Light sensor to sense light level

Question

I have an application where I am trying to monitor a defect in a web based manufacturing process.

This defect can be seen if one side of the substrate is lighted and the transmitted light is viewed from the other side. When the defect occurs a change in the transmitted light is noticed.

I am looking for a commercial sensor which I can use to give me an analog signal.

I also want to experiment with a sensor on a breadboard to get an understanding of how a change of light translates into an electronic signal.

My application will require a very low response time as the speed of my process will result in the defect being in front of the sensor for less than a millisecond.

Any direction on sensors or components will be appreciated.

Answer 1

For stability and accuracy, I would suggest a silicon PIN (P-Intrinsic-N) photodiode. Response time in the microseconds is not too difficult with good design (you'll want the response to be an order of magnitude faster than the time the sample is in front of the sensor).

Phil Hobbs has a good article here on photodiode front end design. You may just need a simple transimpedance amplifier as this one using the OPA128 (there's a mistake in the schematic on the datasheet- the resistor should go to the PD bias voltage, hey nobody's perfect).

if there is plenty of light available (so you can use a really low capacitance PD), but if you need more, that article (and his book) are a good guide.

Answer 2

CdS light-dependent resistors have realtively high sensitivity and are easy to use, but generally too slow to measure something only 1 ms in duration. Sensors based on semiconductor diodes are probably the most appropriate. There are also photo-multipliers, but those are more complicated and usually only appropriate when the light levels are very low.

Semiconductor diodes can be used two different ways to sense light. These are forward mode where the diode is used like a solar cell, and reverse mode where light effectively causes leakage thrue the reverse biased diode. Reverse mode is more linear, but usually needs considerable amplification to get a useful signal. You can also get a phototransistor, which works like a photodiode but includes some transistor action to multiply the raw leakage current by the gain of a transistor, all in one device.

You also need to consider the spectrum of what you are trying to sense. The particular wavelength may dictate a particular sensor. Note that in a pinch LEDs can be used as photodiodes. These usually respond to a narrow range of wavelength around what they emit in normal LED mode.

Answer 3

A photodiode should give you the analog signal you need. As for an exact component you would need to weigh up the amount of illumination, wavelength and probably other things to select the best component.

Answer 4

Can't argue with using a PIN photodiode as others have said. Consider another problem though....

Ambient light and variations in the light level emitted by the illuminating device. You need to factor these into your design.

I'd use a low power laser (divergence angle to suit your requirement). I'd consider pulsing it to overcome ambient light problems and I'd use a laser with inbuilt photodiode so that you can accurately control the light emitted.

This should work better and more consistently than an unmonitored continuous light source.

One more thing, for consistent accuracy try and control the temperature of the receiving photodiode or consider a two phase measurement where you take a reference light intensity measurement of the laser without the "thing" you need to measure being in position