time constant for typical fluorescent lights For a typical flourescent light available in the United States on a a standard 120V 60Hz line, is there information available on the decay time of the light?  For example, if I am running the light a full power (ON) and shut off power (instantaneously) how long before the light is completely off (or reduced by $1/e$...)?
E.g. if the luminance is given by 
$$ L = L_0 e^{-t/\tau} $$
where $t$ is the time elapsed after killing the power: what is $\tau$?  Order of magnitude estimates would be great.  More precision would be even better.  I couldn't find this information easily on google.
 A: EDIT: I'm not sure what specifically you're after. I'll explain a little more why it is difficult to give you a number.  It is also quite possible that the number you are seeking might not be what you think it is....
The decay would be similar to any plasma type reaction for a mercury-vapor gas.  There is a delay between absorption and re-emission of light photons and a typical electrical system driving the energy is not going to stop instantly.  In fact the dominate feature in the decay rate will probably be due to the slow discharge of the ballast system driving the gas.
The common household circuit lighting up a tube usually has a large transformer with a big magnetic field to step up the voltage and resonate the gas to force the plates to conduct through the ionized gas and light up.  It takes a considerable amount of time for this electrical system to discharge and while it does it is going to continue to drive energy to the lamp causing it to light and effecting your "decay" rate significantly.  Here is a 12V lamp driver just for you to see what's involved electrically:

NOTE:  The large cap (0.047uF) and the transformer are going to resonate the lamp energy much longer than the actual natural decay rate of the gas.  In addition the 47uF cap is going to supply power to the circuit for a non-trivial amount of time after the 12V is removed.
For comparison here is an 120V 60Hz ballast design and you'll see there are similar issues.

If you want to look at just the gas decay rate then you might have to excite it with a laser for more precise measurements.
The states that the energy moves through is usually measured and plotted on a Jablonski Diagram like this:

In your case $\tau=\frac{1}{k}$ where k is equal to (in the de-excitation case)
$k = k_f + k_i + k_x + k_{ET} + …= k_f + k_{nr}$
Where $k_f$ is the rate of fluorescence, $k_i$ the rate of internal conversion and vibrational relaxation, $k_x$ the rate of intersystem crossing, $k_{ET}$ the rate of inter-molecular energy transfer and $k_{nr}$ is the sum of rates of radiationless de-excitation pathways.
And a more detailed model can be measured:

The method for measuring a gas properly looks like:

Here is an example of what the data looks like.

Much of this information was taken directly from the research documented here:  www.jh-inst.cas.cz/~fluorescence/support/Lectures/UFCH_fluor03.pps‎
