Have a look at this picture showing a seaside pier funfair decorated with neon lighting. Another picture from the outside shows the spectacular display of light created by loads of neon tubes all around the sides.


But unfortunately when lots of high voltage lighting is installed on an old wooden building it can cause illumination of a different kind - see here   
(That is the same place as the previous two pictures)

WOW!!!!!! 

Holy Crap!!! 

wow.

  Yikes

As much as I love neon, I have seen some bad fires caused by it. I've always been of the assumption that poor maintenance caused many of these fires.

The last two fires in my area involved neon, and in both instances were replaced by fluorescent behind perspex. 

The problem is not with maintenance, but with the fact that whoever installs those high voltage tubes does not give them the caution they deserve

Neon tubes work at 1000-3000V or so. Those are voltages at which many odd things happen :

 - Arcs up to few mm long can happen in open air, and be very stable. Unlike 240V arcs which are rarely over 2mm long, and nowhere as stable. There is a reason why its called high voltage

 - The transformer is a limited current source to about 30mA. When you have a short circuit in 240V, there isno much limit on the current - there will be a several KA bang, and tripped breaker. When its 1KV at 30mA it is a stable burning arc dissipating 30W. Its about the same as a constant flame burning in there instead of the arc, would be about the same amount of heat. Burning for as long as it takes to set something on fire, no breaker tripping whatsever

 - Electrolysis and arc tracking effects go on nicely at such voltages. If 240V takes ages to electrolyse wood, 1KV+ is arcing through wood as if it would be metal

 - Voltage is high enough for electrostatic effects to take place (though to a lower extent because its AC) : fine dust accumulating in the way for a flashover etc, making over time places where a flashover is going to be imminent




And the story of a fire comes down to this :

Designer wanted to make a sign easily, so he used a wood backboard and painted it and installed the tubes

The tubes worked for many years, somewhere a silicon cap (on the connection of wire to the tube) snapped due to old age and exposed the wire

Some dust accumulted there making a conductive path to the wood back board

The dust flashed over and the arc kept going on,, maybe even unnoticed - electrolysing the wood and burning it up to carbon, The arc keeps going on and dissipating the heat in the crbon

Carbon ignites

Fire



So basically :

Pick the corect materials for the job and dont use materials which are flammable or prone to arc tracking. Wood is no-no. Plastic is not too good either but at least it is not outright conductor for such voltages

Avoid "depending" on materials that are going to degrade over time. Installing the sign outside in sunlight ? Use a shrink wrap over the wires for extra UV protectn, use longer stand offs for the tube

Dont let the sign accumulate dirt, and maintain it

Because a neon sign transformer, is current limited, it can be used to make "Jacobs Ladders".

And dont forget that, its output voltage is sufficient to strike the arc in the jacob ladder. Not something you'll be able to do with a fluorescent ballast. (Though maybe you can with the addition of HPS ignitor)

I remember finding a report on that fire somewhere on the Internet. I always thought the fire was caused by too many stuck starters overheating banks of remote ballasts. I remember seeing pictures of banks of (fluorescent?) ballasts in the corner towers. I didn't, however, read the report in it's entirety. Malfunctioning neon lighting could also very well be the cause.

How neon signs can have an operating voltages of several KVs? Isn't the 1-3KV is for starting them?

A quick web search revealed that the Grand Pier was rebuilt, and the cause of the fire is still before the courts.

The tubes clamp the voltage down to their arc voltage, but it is still in the KV+ range. When the circuit is open (broken wire, broken lamp....) the voltage becomes full OCV



Overheating ballasts will eventually short out and trip a breaker, Will something ignite by then ? Thats what i would ask. What can catch fire actually ? Wires isolation. Backboard (if its made of wood). Tar from ballasts. Dust. Flammable materials put on top of the cabinet by users.... Each of those is the question

Besides, it costs very little to connect a thermostat in the ballast bank. Apparently nobody thought much about that all those years ago



In the 60's there was a documented fire in Zim company's building, Tel Aviv, Israel. Quoting from the Wikipedia article (with a picture inside btw)

The fire, which was one of the worst in Israel's history, spread, among all reasons, due to the plywood walls between the offices, and the electric system being illegaly installed in the wood boards of the ceiling. The reason for the fire is assumed to be a fluorescent lamp ballast

I have some more to expand about this :

Most fluorescent fixtures used at the time were open-back. I have a fixture like that in my collection. This is how an open back fixture can start a fire if it is installed on a flammable surface (which, given the description of the wiring, can perhaps be the case)

The ballast in it is a wound coil of wire, not layered accurately inside, isolated by some sort of cardboard (probably impregnated with something). All this is located quite deep between the plates forming the core and outer metal wraparound of the ballast. It would take very little to cause a short of  loop in the coil, which would probably make the ballast heat up and this loop get to the temperature at which it might ignite something. But i dont know whether its likely that the fire/smoldering would get all the way out to the open edge of the cardboard roll, where it can ignite stuff outside the ballast

For all i know those ballasts were cheap, noisy, but reliable and even did withstand quite well stuck starters. Failures with them were rare even after 50 years in use

Those fixtures did have other weak points :

The design of the starter socket is flawed. There is "nothing" (other than the springyness of the contact) that prevents the contact of the starter socket from shorting to the screw that holds the socket to the fixture (earth). Normally the distance between them is about 1mm. Which is quite little. If in some fixture  this happened to be closer than in others, or if some dust or moisture is present, it can arc in this point. Due to the contact prone to arcing being the "live" side of the starter, the circumstances are excellent : The ballast limits the arc current so no breakers are tripped, and the starter provides instant restriking of the arc if it goes out on its own. As the back of the fixture is open and the spot of arcing is on the back of the starter socket, this is few mm apart from the surface to which the fixture is mounted

There are large exposed live contacts in there in general. Especially the same connection on te back of the starter socket. If there is something in the fixture which should not be there (sawdust, steel wire poking from the concrete,....) it can be the cause of arcing or electrolysis on its own

Connecting wires at the time was usually done by wire nuts. The wires of the ballast are flexible, and they dont work well with wire nuts. A loose connection could have heated and arced

Most of those fixtures were without capacitors. But if there is one, it can possibly fail violently as well

A rare and unlikely way of failure which can happen with closed back fixtures too, is if the fixture is installed on wood, unearthed (as many lighting installations at the time were), and there is earth fault in the fixture - The voltage applied to 2 points in the wood can lead to electrolysis and carbonization, though it is not likely (it does not happen not nearly as readily as with neon sign voltages)

Other thing that can be blamed is general badly done wiring, which as in the description, was all over the place....

As for poor installation, I`m not sure about other parts of the world, but neon tubing here is installed by specialized companies and not by your average domestic / industrial electricians.

I`m guessing all the fire precausions are taken into account with each neon installation.

Problem is, not all electricians / neon technocoans / whoever is familiar (or cares) with some small factors. Many of them think like "hey, it works now". Not "What will happen when this plastic isolator snaps after few years, will it make a likely condition for arcing"