Graphs of current vs voltage (or power) for incadescent bulbs?

There are devices built for this that are much better than bulbs, comprising a thermistor inside a vacuum capsule.

I have a citation of

amps = K * volts^0.541.

which you could check against a few cases.

John

Ian replied:

Real ballast tubes are iron wires inside a hydrogen-filled bulb. Somehow the hydrogen provides cooling that lets them deliver very constant current. But they are hard to come by these days.

It seems that some small incadescent bulbs are vacuum-filled (is that an oxymoron?) and others are argon filled. Maybe the argon-filled ones will have better constant-current performance. But I don't know how to tell the difference just by looking.

My requirements are not all that stiff: ten or twenty percent variation in (nominal) 300mA current over, say, 6 to 18V drop would be fine. I thought that a properly chosen bulb (or maybe a series string or maybe a parallel group) may do this for me. If you can name a maker/distributor of these thermistor/vacuum capsule devices, and if they're cheap (less than a few bucks each), I am interested.

Tim.

Why not an LM317 and a resistor?

John

A LM317 does not handle AC voltage all that well :-).

The need is for AC (sine wave) in, AC (sine wave) out. Converting to DC in the middle is possible, but I'm looking for something cheap and simplistic (thus light bulbs)! Some semi-sloppy current regulation is all I desire, and rectifying, smoothing, chopping, and filtering to get funked up current-regulated AC out is way too complicated.

Tim.

Across a bridge rectifier.

Cheers,

Phil Hobbs

I've used CdS photo cells (resistive) controlled by an incandescent bulb tweaked by a comparator.

...Jim Thompson

Or back-to-back, with schottkies?

PTC thermistor?

John

The LM317 does make a nice current regulator.

But should Tim Shoppa have some requirement or desire to use a lamp (wants the regulator to glow?) then I would parallel a bunch of 25 watt A19 "standard" or "soft white" incandescents. (I doubt these particular ones glow much below 18 volts because of greater filament cooling by the gas, although higher and lower wattages and tubular "refrigerator" /"showcase" lamps up to 40 watts [vacuum] do.)

Incandescents usually have a vacuum if very low wattage and/or low current - less than roughly 25 watts per inch of visibly apparent length-plus-diameter of filament. Incandescents of higher power/current than this tend to be gas filled. Lower current incandescents have a vacuum instead of a gas because the heat conduction losses by the gas outweigh the benefits of gas slowing down filament evaporation and permitting a higher filament temperature. This is because a narrower filament has a higher temperature gradient in the gas around the filament, which largely cancels out the lower filament surface area. Incandescents of design that barely has a net benefit by the gas have greatest effect of the gas - current varying less with voltage, and light output and life varying more with voltage than is the case with other incandescents. And these differences in performance variation with voltage are greater at low voltage - especially near or a little below the threshold of visibly producing light - than they are near design voltage.

I once many years ago plotted current as a function of voltage for a gas filled flashlight lamp (HPR52), and found the curve to be almost flat in some range of low voltages - I believe about .3-.8 volt.

- Don Klipstein ( snipped-for-privacy@misty.com)

"Tim Shoppa" a écrit dans le message de news: snipped-for-privacy@g44g2000cwa.googlegroups.com...

So put the 317 inside a rectifier bridge (if the discontinuity isn't too sloppy).

And I've used this idea with great success in Wein bridge oscillators.

But in my latest need, the power dissipation is 300mA x 14V or around 4 or 5 watts. A CdS photocell isn't good for anywhere near that much power.

Tim.

I have a Wagner Lighting automotive lamp catalog that has a table with some of this information, but it's general data, not specific to each lamp number.

I once heard that there are empirical formulas for this involving odd things like thirteenth and ninth powers, but I don't recall exactly what they are.

I have taken the liberty of cross-posting (and setting followups) to sci.engr.lighting, which should yield much better info.

Matt Roberds

Eg. the old Amperite devices. Last I saw them they were going for upwards of $100 a pop.

Here's an old graph of current vs. voltage for line voltage Xmas tree type bulbs:

Best regards, Spehro Pefhany

Any such set of curves would also have to account for time. At a constant voltage, the resistance/current changes as the bulb warms up.

I somewhat remember the "one-size-fits-all rule" at least sometimes saying that current is proportional to voltage raised to the .42 power.

I have found this exponent to be more like .52 with a high current gas filled lamp (9005 halogen high beam bulb), and .57 with a vacuum lamp (either 25 or 40 watt 120V "T10" showcase/refrigerator lamp). I expect this exponent to be less with gas filled lamps with design current a lot less than that of the 9005. I also expect this exponent to vary with voltage when the voltage gets low - often decreases in gas filled lamps, but has to increase towards 1 once the voltage decreases so much that the filament temperature is close to room temperature.

- Don Klipstein ( snipped-for-privacy@misty.com)

...And if it is too high, could you use it in series with a 14V power transformer?

John Perry

I am not to sure exactly what they do, B&K tube testers use a pair of bulbs to regulate something. I recall that they are number 55. B&K says they must be replaced as a pair from the same manufacturer.

They have something to do with the meter. If one of them burns out the meter goes crazy. Thought this might give you a clue. Good Luck, Mark

Hewlett Packard used a #49 lamp for there variable sine wave generator. To get an idea for a specific bulb: measure the 'cold' resistance with an ohmmeter; calculate the 'hot' resistance by dividing the power by the current squared. Most lamp have about a 10:1 ratio.

I find them very useful in design of protection circuits.