Incandescent lamps on 230 V DC? 1

[Skogsgurra]

What are the problems (if any) if one wants to run normal 230 V lamps from a DC voltage with 230 V tension?

Of course, the switches and fuses may be a problem, but the lamps as such? Is there any knowledge out there?

 

[VEBill]

Based on my understanding, it shouldn't be a problem.

The explanation for RMS voltages always mentioned the equivalence for heating (and obviously incandescent lighting).

 

[Skogsgurra]

Yes, that is understood. But otherwise? Better safe than sorry. That's why I ask. I mean, the glowing wire is quite thin and what happens if you have a DC voltage across the filament? Do the ions migrate from one pole (+) to the other (-)? With AC that doesn't happen. But DC?

 

[benta]

There seems to be consensus that supplying incandescent lamps from a DC source shortens lamp life.
Try googling for:

filament electromigration lamp

 

[dpc]

I don't think there would be much of a problem. The filament failure mode might be a bit more exciting than an ac system since you have no zero crossings to help extinguish the arc.

You could check with lamp manufacturer.

 

[VEBill]

To counter any possible 'ion migration' effects, perhaps once a month you could reverse the polarity of the applied DC voltage... ;-)

More seriously, if you're worried about bulb life then drop the voltage ever so slightly.

 

[Skogsgurra]

Thanks benta. There seems to be a problem - but not quite the one I thought. Does material migrate IN THE FILAMENT? Or have I misunderstood something? I thought that the ions could be moving in the vacuum.

 

[btrueblood]

There is no significant ion migration; filaments "burn out" due to evaporation of the filament (the black stuff that collects on the walls is tungsten). There is anecdotal evidence of light bulbs surviving much longer on DC than they can on AC (see stories of the Edison bulb); the explanation is that the AC current causes the filament to vibrate and subsequently fail due to fatigue.

 

[Myndex]

It's an interesting questions considering that Edison developed incandescent lights using DC...

But it seems that running a lamp on DC will live only 20% of it's AC operated counterpart.


This from Tektronix:

AC VERSUS DC OPERATION

Since lamp life is primarily a function of filament evaporation, there is a minimal life loss from ac or intermittent operation. Failures under ac operation are generally due to "crystallization", although some filament evaporation is present.

However, because of filament evaporation, lamp life under dc operating conditions is around 1/5 the rated average life specified. Hot spots are developed as the filament evaporation accelerates from localized increased resistance.

Electromigration on the tungsten filament results in "notching" or a sawtooth surface.

Most Tektronix incandescent lamp applications are under dc operating conditions. The increased filament evaporation causes a substantially lower life expectancy so provisions must be made for easy replacement.

Figure 2 illustrates, among other things, the tradeoffs between applied voltage and average life ratings for lamps operated under ac conditions. This graph is accurate only for lamps at 5000 hours or less rated life. Applying 125% of rated voltage will yield 0.07 x rated average life hours of operation.

For dc applications the suggested formula for life hours is:


life = 1.8 x 104 hours x (d2.5) EXP(f/KT)

where d is wire diameter/N CM
f = 2.72
K = 8.6 x 10-5
T is filament temperature in °K

In general, ac operation is recommended with a maintaining current to avoid high inrush current.




Andy

 

[VEBill]

"...running a lamp on DC will live only 20%..."

Sounds rather bleak...

If this is really true, then the lamp manufacturers should include within their incandescent lamps a small IC to switch the DC into (square wave) AC. A five-times (?) improvement in life span might, for some applications, be well worth the cost of such a trivially simple and super-cheap IC.

This subject has me thinking about those ultra-expensive lamps used in video projectors. Those bulbs retail for as much as Cdn$600 (~US$500). I wonder what the trade-off is between using raw household AC (with its obvious voltage variations), or using more tightly-regulated DC, to power the lamp?

 

[Warpspeed]

My instincts tell me that design operating voltage might have a large effect on ion migration. While a lamp designed for 12 volts could possibly work equally well on ac or dc, 230v is a whole different matter.

My only practical experience with this was with a 240v system in Australia that used batteries to provide emergency lighting in the event of mains failure. The biggest problem was variation in battery voltage, especially when the batteries were taken straight off charge, and placed under lamp load. Contactor controlled diode voltage droppers were used to reduce the initial voltage.

As an emergency lighting system hopefully only has to operate for short intermittent periods, overall bulb life is probably not a huge issue. If you plan to run lamps continuously off high voltage dc, a lower design operating voltage might present fewer problems.

 

[SolderingGunslinger]

If the incandescent lamp is rated 230V RMS, then there shouldn't be a huge difference in the filament life, IMHO. If the lamp is rated 230V Peak to Peak, then the operating the lamp from 230 VDC will indeed shorten the lamp's lifespan.

Back in the early days of Edison lamps, one of Edison's workers noticed that after some hours of operation, you could see the "Shadow" of one of the filament supports on the inside of the envelope. Edison was a proponent of DC power for household use. Tesla championed AC. Anyway, the carbon filament early Edison lamps used "Boiled" away and the carbon condensed on the inside of the envelope of the lamp...except where the filament support connected to the positive side of the supply collected the carbon, leaving the shadow.

Edison did an experiment, placing a second electrode within the envelope and noticed that current flowed through the otherwise unconnected "Plate". As his system was DC, he had no idea that he had discovered a rectifier, saw no commercial use for the interesting phenominon and did nothing with it.

How much tungsten boils off the filament? I haven't ever done a study, however I would guess, running your lamp off DC might indeed cause a "Shadow" to appear on the lamps envelope.

You didn't mention your reason for operating the lamp from a DC source. Whatever that may be, good luck.

I remain,
The Old Soldering Gunslinger

 

[Myndex]

If this is really true, then the lamp manufacturers should include within their incandescent lamps a small IC to switch the DC into (square wave) AC.  A five-times (?) improvement in life span might, for some applications, be well worth the cost of such a trivially simple and super-cheap IC.

Or, they could use an LED, which runs really well on DC....

A

 

[Warpspeed]

Gas lights and candles have none of these problems hehehe.

 

[Skogsgurra]

Ever tried to run a gas light on DC? ;-) Or a candle?

 

[SolderingGunslinger]

Ever tried to run a gas light on DC? Or a candle?

There are those who think I've been in broadcasting since the time you used a wooden match to light the TV set's pilot light. :-D

 

[logbook]

The bulbs in cars seem to work ok on DC and they have to contend with continuous vibration and environmental temperature excursions as well.

 

[Myndex]

logbook: sure, bulbs do "okay" on DC, but those same bulbs would last up to 5 times longer on AC.


Here's an idea - why not take the car alternator's AC output and use it for the lamps? Probably never have to change a bulb...


A

 

[Skogsgurra]

I don't know if a measly few hundreds of hours is "to do OK". A car isn't in use for many hours/year and I think that I have to change bulbs at least once a year.

Also, I still think that ionisation plays some role - and that gets worse with higher voltage. 230 V is a lot more than 12 V.

 

[VEBill]

"...gets worse with higher voltage."

Hmmm...

The voltage divides itself across the length of the filament. Doesn't that mean that each section of the filament would 'have no idea' what the overall voltage is?

One can imagine making shorter filaments with proportionally less voltage and the conditions on those shorter filaments would be identical to the full length filament with the full voltage.

This equivalence might not apply during start-up.