lamp air flow sensor

Halogen bulbs can run hotter for a lot longer than normal tungsten bulbs, which gives the recrystallization more of a chance. The halogen redeposition mechanism is basically to keep the envelope clean--once the tungsten iodide is rattling around inside the envelope, it isn't selectively redeposited on the hot spots. Normal bulbs don't have that mechanism, so their envelopes darken progressively with use.

It's the high pressure inside the (very strong) quartz envelope that makes the redeposition selective, by retarding the diffusion of tungsten vapour away from the hot spots.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

Hmm I'm not sure of that last sentence. I'd think that one thermistor is run at low current to measure the temperature of the air. The other (smaller) one is run hot.. to see how much cooling the wind does. The question is how much power should you put into the hot one? Maybe the current into the hot one is servoed to be some constant temp above the other.... It's an interesting question.

George H.

I think that just adds more calculations, run them the same and when there is no flow they are the same resistance

the more power, the more sensitive & the less temp range measurable. And of course after a point the more the thing degrades over time.

NT

A random aside as you seem to be something of a lamp-expert, do you know anything about studies of the longevity of small (intact) incandescent lamps, e.g. grain-of-wheat, run at various currents?

I figure they're not as unreliable run at low currents as some would imagine, there's that one Edison bulb in the firehouse that's been running for 107 years or something.

For a smaller modern bulb at a couple of mA idk 10,000 hours, maybe? Would depend on if and how it's cycled I guess.

Asking for a friend...

From the GE book on neon glow lamps small neon lamps seem to be just about indestructible other than physically, can blast them with inside-a-reactor levels of gamma rays for months with little detectable change to their parameters

en in

logen

e

that

n

Filament lamp life is inversely proportional to voltage to the power of abo ut 3.5. So if you want 10x rated life, you'd work it at around 82% of rated voltage*. That will do it no favours in light output or efficacy though.

*The further you get from standard running conditions the less accurate thi s figure becomes.

NT

Excess current kills them. And they don't last forever at rated current, so better not used where they'll be on all the time. They were once considered a viable room lighting option.

NT

Longevity is also DC / AC dependent AFAIK. I have been bitten by running a halogen lamp on DC, got very short lifespan. When searching for clues I found and articles with nice pictures of unidirectional smearing of the crystals, in one direction, when using DC. Adding a relay to invert the pins between the 8h power cycles saved the day...

Arie

pen in

mal

w

dc lamps also tend to fail at one end, probably due to a valve-like effect.

NT

Some articles about this effect, searched for incandescent & electromigration:

"Observations on both electromigration and the Soret effect in tungsten have been recorded using silhouette electron microscopy and electron microscan techniques. The surface structure developed when a tungsten

A/cm2 depends on whether the current is direct or alternating. An alternating current produces a smooth surface structure while a direct current produces a steplike structure due to electromigration of tungsten ions. The direction of the electromigration of tungsten ions is toward the cathode in agreement with the theories of Huntington and Fiks. Soret effect diffusion has been observed in

Either electromigration or Soret effect diffusion can produce pronounced changes in the original geometry of the conductor resulting in large current density variations over a short wire segment."

Arie

en in

The last time (a quarter century ago) I looked into this, there were vacuum-filled bulbs for flashlights, with very short lifetimes but high power efficiency (like, 2 hours and they start getting black inside) as well as 'krypton' higher-power lamps. The 'krypton' was less energy-efficient but better (less red) color. The halogen was also bette r color, but longer lifetime (and less tolerant of overvoltage).

Halogen lifetime against blackening the envelope depends on them running at full design power; if you put a dimmer on those the lifetime can be terrible. I've rewired some of my home lamps to remove dimmers, it improves both halogen and LED performance.

Lamps that are hard to replace (indicators) are often run at low currents and last for years but have TERRIBLE power efficiency. My auto instrument panel had three of eight indicators dead, when it was relamped at circa 30 years of age, if that means something. Of course, such age data depends on the technology in use decades ago.

Alas, the fill gas in those is Ne and He, and there is a high-likelihood failure mode where the He diffuses out. Materials and the technology of making glass/wire seals matter a LOT. Quality vaires. Does GE still make those? Who DO you trust?

Yup, they're vulnerable to electromigration because the W is run so close to its melting point (3200 vs 3400 K roughly).

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

The halogen redeposition mechanism selectively returns the tungsten to the thin spots on the filament that are running hotter. Keeping the envelope clean is an incidental advantage.

The fact that the redeposited tungsten tends to deposit as large crystals is a weakness of the scheme.

The do have a filling of inert gas, which does slow the darkening process.

The main selective mechanism is that the tungsten halides are stable at room temperature, and only come apart close to the filament when it is hot.

Tungsten halide bulbs are small so that the silica shell runs hot enough that the tungsten halides won't condense on the inside of the bulb (at least not when the lamp is running).

--
Bill Sloman, Sydney

So I got a root canal today. I usually find some mechanical thing to think about when getting dental work. Today it was a thermistor air flow sensor. (so thanks for that.) My constant temp difference idea would be hard, I'd have to take the ratio of the current to the thermistor voltage, to find out what the temperature is. A bridge would give the difference, but I think you need the absolute value too.

So some switching thing? Or measure a bunch of things with a uC?

George H.

why make it so hard? Fixed v or i to thermistor, with an opamp reading the other variable i or v.

NT

Thankfully I don't have to make voltage references out of neon lamps designed to withstand the prompt gamma emitted by a nuclear explosion at some absurdly close range, I suppose someone did at some point and I surely hope they were well-compensated for the job!

In WW1 hot wire microphones were used to locate the infra sound from distant artillery.

piglet