I want to do a timer that will turn on and off mains powered light bulbs in predefined intervals using a triac. For added benefit I would like to add a soft start for incandescent lights. When incandesent light is cold it has low resistance, leading to high inrush current that is damaging the filament. Since I will have a MCU controlled triac I want to implement a slow turn on to give the filament time to heat up without the inrush. I have no idea how quickly the filament heats up. If it is fast enough switching at zero crossing is all that is needed. But if the filament needs more than one full mains cycle to heat up it is needed to slowly increase the conduction angle. I cannot find any info related to this or a similar project.
Are they special light bulbs of 100 years old ?
If not, then don't worry about it, what you propose might make it worse.
Unless its very high power with a thick filament mains bulb filaments heat and cool very fast so you can't use zero-cross switching, you need cycle-by-cycle triac operation to prevent flicker.
It doesn't heat-up that fast. An incandescent seems to come to full-brightness in about 1/10th of a second. I've made some sound activated lighting effects and if I don't hold it on for about 1/10th of a second (or more) I can see it's not reaching full-brightness. Of course, you could go more-slowly.
You should be able to find a dimmer project and then it's just a matter of fading-up the brightness...
A dimmer works by triggering a TRIAC at a known time after the zero crossing. Then a TRIAC latches-on and won't turn-off 'till the next zero crossing (when current goes to zero).
Of course, your zero -crossing detector (and your TRIAC) has to be isolated. Most people use an optical isolator for both, but when I built a dimmer* I also built the power supply so I used the transformer secondary to get an isolated signal for zero crossing.
*A long time ago, with a different microcontroller, I built a "sunrise dimmer" that fades-up a lamp over a 10 minute period. (My sound activated effects are just on/off with no dimming.)
Get some led bulbs !
Not suggesting you should or shouldn't slow start your incandescent lamps. However unless you are turning them on or off very very often the added life is negligible.
If you want to increase the life of an incandescent bulb you should reduce the voltage a little bit. The live of a bulb varies as the 12th power of the voltage ratio.
So a 5% reduction in voltage will yield nearly double the life.
Update:
All three incandescent lights I have tried need about one mains cycle to heat the filament:
Yellow trace is mains voltage, blue trace is voltage over 10R current sense resistor (meaning 100 mA/div). Turned on simply by plugging into the mains socked.
extending life of incandescent light, put a diode in series to the bulb. Which will cut its light o/p but make it run cooler. I then put a switch around the diode so, if I want full brightness I can bypass the diode.
You'll want a diode with at least 600 PIV.
When was the last time you were sitting in a room and the lamp failed?
Applying voltage at 0vac is the best approach, not ramping.
The time spent at low resistance where stress is greatest, the filament is most 'vulnerable'.
This explains why lamps fail (almost always?) at turn on, flick the switch and "POP", probably even coincides with peak volts.
Hi,
Many, many, many years ago, when slide projectors were in their heyday.
Some of the more powerful units that were on the market, had the habit of blowing its high power globe just as you were about to bore the life out of the family with the holiday snaps.
A magazine here called Electronics Australia had a project that you plugged in between the mains outlet and the projector, this provided a soft start for the lamp.
The condition was you had to make sure you turned the projector power switch on first.
This was the late 1960's and it used a Thyristor as the soft start with a bypass relay that switched in after the Thyristor had done its soft start.
Any ventilation fans also got the softstart treatment, but for the 5 seconds or so it didn't seem to cause any problems.
Tom... 
This topic was automatically closed 120 days after the last reply. New replies are no longer allowed.