From long time ago, when I wasabout 12..13 years old (so it was around 1988), I remember doing some electronic experiments with my classmate in his father's home shop. And in between the lot of different junk (what usually get accumulated over years on these places) wi find some sort of fluorescent tube. If somebody will catch up and recognize something he have, or at least know some literature about and share some information, I would be glad.

What I remember quite well:
- It was glass tube, about the size of F4T5, maybe only a bit smaller (his father have these as well, so I remember the raw compare, but I don't remember it side-by-side)
- It was coated by a phosphor layer
- It have one solder joint tip on each end, around were brass sockets. But I don't remember, if the wires were soldered directly to the brass, or there was some insulator between
- As far as it was visible, there was a filament going axially though the tube from one contact to the other one
- It was labeled as "4W"
- It was rated for DC supply directly from the vehicle battery, either 12V or 24V (I recall 12V, but I'm not as sure if it was not 24V - see below), without any ballast (as far as we understood it)
- When tried form a DC source of that voltage, it first start to glow incandescent dark red to orange-red (as it warm up) and then (afer about 2second) it jumped to either cool white or something between cool white and daylight (for sure not an incandescent color; It was very similar to the then ordinary halophosphate CW fluorescents illuminating the workbench)
- I don't know, if it was polarized (worked only in one polarity or both; if the polarity was somehow marked on it)
- At that time (~1988) it was already assumed as obsolete, kept only as a curiosity, as already then were quite common inverter based fluorescent fixtures, so made latest in 70's
- I have no clue about the make, neither country (even area) of origin. If it was from early 70's, it may have well originate from the "western world"...

Based what I know today, I may say:
- I've never seen anything like this again (real life, nor in different literature, lighting related internet pages,...)
- Based on the behavior, it was not an incandescent (neither LED, obviously), but either low pressure discharge, or free electron based fluorescent (the visible light originate from the phosphor)
- For either of the above concepts, the 12V seems to be too low voltage, that's why doubt, whether it wasn't rather the 24V - another common voltage in vehicles at that time...


So if someone here recognize something, ask (maybe I would recall and either confirm or reject), share info about you may remember seeing something very similar somewhere,...

I would very like to "assemble the puzzle" and get better view, what that may have been.

Definitely a low voltage flourescent lamp. I did send one to Sailormoon for his collection, and probably still have one around. Mine though are a single ended item with a bayonet socket. They were normally operated with a resistive ballast in line with the supply, generally running at around 1A of current.

And do you have some specification?
What are the operating voltages, currents, ballast specifications,...
I remember we didn't use any ballast, it light well from stabilized voltage source. But it may be, then in the car some sort of ballast was expected. Or the one we had may included some ballasting resistor (incandescent,...) inside...

The ones I have operate off 28V DC, with a ballast of 2 resistors in series. One is 12ohm and the other is 27ohm, with the lamp across the 27ohm section. These lamps were used in aircraft as cockpit map lights, and were operated via a resistive dimmer. Not terribly bright, but they were used behind a woods glass filter to illuminate the maps only. They run pretty hot.

Maybe incandescent lamp with arc that strikes parallel to the filament, like the ends of a good fluorescent with stuck starter. Then it's current is limited either by some ballast resistor (how much power it would take ? not more than 4W, so the ballast probably no more than 2W, so can be just a ceramic block resistor inside one of the endcaps) / some kind of saturation that prevents the arc itself from taking higher current if that is possible

The thermionic discharge (described by Ash and how it look like) need the supply voltage to exceed at least the cathode fall, what is about 15V (as far as I remember), so such lamp wold not operate at 12VDC. And I'm still ~80% convinced, then we run it at the 12V only (because I remember having the maximum output voltage of the regulated power supply to be 20V - I find some paper notes abut it from that time; but there is stilla possibility, then we modified it and I forgot that)


The split of 33% of the input voltage for the ballast and 66% for the lamp seems realistic, what would mean ~1.3W from the total 4W for the ballast resistances and 2.7W for the main lamp itself - that would correspond to the light output of the 12V 5..10W incandescent lamps.

They were used for the UV output, to make the paper glow in the dark without spilling light. they did have a red glow as well, but not much.

Is there a way to make a discharge that would have some "internal" limit on its current without a ballast ? such as very limited amount of charge carrying partcles etc ? (to get to some kind of saturation condition in the discharge itself)

Simple answer is no. Discharge lamps are current sinks, and you need to keep the current constant, even as the voltage across the tube increases or decreases.

There is still a question, if some cathode chemistry is possible to make the cathode fall way smaller (let say ~5V), so a discharge lamp may operate at the 12V supply.

@Ash: That would be the state of 100% ionization (all atoms of the fill are ionized), but that would require either very low pressure (virtually a vacuum) and/or very high current. To be able to use that in normal operation it would require extreme good pressure control of the nearly vacuum, what is not feasible. With rather normally attainable and controllable gas pressures such saturation would happen on currents in high 10's of A in the T5 tube.
Such state normally happen in "vacuum" inscandescent lamps and another vacuum electronic devices as a parasitic effect - there is small current carried by the ions from the remaining gas, but it is rather unwanted (in CRT's it cause deterioration of the screen phosphor, so a layer of Al is necessary to protect it, in all free electron devices the ions degrade the cathode).
Moreover it won't make much sense to complicate the manufacture by such accuracy, as energy-wise the effect wold be the same as series resistor ballast: Dissipate the extra voltage as the heat. Moreover I'm not sure, whether 100% ionized plasma is able to generate any light at all - normally the neutral atoms get excited by the electron and return to the lower energy state of these then emit the radiation...

But if you fill it with mostly isolating gas (nitrogen...) so that you have to control the part of the active gas in the mix, and not the pressure

Pressure determines the discharge voltage. You need it to be almost a pure vacuum to have space charge limiting controlling current. Either you need a cathode that is limited in emitting area ( and which will burn out in fairly short order) or you need a control grid to control current ( like in a thermionic valve). In either case this is more complex than a choke ballast, and is going to be difficult to make in a package that will be compatible with existing tubes. It will need at least more leadout wires through the glass, and these are both prone to leakage and difficult to make in a small area.