers both questions of gap-setting, and of gap & surface conditions.

Why wonder when you can measure?

Cheers, James Arthur

Oh, another question: any fried bugs piled up beneath this thing?

Cheers, James Arthur

;-) At that power level, they tend to vaporize rather than just fall down dead. Used to have a problem with flies getting into the air- variable neutralizing capacitor in a 1kW AM broadcast transmitter, and vaporizing themselves while drawing a pulse of current large enough to engage the over-current relay. Things got more exciting one day when the protection circuits didn't. One of the results was a track melted in the ceramic core of a 200W wire-wound resistor, molten ceramic dripped down on the deck below. Didn't find the fly carcass, though.

Cheers, Tom

That's because the fly made an ash out of itself. ;-)

Molten ceramic? That's good stuff! (But don't show this to Tim Williams ;-)

Grins, James Arthur

Yeah, Tim will demand a royalty payment from him. ;-)

Feh. Examine the remains of three 20 kV lightning arrests mistakenly connected across a 66 kV transmission line.

The fireball looked pretty cool too. But not to the lineman who was standing practically underneath them closing the switch.