3
\$\begingroup\$

I am currently working on a project that uses an IR LED and a photodiode in a proximity sensor configuration (shown below). My question is, I have read some sites that state that your human body reflects IR light transmitted by the IR LED hence when you touch or move your hand close to the proximity sensor the circuit will activate. But other sources say that the human body does not reflect IR light but rather refracts it. So which is correct does the IR light emitted by the IR LED gets reflected or refracted towards the photodiode when your finger or hand moves closer to the sensor?

The second part I noticed is that even if the IR LED is removed the sensor still activates because our bodies do also emit radiation which the photodiode can detect. Provided the threshold voltage is low enough. So then what's the point of having an IR LED in the circuit if the circuit already activates when a finger gets closer to the circuit, or am I understanding this wrong?

N.B The circuit I built is going to be used to detect when a finger or hand gets closer to the sensor.

enter image description here

\$\endgroup\$
  • \$\begingroup\$ ”bodies do also emit radiation” Is your photodiode that sensitive? Is it black in order to block out visible light? \$\endgroup\$ – winny Sep 6 '18 at 15:16
  • \$\begingroup\$ Refraction is the effect of light (or EM radiation in general) bending through a lens, the lens is something that is at least partially transparent to the light (or EM freq) in question. The human body emits IR due to its temperature (black body radiation). \$\endgroup\$ – Oliver Sep 6 '18 at 15:17
6
\$\begingroup\$

To what's already been said here, I want to add a bit more.

  1. Human body temperature is about 310 K. The blackbody spectrum at 310 K peaks at about 10 um: enter image description here (source)
  2. The photodiodes used in this kind of circuit is typically a silicon device, with a bandgap energy corresponding to about 1.1 um. The responsivity falls off strongly at longer wavelengths: enter image description here (source)

    So this detector will not have any practically useful (or confounding) response from the radiation from the human body. In fact, if it did, it would also be detecting radiation from other nearby objects of similar temperature, such as the PCB it's mounted on, the metal or plastic housing around the actual photodiode, etc.

    1. The reflectance of human skin depends somewhat on the color of the skin in question, but at the near-IR wavelengths that will most affect your photodiode, we're all about the same. We are not great reflectors, but probably good enough to have a measurable effect on your sensor: enter image description here (source)
\$\endgroup\$
  • \$\begingroup\$ Once again, earning your pseudonym! \$\endgroup\$ – Scott Seidman Sep 6 '18 at 20:59
  • \$\begingroup\$ You correctly say that at 310K the peak is near 10µm, but the illustration shows the default setting from your source with 5700K from surface of the sun. I cannot seem to upload the correct screenshot here... \$\endgroup\$ – Jürgen Weigert Jan 27 at 19:14
5
\$\begingroup\$

The detector must have a black lens which passes IR and blocks stray visible light. aka "Daylight Blocking Filter"

Skin reflects some IR, even if you are black-skinned. It also conducts IR so it is semi-transparent, which is good for finger sense heart rate and oximeters.

But since the dielectric constant at this wavelength is greater than 1 it also refracts. ( plastic is higher than glass so they can use thinner lenses for eye-frames)

But skin is also an irregular surface compare to the wavelength, so it also scatters light.

We also generate IR as heat, but this is a longer wavelength than most IR LEDs and PD's can detect. FIR or Far InfraRed is (>5um) but I'm not sure of the peak wavelength.

The body also absorbs infrared heat both short IR or SIR from tungsten heaters and far IR or FIR from our partners bodies. It gives " black body effect" a whole new meaning to some, but some researchers, including myself, know of the healthy benefits of FIR and some say married couples live longer if they sleep closer. All Doctors of Acupuncture know this because they use FIR lamps.

Another play on words is "Skin effect" because we know the longer wavelengths penetrate deeper, an important quality of FIR for therapeutic reasons. Often it is synthesised with many mineral phosphors including Germanium with a short IR back heater.

I would call it a "diffuse surface" which gives lower resolution monochrome images or scattering of incident light at and below the surface.

\$\endgroup\$
4
\$\begingroup\$

Without giving you the actual numbers here I can tell you from my own experiences with such IR circuits that the IR light is reflected from your body and that your body does not emit enough IR itself to fail this type of circuit. But sunlight IR is a big problem that you need to consider.

\$\endgroup\$
  • \$\begingroup\$ For the sunlight part it trail and error. I'm going to tabulate various thresholds and record which threshold value give me the most optimal performance so that part I'm not to worried about \$\endgroup\$ – Joey Sep 6 '18 at 15:36
  • \$\begingroup\$ I made best experiences with transmitting some IR pulses. Then you get almost rid of sunlight IR. (That’s what is used in remote TV controls). \$\endgroup\$ – Stefan Wyss Sep 6 '18 at 15:43
2
\$\begingroup\$

The second part I noticed is that even if the IR LED is removed the sensor still activates because our bodies do also emit radiation which the photodiode can detect. Provided the threshold voltage is low enough. So then what's the point of having an IR LED in the circuit if the circuit already activates when a finger gets closer to the circuit, or am I understanding this wrong?

Provided the threshold is low enough. When that happens, the circuit will also trigger from things like stray sunlight.

So the threshold is set higher than the highest ambient illumination you think is likely, and this makes it impossible to trigger off your body's emitted IR. Then you need an LED to provide even more illumination so the circuit can trigger at all.

Note that putting an IR filter in front of the detector helps, since it eliminates ambient radiation in the visible, and improves the amount of unwanted power falling on the detector.

\$\endgroup\$
2
\$\begingroup\$

Like any imperfect mirror, some IR light will reflect, and some will be absorbed. I don't, off the top of my head know the ratio -- but even if most is absorbed, some is still reflected. Thus, if the ONLY optical path between the LED and the detector involves a reflection off a person, the detector will become MORE activated -- assuming the person is not blocking a bigger source of IR light, such as the sun.

Many detector circuits modulate the LED, and then you look for a change in the detector at the modulated frequency. This way, you know you're only looking at changes involving the light source of interest.

\$\endgroup\$

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.