BCCs:  Some possibly useful stats re incandescent filament visibility 
at reduced voltages in 'table' at end.

>> I have some experience doing this in a 12V, 100W application.
>>  I think the most practical thing is to make a graph of
>> steady state current vs voltage.

> You wouldn't happen to still have that graph would you?

Would be very easy to generate for a given bulb.
And it will vary with bulb and circumstances.

Bulb temperature will vary with power dissipated.
Power dissipated is function of current squared and resistance.
Resistance is a function of temperature.
Current at given voltage will vary with temperature.
Radiation loss is probably reasonably predictable based on temperature 
alone.
Some second order effects will occur due to convection and conduction 
which will depend to some extent on bulb envelope dimensions - both 
distance from filament and envelope shape.

Oh dear :-) !!!

I'd suspect that taking a variable power supply and the bulb of 
interest would produce more useful results in a few minutes than you'd 
get from pages of tables.

I'd also guess that 1% of nominal voltage would produce no output and 
that 10% of voltage may produce something.

Oh, alright then ...
Mumble ...

T-10 12V 5W nominal.
Small auto dash etc bulb.


V    mA

12   340    nominal
11   330    much the same
10   310    dimming a bit
9     280    and a bit dimmer and yellower
8    270    yellowing
7    250    more
6    230    visibly down
5    200
4    180    very notably down but still very visible as lit bulb
3    150    dim - visible easily at 20 feet in workshop lighting
2    120    visible at 6 feet in workshop lighting
1v5 100    visible on inspection in lit room, clearly visible in dark
1v2    90    On close inspection in room light. Clear in dark
1    80    Not in room light. Whisper of light in dark.

Even at 1v there is no visible surge at turn on - current peaks to 300 
mA + before stabilising.

SO
WITH THIS BULB
Voltage of 10% nominal gives filament which is perceptibly on if 
carefully inspected in room light. Daylight probably not so. Notable 
filament glow in dark.

5% of nominal is reasonably well off.

Note how bulb is a reasonably good current source for up to 20% 
reduction in voltage from nominal.

Note that Ibulb is almost 30% of operating for V of 10% of operating.

Current limiting / feed seems a better way to control offness.
10% of on current would ensure substantial visible offness for this 
bulb.
20% probably would too.
Even 30% is hardly perceptible.



        RM





-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist