AC Soft Start Incandescent Lamp switch (*FINAL EXPERIMENTAL RESULTS*)

[SeanB]

I do have the circuit, drawn in what is now called DaveCAD, but it is at work. Not exactly complex though, using 6 different components and a little bit of well butchered Veroboard. Resistors were really derated to fit my definition of safety margins, and they always ran cool. It was built from parts from the junk box, even the thermal fuse was a spare unit I had around for use in fans.

[hgg]

Shorting out a soft start thermistor makes it even more pointless.

I don't want to save the bulb and stress the thermistor... 
After the thermistor does its job, it is no longer needed.
I am looking for the the most efficient way to protect the
bulb and all the other  components as well, so that the
whole solution will last.

[IanB]

I don't want to save the bulb and stress the thermistor... 
After the thermistor does its job, it is no longer needed.

Several people in this thread have told you that putting a soft start on a light bulb will not extend the life of the bulb.

The life of a bulb varies inversely (and exponentially) with the operating temperature of the filament. The hotter the filament, the shorter the life. All a soft start will do is to delay the inevitable bulb failure by a few extra starts.

Furthermore, all components are stressed somewhat in operation. If you put a relay in there you will be stressing the relay to save the thermistor.

If you really want to extend the life of the bulb the best option is to put it on a dimmer circuit and operate the bulb at lower brightness. Drop the brightness by 10% and the bulb life will probably be doubled.

[hgg]

I have yet to see any actual data that support your rather strong belief.
People from the Apollo space program have a different opinion.
If you have any actual data, you will save me a lot of time.

[IanB]

I have yet to see any actual data that support your rather strong belief.
People from the Apollo space program have a different opinion.
If you have any actual data, you will save me a lot of time.

You could start your researches here:

Why does a light bulb burn out?

[hgg]

I've read it before.
They have not used any kind of soft start circuit in their experiments.

Please read the following more carefully:
http://technology.nasa.gov/?t=open&this=/success_story/larc/9/200669

[Rufus]

I have yet to see any actual data that support your rather strong belief.
People from the Apollo space program have a different opinion.
If you have any actual data, you will save me a lot of time.

Seriously?

"It was originally developed during NASA's Apollo program to protect the Saturn V launch vehicle from electrical current surge"

No mention of lamps at all. NASA develop some kind of thermistor. An shyster claims putting it in series with a lamp will save money. NASA keen to show their government funded work provides commercial spin offs publish one page on the work of this shyster and you claim that counts as approval from the Apollo space program.

I am not surprised data on the effect of soft starting is hard to find. Lamp wear out and failure mechanisms are well understood and the effect of soft stating quite predictable. The only people who could benefit from such data are the shysters trying sell soft starters and if they have any they haven't published it because it doesn't support their claims.

[IanB]

So on the one hand we have a detailed analysis of bulb operation and the physics and metallurgy of the bulb filament, with photos, electron micrographs, experiments and calculations you can perform yourself.

On the other hand we have a short paragraph making a claim with no data, no details, and no experimental evidence to support it.

And you would rather believe the unsupported claim than believe the science?

It seems you need to to the experiment yourself. Here is how you can do it:

In a laboratory, set up 100 light bulbs. Put a thermistor in series with 50 of them, and put an ordinary 10 ohm resistor in series with the other 50. Rig up an automated relay system that turns the bulbs on for an hour, then turns them off for an hour, then turns them on for an hour, 24 hours a day.

Watch the bulbs over a period of months and record the failure rate in each group. See whether one group fails at a faster rate than the other.

When you have done the experiment come back and let us know the results.

[IanB]

Seriously?

"It was originally developed during NASA's Apollo program to protect the Saturn V launch vehicle from electrical current surge"

No mention of lamps at all. NASA develop some kind of thermistor. An shyster claims putting it in series with a lamp will save money. NASA keen to show their government funded work provides commercial spin offs publish one page on the work of this shyster and you claim that counts as approval from the Apollo space program.

Here's the kicker:

Public Release Date: 0000-00-00

The "invention" does not even have an official publication date.

[BravoV]

The direction of this thread's discussion is becoming more and more towards religion rather than an applied science.

.... "I want to believe ..."

Audiophool was born this way. 

[hgg]

It sounds almost as if you are working for the "energy saving" CFL bulb industry.   

Anyway, thank you for your help.

[SeanB]

Lower the voltage a little and the bulb lasts longer, but the light output is redder and lower intensity. I run a lot of 12V 20W dichroic lamps at work, and they have a long life with a reduced voltage, failing only at turn on. Failure while operating is invariably caused by a failing socket that arcs and burns the lead wire. Life is also strongly dependant on manufacturer and plant, a lamp made in the EU, USA or almost any country other than China lasts a lot longer, and the PRC made lamps have a widely varying life depending on the manufacturers QC. Best lamps are made in Poland by Phillips currently, as they are about the only ones still left in EU making lamps.

[hgg]

Best lamps are made in Poland by Phillips currently, as they are about the only ones still left in EU making lamps

Nice to know.  Thanks!

[SeanB]

Youy cannot buy then in EU though, too much mercury in them, which gives the best life. Worst lamps are those low mercury ones, as they always go mercury starved and glow pink.

[hgg]

Reviving an old topic just to show the results of an experiment that lasted many months.

I was trying to find a simple soft start for incandescent lamps and in the end I used in series
thermistors.  I made a simple 555 timer circuit that was turning on and off a relay every
five minutes and switching between two bulbs.  One of them had two thermistors in line.
Its hard to find thermistors with exactly the value you want, that is why I used two in series.
The time interval was chosen in order for the thermistors to return from 90 degrees Celsius,
back to the ambient temperature.   Each 40W bulb was 5 minutes on and 5 minutes off.
All bulbs were from the same batch.  OSRAM 40W / 400 lumen.



(Image has been processed in order to show everything equally exposed. 
The bulb is way brighter from what it looks like in the photo.)

The bulb with the thermistor has a nice soft start and was reaching full brightness after 90 seconds.

The results:
The protected bulb has already outlasted two unprotected ones and it keeps going strong!
Each bulb lasted for exactly 2 months.   (or 720 hours of continuous operation and 17280 switchings.)

So, for the life of the incandescent, inrush current and thermal stress is just as important as filament
evaporation. I will continue the experiment to see how much longer this bulb will last because I think
that inrush current is more important.  Its been already on for at least 1440 hours.  I will report back.

Bulb Miser and NASA were right I think.   

[SeanB]

How about this lamp? Still going strong, and made circa 1939.

[hgg]

Its barely lit but nice.    
Do you know if it was shinning brighter before?

[SeanB]

It was dimmed to show the filament. Otherwise it is a 16 candlepower lamp (40W equivalent) and has a SSR and a small 10R thermistor ( ex PC power supply one) to soft start it. It has been doing faithful work as a wake up alarm now every morning for a few years. I took a clock radio and cut the board in half, removing the radio side, and put a hockey puck size Opto22 SSR in the space now opened up and fed the mains through it to the lamp. I had to replace the radio with a new one because the old one failed after 20 years of use.

I do have a lifetime supply of spare lamps as well, and if I am unhappy with the direct light I just put a inside frosted one in instead.

[hgg]

Excellent!   

[Wim_L]

Hi,

In my opinion the best artificial lighting available in the 21 century is still the humble incandescent lamp
for many reasons.  The only problem is that apart from the fact that they are slowly becoming extinct,
after the introduction of the toxic "energy saving" bulbs, the lifetime of the incandescent ones for some
strange reason started to decline dramatically.  <-- (conspiracy theory)   

Many countries now run slightly higher mains voltages than they used to. This makes the lamps run hotter. There may also be the effect of manufacturers trying to make their bulbs more efficient, and the hotter you run a bulb, the more efficient it gets (but lifetime is much shorter). As mentioned before, colour temperature is affected too.

That's also why dimming dramatically extends lifetime - and lowers efficiency. If you use less power per bulb, but need to use more bulbs to get sufficient light, you could end up spending more on increased electrical power than you save by making bulbs last longer. One exception is for bulbs that are hard or expensive to replace for some reason, in that case running a bulb at low power to dramatically increase its lifetime may be a good idea. Otherwise, select bulbs so they can be run at full rated power.

Dimming a halogen bulb too much isn't always a good idea either... http://www.teklight.com/faq_s.html

[hgg]

That's the thing though.  I've tried to protect the lamp but with minimum dimming.
The protected bulb was emitting 132 lumen lux @ 1m and the unprotected 148 lumen lux.
Very subtle difference in luminosity but a huge reduction of inrush current.

[edavid]

That's the thing though.  I've tried to protect the lamp but with minimum dimming.
The protected bulb was emitting 132 lumen @ 1m and the unprotected 148 lumen.
Very subtle difference in luminosity but a huge reduction of inrush current.

You should add a resistor to the unprotected side to match the hot resistance of the thermistors.

P.S. Very nice test setup!

[hgg]

You should add a resistor to the unprotected side to match the hot resistance of the thermistors.

You are right, that would be more accurate.  I will do it.

[ASowa]

You should also increase the on time.  Right now the protected bulb spends 30% of its on time at reduced wattage which is not isolating the soft start mechanism.

I also believe you meant to say lux rather than lumens.   Lumens is not a distant dependent unit.

[retrolefty]

How about this lamp? Still going strong, and made circa 1939.

And the back story, this firehouse is about 30 miles from me.

http://en.wikipedia.org/wiki/Centennial_Light