Connect with us

infrared LED

Discussion in 'LEDs and Optoelectronics' started by nsak, May 15, 2012.

Scroll to continue with content
  1. nsak

    nsak

    10
    0
    May 15, 2012
    I have an infrared led that I use to identify a passing object. The light reflects on the object and triggers a IR led receiver.

    My problem is that I cannot restrict the light in a narrow angle. As a result the receiver is also triggerted from stray light. Normally I filter this out as the pulses from my interest object are different from the stray ones. Not always however...

    The question is what is the best approach to restrict the LED light so that it goes only in the intertest direction and does not trigger its nearby receiver. Any cap perhaps that can do the job?
     
    Last edited: May 15, 2012
  2. duke37

    duke37

    5,361
    768
    Jan 9, 2011
    What will a cap do other than keep your head dry?

    An led is small so if it is fitted to a tube the light will be collimated. A tube 3mm diameter and 10mm long will restrict the light to a few degrees. Adding a collimator to the receiver will also help.

    If the transmitter is driven with a pulsed signal and the received signal is locked to this, then the background signal can be greatly attenuated.
     
  3. KrisBlueNZ

    KrisBlueNZ Sadly passed away in 2015

    8,393
    1,268
    Nov 28, 2011
    I guess a parabolic reflector would be best, or failing that, a tube.
    Are you using pulsed infra-red? This is the best way to distinguish your reflected light from ambient light. It's the method used in infra-red remote controls. You pulse the LED at a fixed frequency and use a receiver with a bandpass filter in it. If you choose a standard remote control carrier frequency (36~40 kHz) you can use an IR receiver-demodulator that includes the bandpass filter and AGC and gives you a logic output. This might be too sensitive, though. If you stuck a short tube on the receiver, that might do the job.
    IR remote control receiver devices: http://search.digikey.com/us/en/cat/sensors-transducers/optical-photo-detectors-remote-receiver (click "Optical - Photo Detectors - Remote Receiver").
     
  4. KrisBlueNZ

    KrisBlueNZ Sadly passed away in 2015

    8,393
    1,268
    Nov 28, 2011
    duke37: snap!
     
  5. nsak

    nsak

    10
    0
    May 15, 2012
    thanks a lot guys! Cap, I meant not a wool one, rather fitted with some lens.

    I wonder if you have any providers to suggest to look into for these tube fitted leds; we normally use Farnell but haven't spotted anything suitable

    Yes we use pulsed infra red, 36KHz; everything works fine, only issue is with stray light triggering the receiver now and then

    thnaks again!
     
  6. duke37

    duke37

    5,361
    768
    Jan 9, 2011
    Great minds think alike!

    If you are using a phase locked detector, background light should have little effect. If the background is too strong the detector can go into saturation. A collimator would help in this repect.
     
    Last edited: May 15, 2012
  7. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,388
    2,774
    Jan 21, 2010
    One obvious solution is to modulate the LED with a signal and detect only signals that are of the frequency of your modulation (use a bandpass filter).

    This will increase your sensitivity by eliminating the DC bias changes caused by ambient light.

    All the focussing stuff will help too.
     
  8. nsak

    nsak

    10
    0
    May 15, 2012
    I' m confident its not about ambient light; its about LED light reflecting and straying its way to the receiver
     
  9. nsak

    nsak

    10
    0
    May 15, 2012
    any suggestion as to the tube material/ thickness?
     
  10. duke37

    duke37

    5,361
    768
    Jan 9, 2011
    The way I would do it would be to take some paper and black it with ink or find a black area in a newspaper and cut it into strips.
    Take a 5mm (?) drill, wrap the shank with plumbers PTFE tape and then wind the paper round the drill and glue with PVA adhesive. Several turns would give a strong tube once the glue has set. They used to make chairs from papier mache.
    The PTFE tape should allow the tube to be slid off the drill.
    An alternative to the drill is a wooden dowel, if this is well soaked with water, it will expand and when dried, will allow the tube to be released.
    Attach the LED to the tube with hot melt glue (this can be removed if necessary).

    A wooden block with a hole drilled through it would also do and would perhaps aid the fixing. Put some black ink down the hole, not shiny black paint.
     
  11. KrisBlueNZ

    KrisBlueNZ Sadly passed away in 2015

    8,393
    1,268
    Nov 28, 2011
    duke37, nice ideas!
    Re avoiding bogus triggering, obviously put the emitter and detector as close to the items being detected as possible. Also, if these reflective surfaces are always facing exactly the same way, you could angle the emitter and detector at say 30 degrees off the perpendicular, one on each side of the perpendicular, and space them the appropriate distance apart so that only a reflecting surface that's exactly the right distance away from them will reflect the transmit beam into the receiver tube. If that's unreliable, you could try reducing the angle between the perpendicular and the transmitter and receiver, to make the system less fussy but still able to reject reflective surfaces at different distances.
    What are the objects you're detecting? This sounds like an interesting project.

    As an aside, why do people who ask questions always seem to post the absolute minimum of information about what they want to do? Why say you want to detect "a passing object" instead of "a shiny box moving past on a conveyor belt" or "a car driving past" or something else useful? Not complaining at you specifically, just wondering...
     
  12. nsak

    nsak

    10
    0
    May 15, 2012
    thanks Duke!

    there's a a sort of product development project on going and we need a a bit more of an "industrialised" approach, but I surely get your point and will look in the correct direction.

    I thank you all folks for your stimulating feedback !
     
  13. nsak

    nsak

    10
    0
    May 15, 2012
    Kris, sorry. I just had in mind not to bother you with all sorts of details. I' m only happy to give you further info.

    What we want to do is develop a system that can identify humans stepping in and out of a space (pls note, in and out...) . The module is meant to be placed near the door, entrance, etc; we need to manage some 70- 80 cm of a distance.

    As an altrernative approach we are using a camera; reading pixels and assuming from there changes what the motion is at the entrance. We dont need pattern recognition and stuff like that because we dont have processing capacity for it; the module is very lean, the m/c on it is the one used for radio transmission of data and its cant possibly also manage things like that.

    I' m not sure which approach will be better; I favor the camera as it doesn't restrict me in terms of distance, or compromise me in terms of installation.But there a lot of issues there; we might have to consider adding a further m/c to leave the radio to do its business and not the image analysis stuff. And then what about ...burglars ? (I mean motion in the dark...)

    Happy for any further insight and sorry again for mystic attitude -)
     
  14. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,388
    2,774
    Jan 21, 2010
    When you say humans, are you specifically excluding cats, dogs, trolleys, etc?

    And do you want to distinguish between entering and exiting?

    Is there any reason why you don't want to (say) detect a beam being broken?
     
  15. nsak

    nsak

    10
    0
    May 15, 2012
    Steve, a broken beam does not carry any directional information; i need to know the direction

    also, no dogs, cats and so on
     
  16. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,388
    2,774
    Jan 21, 2010
    One broken beam doesn't convey direction, but 2 do.

    How do you intend to determine direction from your current method?

    How will you discriminate between humans and not-humans?

    Height of the sensor will help in some respects. People are taller than most cats and dogs.
     
  17. nsak

    nsak

    10
    0
    May 15, 2012
    ...there are two beams but I need have the receiver on the same side of the door, not on the opposite, which is where the LED problem starts.

    We have singled out non humans, becase, due to height, the signal is different (if any at all); same can apply to noise although I would be more confident if I could shield the led and avoid the noise
     
  18. KrisBlueNZ

    KrisBlueNZ Sadly passed away in 2015

    8,393
    1,268
    Nov 28, 2011
    Can you place a reflector on the other side of the door? How about two narrow beams of IR light using different modulation frequencies, reflecting off a surface on the other side and received by one or two receivers. If the reflector isn't perfectly smooth there will be some scatter so the reflector will not have to be exactly perpendicular to the beams.
    Have you considered ultrasonic? If the transmitter-receiver is not perpendicular to the direction of entry you can detect direction by the direction of the Doppler shift, but it would be hard to avoid detecting animals and maybe movement inside the room or outside the door.
    When you explained your application in more detail, we were all able to make interesting suggestions. Why not tell us more about the application, including what you've tried and the problems you found. It sounds like an interesting project. Someone here may know a trick you don't know about. All we need is details!
     
  19. nsak

    nsak

    10
    0
    May 15, 2012
    Here' s a summary of what we have tried

    ultrasonic; not working beyond a meter, a lot of reflections. The solution would require two modules making it bulky. LEDs appear more promising more compact and can work with just one one receiver

    pir sensor; many advantages (compact, passive element, …) but sequence in- out was not possible to be identified well in all cases; even when equipped with fressnel lens we would not get a clear signal

    camera; a lot of software to read pixels, and analyze their property changes; we need at least 3 fps and we have not managed that far with the current on board m/c (embedded on a radio chip) to manage this; will are now expanding the design with a supplementary m/c to manage the images and the inference engine; a lot of advantages, a big development pain ...

    ir leds; everything working well, with the exception of stray ir light; Assisted by this discussion we are trying to contain this with tubes,.lenses, etc, althougfh we would be ver6y happy if we could get oursleves a marekt ir led with all thee provisions embedded.

    Still it can not compete with camera in terms of installation easy/ flexibility. Out of the question to use reflectors , etc, on opposite side.
     
  20. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,388
    2,774
    Jan 21, 2010
    If you use modulated IR LED light (perhaps even laser) pointed at a reflector (like a bicycle reflector) then you can eliminate stray light easily.

    You may be able to get away with a single light source and 2 detectors.

    If you try to detect reflections you will have issues with objects of differing reflectivity passing by. The same thing is true to a point with a reflected light beam, however very few people wear highly reflective pants. :)
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-