I want to build something like a fog-sensor.

I'm using a laser and on the other end (~20m) I have a LTR-536AD photodiode. I hooked the photodiode to an arduino and was monitoring the voltage.

My problem: When there is ambient light the analog reading is ~640. When I hit the sensor with the laser it is about ~620.

How can I increase the resolution?

  • 1
    \$\begingroup\$ Please show a circuit and explain your system a little more thoroughly. When you say, "my analog reading is x," that's kind of meaningless without more context. Also, the LTR-536AD is an infrared photodiode... are you using an infrared laser so that the wavelengths are relatively matched? \$\endgroup\$ – user205769 Nov 28 '18 at 21:33
  • \$\begingroup\$ 640 what? Elephants? \$\endgroup\$ – TimWescott Nov 28 '18 at 23:26
  • \$\begingroup\$ On a side note, if you're using an IR laser remember that you won't have a blink reflex from direct light hit or specular reflections and neither will anybody else that happens to (unknowingly) walk into the beam. Retinal burns from NIR laser light are generally permanent, and even some NIR LEDs are now strong enough to cause problems, especially at night when pupils are dilated. \$\endgroup\$ – isdi Nov 29 '18 at 3:58
  • \$\begingroup\$ @TimWescott 620 units... one unit is 4.9mV so it some voltagereading from arduino. \$\endgroup\$ – Felix Kunz Nov 29 '18 at 17:02
  • \$\begingroup\$ You just need to post a schematic. There's not enough information here. \$\endgroup\$ – TimWescott Nov 29 '18 at 17:16

You need to measure only the change in some property caused by your "fog". This typically is a change in intensity, converted to a voltage by a photodiode and a transimpedance amplifier and measured. You're sensing changes in \$\mu_s\$ (scattering coefficient) which results in changes in intensity.

You need to eliminate any effect not caused by the "fog" - change in ambient lighting etc.

Some suggestions are:

  1. Chose a wavelength where intensity of the detected beam is maximally changed (@Janka correctly suggests a short bluish wavelength - but this may not be possible as sources of blue laser light are more expensive)
  2. Filter out light that is not of the above wavelength - if you're using blue light, any red light at the detector is noise and needs to be rejected. Use an optical filter to eliminate it. The narrower the better, but narrowband filters are expensive. (Need to make sure that temperature change in the laser doesn't change the wavelngth to wander out of the passband.
  3. Modulate the light. This allows you to ac couple the amplifier and now dc offsets are not a problem. (Don't want to measure a small change on top of a large offset) When you detect the light use a lock in amplifier and you can narrow the detection bandwidth (ie lots of averaging) so your SNR improves as the BW decreases. You can modulate a frequency which is not related to the mains power and can eliminate interference from room lighting. Avoids \$1/f\$ noise in the detector.
  4. Alternate between signal and background. Measure your detection background for a second, turn your laser on then measure again for a second.
  5. Maximize your signal. Use several mirrors to make the path length thru the fog as long a possible.
  6. Make you light source a bright as possible. Then you are well away from thermal noise sources in the detecting circuitry. You want to be shot noise limited.
  7. Collimate the light from the laser - use a lens to take that parallel rays onto the photodiode - to make the angular field of view (what the photodiode sees) only that of the source.
  8. Put it in a box where the fog can enter but EMI or stray light cannot. (Just like a smoke detector)
  9. Make everything solid. The system shouldn't be sensitive to temperature, vibration etc.
  • \$\begingroup\$ Good ideas, i also was thinking of some of your points. But my problem istn the stray light. When i fully cover the sensor i get arround 660, so it is mainly a problem of resolution. i want to scale the region between 660 and 620 \$\endgroup\$ – Felix Kunz Nov 29 '18 at 16:58

Use a lens tube to keep the ambient light out.

In addition, a better way for fog measurement is using a blue LED and measure how much the light is diverged by the fog. You can achieve this by using a staggered line/spiral of several photodiodes. This gives you a more accurate measurement. A cheap camera module is an alternative nowadays.

(Why blue? Because it's diverged the most.)


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.