Department of Energy Argonne National Laboratory Office of Science NEWTON's Homepage NEWTON's Homepage
NEWTON, Ask A Scientist!
NEWTON Home Page NEWTON Teachers Visit Our Archives Ask A Question How To Ask A Question Question of the Week Our Expert Scientists Volunteer at NEWTON! Frequently Asked Questions Referencing NEWTON About NEWTON About Ask A Scientist Education At Argonne Instantaneous Light Intensity Function
Name: Emanuel G.
Status: educator
Age: 50s
Location: N/A
Country: N/A
Date: 10/16/2004

Dear Scientist!

Though I by my own do a bit with some science, I (we) have a problem that I (we) cannot answer. We are in collaboration with a photoelectric laboratory in Greece, and there is a problem with measurement of THE INSTANTANEOUS LIGHT INTENSITY FUNCTION,y(t), of a fluorescent lamp fed from 50 Hz line via magnetic ballast, the reasons for which we ALL cannot understand. When a photoDIODE is used for the electronic measurement of y(t), then the waveform registered by the oscilloscope is close to

y(t) = A + B|sinwt|

where the time-depending part is the ripple (light flickering), and when photoSENSOR is used, then the waveform is close to

y(t) = A + Bsin(2wt).

That is, a smoothing of the singular waveform is done in the latter case, and we do not see why the photoSENSOR (should be also quick enough) smoothes the wave.

We would be very grateful if this message would come to a physicist who know such measurements and who would be in contact with us (me) trying to explain the situation.

Probably the two sensors are sensitive in different wavelength ranges. Several phosphors are used in fluorescent lights, to make a spectrum that, overall, looks white; the different phosphors fluoresce at different wavelengths, and they have different decay times. Generally, the longer-wavelength phosphors have longer decay times. See Wilkins, A. J. and Clark, C.; "Modulation of light from fluorescent lamps"; Lighting Research and Technology 22 103-109 (1990).

Tim Mooney

Yeah! It's out of my area of expertise. If I had to make a guess (I would not bet a lot of money on it.), I would look into the power supply electronics. Sounds like the diode circuit that depends upon |sin(wt)| has a signal inverter that converting the (-) phase into the (+) phase thus doubling the time dependence. The inverter time constant is probably not the same going from one polarity to the other, i.e. (-) ----> (+) not equal to (+) ----> (-). This would double the time constant. In the other case the intensity does not distinguish whether the electrons are moving (+) --> (-) or (-) --> (+) thus doubling the frequency of the flicker. I admit this is an ex post facto "explanation". Sounds like there's a need for an EE major to look at the circuit diagram.

Vince Calder

Emanuel- bad news first: I do not know specifically what kind of device your "photosensor" is. In my terminology, "photosensor" is a general term for anything which converts light levels to a signal in some other medium, such as electric current. It includes many things such as photodiodes, photoresistors, photovoltaics (solar cells, photodiodes used in power-generating mode), photocathodes in photomultiplier vacuum tubes, phototransistors (merely a photodiode in parallel with the collector-base junction of a transistor), and presumably others.

Good news- I recognize "photodiode" as something specific. Photodiodes tend to be fast and accurate. They are usually linearly responsive to light intensity over more than 3 orders of magnitude, and usually have response speeds of < 1msec to < 1nsec, depending size, load-impedance and bias. (This supposes that you measure the current through them at zero volts, or externally apply reverse bias. If you let them generate their own voltage like a solar cell, and you only measure that voltage, these advantages do not apply. I am betting you used it right.)

So I tend to believe your "y(t) = A + B|sinwt|" is the accurate measurement of light levels in the millisecond time range. If your fluorescent tubes are new and working well, your sinwt is 50Hz.

The sloppiest, least-often understood kind of photosensor is the photoresistor. (also called photoconductor) They are often made of grainy yellow Cadmium Sulfide (CdS) printed on a 1-cm-wide surface between interdigitated many-fingered electrodes of printed silvery metal. So the CdS looks like a yellow squiggle or zig-zag. This semiconductor has few free electrons or electron-holes normally, so there is a high resistance between the two metal electrodes. When light falling into its crystals creates new electrons and holes, which live long enough to drift between the metal electrodes at a speed proportional to the voltage the user has applied. So it is a conductance proportional to light level, or, a resistance inversely proportional. If you are careful, you will apply a large voltage (say 10v) and read the current through it by the voltage drop across a smaller resistor (< 1v). That way the response may be somewhat linear with respect to light level. However, the proportionality varies with light level. It is more sensitive at low levels than high.

Guessing- I tend to guess that your "photosensor" is such a photoresistor. Well, it has lots of differences from silicon photodiodes. For scientific measurements, it has distinct inferiorities. Their response speeds are usually around a millisecond, and sometimes at low light- and carrier-levels over 10 milliseconds. So it's my guess that this slowness is filling up the deep, sharp valley between adjacent half-sine waves in the ( Absolute Value of [sinewt] ) wave. Then it looks like a sine wave of twice the frequency. (Taking the absolute value and low-pass-filtering is a known way to make a frequency doubler. Another way is to use multipliers or non-linear devices to make the square of the signal. Some of that could be happening here, too, but I cannot tell from your description.)

Also, think: the photodiode sees light from blue (0.4um) to near infra-red (1.1um). The photoresistor only sees light in a narrower band, perhaps 0.45-0.60 um. Your fluorescent tube emits several different wavelengths in that range, with different time behaviors. Colors emitted directly from the electrified mercury gas will track the absolute value of current fairly accurately. Colors emitted from phosphors on the wall of the tube are often delayed, having any exponential decay rate from 1msec to 1 second.

In short, both your photosensor and your fluorescent lamp have complex and uncertain behavior. Your photodiode has simple behavior; so it can help you understand the lamp. But your photosensor can't help until you do much work to understand it in detail, and perhaps it is not much help then either.

LED's often provide light output proportional to current, and their response speed is usually faster than 0.01millisec. Using square-waves applied to LED's of a few colors, you could check out the response speed of your "photosensor", the sensitivity change at high and low light levels. and it sensitivity with respect to wavelength. Start with a green or yellow LED to match the color range of the photosensor.


Jim Swenson

Click here to return to the Physics Archives

NEWTON is an electronic community for Science, Math, and Computer Science K-12 Educators, sponsored and operated by Argonne National Laboratory's Educational Programs, Andrew Skipor, Ph.D., Head of Educational Programs.

For assistance with NEWTON contact a System Operator (, or at Argonne's Educational Programs

Educational Programs
Building 360
9700 S. Cass Ave.
Argonne, Illinois
60439-4845, USA
Update: June 2012
Weclome To Newton

Argonne National Laboratory