I wouldn't think that 12 volts would be enough to strike the arc. Perhaps the "12 VDC inverter" shown can produce a high voltage to start the arc before supplying the 12 volts running voltage. But the capacitor in series method is fairly obvious.

I agree.

The cap applies 170 volts peak to strike the arc and then limits the current once the gas ionizes. The inverter does the same. It applies a voltage high enough to strike and then limits the current.

The lamp, itself, sets the final voltage drop in both the cap and inverter versions.

 

Try firing up a GTL3 with 10.5-12 volts DC and within seconds you'll be smelling ozone while a runaway arc eats its way along the filaments, the supports and eventually the bulb's quartz envelope. No high voltage needed. You might try putting a moderate wattage incandescent bulb in series with the lamp to limit the arc current. A standard mercury lamp requires 100-250 volts to strike the arc. Once ionized however, the arc path has a very low impedance so that 10-12 volts is enough to sustain discharge. As I mentioned in my post, the mercury vapor in the vicinity of the filament will ionize without high voltage due to collisional pumping between hot and cold mercury atoms. There is no need for the arc to "strike". Actually, mercury vapor has such a low ionization threshold that simply shaking a 1-2 cc pool of it in a glass tube under vacuum will create quite a light show. Back filli the tube with a little neon and you have a wonderful little desk toy that lights up red and blue when gently shaken in subdued light.

As to the photo of the Chinese GTL3 equivalent, yes, those white elongated strips in the envelope appear to be thermistors that have a low cold resistance that increases with increasing temperature. You have a good eye for detail. They could also be a ceramic material with a low work function (low threshold of thermionic emission) such as barium or thorium oxides designed to draw the arc away from the filament once it's going. They could serve both purposes. Either way, this variant of the GTL3, being self ballasted, should light up with just a 10-12 volt DC battery source

But then, the bulb pictured isn't a GTL3, it's a GTL3 replacement. I have many brand new Ushio GTL3's and not one of them has anything other than a filament crimped to the lead in wires.

BTW, I have roughly 30 pounds (13.61 kg) of scrap Hoya U325-C shortwave UV filter glass that I scored for free from a local UV lamp manufacturer. Using a tabletop glass kiln and machined graphite molds, I created a cylindrical black glass UVC filter that slips over a GTL3 bulb, turning it into a decent 254 nm isotropic UV source. Add a parabolic reflector and you have a kick ass UVC flashlight/mineral light.

 

Hi Richard,

Yes, I'm very familiar with the sunlamp you described. I also use one to expose Kodak KPR3 photoresist. The lamp itself is a standard quartz HID mercury vapor capsule with a beefy tungsten incandescent filament wrapped around it. The filament and the mercury lamp are connected in series across the AC line. At turn-on, the filament is cold and is electrically a very low resistance, putting the entire mains voltage across the mercury lamp. A tiny "tickler" electrode in the neutral end of the mercury lamp is connected to the line voltage through a high value resistor. The initial ionization in the tube occurs between the tickler and the neutral electrode and quickly spreads through the discharge tube, striking an arc across the entire tube. The mercury lamp now becomes a low impedance, drawing increasing current at an exponential rate. Were it not for the filament in series with the arc lamp, the lamp would explode. The filament now heats up, increasing its resistance and limiting the AC current through the lamp. As the mercury lamp heats up, the vapor pressure increases, decreasing the current draw through the filament. This is the point at which light output from the mercury lamp increases and the filament grows dimmer, causing the light color to shift from white to bluish white.

 

I agree.

The cap applies 170 volts peak to strike the arc and then limits the current once the gas ionizes. The inverter does the same. It applies a voltage high enough to strike and then limits the current.

The lamp, itself, sets the final voltage drop in both the cap and inverter versions.

While it may seem strange to think of a discharge lamp operating at such a low voltage, consider that when operating a conventional 18" fluorescent tube has only about 60 Volts across an arc gap of around 16". The fluorescent tube requires a considerably higher voltage to strike but in these GTL3 lamps the gap is a few millimeters and indeed once operating the voltage is only about 10V and once heated the arc will strike from less than 15V. I have one of these lamps right here on the table and have lit it up with a pair of 9V batteries in series and a resistor.

I realize it's a strange lamp but there is no magic going on here. Heated mercury vapor will ionize a rather low voltage and once ionized the discharge is a very good conductor.