I'm trying to develop a device that can detect or "watch" an infrared beacon that is freely moving side-to-side anywhere within a 180° field of view (-90° to +90°, with center/forward at 0°); i.e. the sensor should see the beacon if it is directly to the left (-90°), to the right (+90°), or anywhere inbetween, but not if the beacon is anywhere behind the sensor.

I've glued a little IR sensor (TSOP38238) on top of a micro servo (Tower Pro SG92R), and I have these wired up to a Circuit Playground Express with CRICKIT. I'm programming on the board via Arduino, and the software seems to be working great -- the servo will continuously sweep the full field of view, and if the specific IR signal I'm transmitting (38kHz, NEC IR protocol with 32-bit payload data) is received by the sensor, the servo's current position/angle is logged immediately.

However, my problem is restricting the IR sensor's vision. For example, if the servo with IR sensor are facing left (-90°), and the IR signal is being transmitted from front and center (0°), the IR sensor detects that signal and thus determines that the source of the signal came from the left (-90°). This example is real and practical, but the problem is in fact far worse; the sensor is detecting IR transmissions from all 360° -- there is no direction the servo and IR sensor can face that it will not detect the source IR signal being transmitted!

The only thing I know to do is to fabricate blinders to surround the IR sensor. If I can somehow block (reflect, absorb, whatever) any IR signal to which the sensor does not have a direct line-of-sight, then it should help prevent these undesirable signals. Or even if I can't block the signal, perhaps it's possible to corrupt it to cause a change in frequency or payload data.

I've constructed some blinders using craft sticks and super glue, with several layers of aluminum foil (like you use in the kitchen) taped to the outside. I added the aluminum because of this SE thread I found. But still, the IR signals seem to pass right through.

So what other techniques or materials are possible here to restrict IR signals to line-of-sight?

The blinders I constructed look like the following. The wooden side faces the inside (faces the IR sensor), aluminum side faces outward (faces the world):

Outside of panel 1, inside of panel 2

And the following is how the blinders are mounted onto the servo, with the IR sensor inbetween. The intent is to blind signals coming from the sides, but still allow the signals to reach the sensor if coming from different heights:

Parallel panels as mounted on servo

  • \$\begingroup\$ With an estimated 80dB of AGC gain the Rx reflection threshold can easily span a 100:1 range. If your backside range is only 2:1 distance of the front side you can see the problem of detection by error rate. Even if you reduced the Tx current to 1% of the existing current you will still not get decent proximity detection without false positives or true negatives. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Apr 7 at 19:39
  • \$\begingroup\$ Maybe I should have pointed out I'm brand new to electronics, as that comment went clear over my head starting with the first sentence \$\endgroup\$ – ardnew Apr 7 at 19:46
  • \$\begingroup\$ AGC is automatic gain control like that used in radios. 80 dB means a 10,000:1 dynamic range in signal levels, If you can define the actual Geometry in your question, it may be possible to solve this. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Apr 7 at 19:50
  • \$\begingroup\$ @ardnew if you look at my answer you will see that AGC is linked (as wall as COTS) to wikipedia pages explaining what they mean. Beyond that, I don't think you need to know much more then what you want are the type of detectors that count people (the type I linked to). Not the type of detectors used for remote controlling your TV set (the type you have). \$\endgroup\$ – st2000 Apr 7 at 21:36
  • \$\begingroup\$ @ardnew, I should add, once you get the correct detectors, you can likely either use much smaller blinders or simply do away with them. Also, since you are new to stack exchange, you should eventually pick a correct answer so as to let the community know what worked for you. The point being you are setting things up so that the next person asking this question will already have the correct answer. \$\endgroup\$ – st2000 Apr 7 at 21:39

Almost with out exception COTS IR remote control receivers contain an AGC circuit. The Vishay part used contains such a circuit (see block diagram at the bottom of page 1). This feature allows for dependable detection regardless if the IR transmitter signal is received directly or if received bounced off a wall.

What is needed in this application is a beam interruption detector. These are less common but Vishay makes several including this one. Note that there is no AGC circuit in the block diagram. When the signal drops below a certain threshold the output goes inactive.

  • \$\begingroup\$ Although it has no AGC, it probably uses a precision limiter that handles the also huge irradiance max:min ratio of 50k : 0.4 mW/m² or a 100 dB range \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Apr 7 at 19:37

I have solved this problem before using a PIC to transmit unique TDM spaced codes to each 5mm IR LED and then recessed IRDA Rx's going to a PIC.

First, you must define the room geometry, distances involved, range of motion, reflection materials and distance, then how you want to detect orientation by some truth table and desired Tx locations. Like North-East-West-South or front-back?

IRDA and IRDA2 Rx-TX optos are designed for speed but short range but can be made to work for a small room at a suitable pulse duration. There is no need to use the actual IRDA protocol as I invented my own protocol with a unique Tx code.

Since optical path loss is a point source with power spread by the square of radius distance, the signal reduces to 10% or -20 dB for a change in distance of any units from 1 to 10x the range. (The same applies to RF aka Friis Loss) for line-of-site.


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.