Connect with us

Movement sensor

Discussion in 'Sensors and Actuators' started by Inverness, Jan 11, 2019.

  1. Inverness

    Inverness

    12
    0
    Oct 30, 2018
    Hi.
    I have a project with a very specific usage in mind.
    A coin falls through a hole that is just big enough for it, into a room with an acrylic glass wall on it.

    I want to detect when the coin falls through. I'll be only one coin at a time, and there shouldn't be one following so closely that a sensor should have problems picking it out.

    Since it lands in a "room" I'm not keen on using sensors that tend to be very broad, such as PIR. It may trip from someone walking around since there is a clear section, or from coin movement in the room after it drops through the hole. I could use some sort of cone or similar to limit its "sight" but it's not a very permanent solution

    I've thought about a small switch, but I'm having trouble imagining how I could fit it. There's also a range of different weights to the coins and I'm afraid that it's possible the coin is light enough that it may not trip the switch.

    Laser based detection would be optimal, but I've only found tutorials where you need to correctly position the laser diode to shine on a photoresistor. For this purpose that's too fiddly and some might see lasers as a risk when eyesight is concerned. (It's not that I personally wouldn't use lasers, but for what I am making it isn't optimal)

    I'm thinking of photoelectric sensors in some form. Something like this would be perfect, but it costs a lot and uses proprietary interfaces to work normally. I'm also not confident in my knowledge of photoelectric sensors that this would work. I've been looking at this. But I'm not confident in how they work, or if it's the same case as the PIR, with a very broad detection range.

    For size reference, this is not a gigantic machine. It's meant to sit on a desk and it's about 30x30x25 cm/12x12x10 in (last number being height)

    Is there something I could try? I intend to use an arduino mega to power the other electronics of this project.
     
  2. Harald Kapp

    Harald Kapp Moderator Moderator

    9,219
    1,856
    Nov 17, 2011
    The "switch" can be constructed from two elastic sheets of metal or wire to both sides of the coin slot. when the coin falls through, it connects to both elements simultaneously, thus closing the contact.
    You'll need a simple debounce circuit (Google that) which in a simple form could be a monostable multivibrator based on a 555 timer chip (lots of schematics on the net for that).
    Without debouncing, the coin will make multiple contacts while passing through the slot which will give you false readings.
    Easy to build from scratch: you'll need a light emitter (a lamp, an LED, an infrared LED if the light shall not be visible) and a receiver. The receiver circuit is known as "photoelectric sensor". It can also be built from a 555 chip, but here too you'll find lots of circuits on the net.
     
  3. Externet

    Externet

    711
    145
    Aug 24, 2009
    Last edited: Jan 11, 2019
  4. Inverness

    Inverness

    12
    0
    Oct 30, 2018
    Thank you for your reply. I've thought about using something like this, but I failed to mention that the coin slot in question is horizontal. The coin passes over a horisontal hole and needs to fall without any slowdown. But might be applicable to other parts of the project, great to know you need something to filter the signal. The second suggestion holds promise. I see in the circuit you link to uses an LDR to measure the light. Would this mean that ambient light could potentially trigger a false positive? Is there an LDR that is only sensitive to IR maybe?

    Thank you for your reply. This is what I want but as explained above I didn't mention that the slot is horizontal, so the gap wouldn't be big enough. Building my own might be the solution.
     
  5. Bluejets

    Bluejets

    3,452
    698
    Oct 5, 2014
    Don't like to rain on your parade but gravity doesn't have any horizontal effect.

    For the IR opto if any slot is unacceptable, a reflective unit could be used but adjustment will have to be allowed for depending on the color of the coin.
    As an example a reflective surface (mirror) on the opposite side of the sensor and arrange the sensor to operate in the opposite mode. i.e. "loss" of ir light, coin detect.
    I assume coin will be silver or gold but I'm willing to bet requirement will be for something else like corroded copper.
    Lastly there does not appear to be any requirement that would involve any microcontroller.
    Also, microcontrollers should not be looked upon as a source of supply to any other circuitry except low current demand sensors and the like.
     

    Attached Files:

    Last edited: Jan 11, 2019
  6. Harald Kapp

    Harald Kapp Moderator Moderator

    9,219
    1,856
    Nov 17, 2011
    ;)

    The orientation of the slot and its size are independent from each other.

    The more an optical solution as suggested in the second part of my answer is in order to minimize any friction. An optical sensor can easily interfaced to an arduino.
     
  7. hevans1944

    hevans1944 Hop - AC8NS

    4,214
    1,994
    Jun 21, 2012
    Please tell us what you are trying to DO with regard to a coin and a horizontal slot through which the coin somehow passes.

    Do you want to detect when the coin has completely passed through the horizontal slot and is in the process of dropping into the enclosing "room," or do you want to detect when the coin has dropped some distance below the horizontal slot
    (presumably under the influence of gravity), or do you want to simply detect the presence of the coin in the horizontal slot without regard to what happens to the coin later? Why do you want to detect the coin? How is the coin inserted in the slot? What size is the coin? Is the coin round, or can it be some other shape? How thick is the coin? How heavy is the coin? Are there coins of different denominations to detect? Does the denomination of the coin matter? Are the coins legal tender, or are the coins merely tokens with no intrinsic value? Helpful minds want to know the answers to these questions.
     
  8. BobK

    BobK

    7,607
    1,648
    Jan 5, 2010
    I might not understand the problem correctly, because this sounds absolutely trivial to me. All you need is an LED and some form of photodetector placed such that the falling coin interrupts the light from the LED to the detector. Am I missing something?

    Bob
     
  9. hevans1944

    hevans1944 Hop - AC8NS

    4,214
    1,994
    Jun 21, 2012
    I think we all are. That's why I asked all those questions in post #7, just to try to define WTF the problem really is if a simple LED and photodiode or phototransistor optical interrrupter or reflector configuration is unacceptable to the original poster. And then there is the mention of an LDR (Light Dependent Resistor, commonly implemented as a cadmium sulfide or CdS cell) by the OP in post #4. The following quotation leads me to believe the OP doesn't know diddly about optical sensors or the relatively slow response times of LDR devices:
    Poorly defined problems almost always lead to poorly implemented solutions. A better description of what the OP is trying to DO would help define the problem and possibly point to an obvious solution. Maybe something as simple as a load cell that would detect the weight change in the "room" contents when each coin is dropped would be an acceptable solution.
     
    Harald Kapp likes this.
  10. Inverness

    Inverness

    12
    0
    Oct 30, 2018
    This is what I said to the optical solution offered by Harald Kapp, and my questions regarding it.

    No, thank you, I don't. Which is why I am here.

    It matters that the slot is horizontal as opposed to vertical, because a vertical slot means that the device linked by Externet would be perfectly suitable.

    A coin on top of the surface is slid over the hole which it drops through. The coin needs to be registered and counted. There are different coins of different denominations that are legal tender. But the slot sensor doesn't need to differenciate between different coins because I have one slot for each coin. A spinning plate picks up one coin at a time and slides it over the holes, starting with the smallest size first. If it doesn't fit, it doesn't drop. I just want to register each coin once in order to count it. This is also why I am using an Arduino. See attached image.

    I have never said that such a solution is unacceptable.

    Since "OP doesn't know diddly about optical sensors" I have some questions as outlined in post #4. The circuit linked to in post #2 features an LDR, but I'm wondering what exists in optical sensors, and trying to narrow the selection down by trying to state my concerns or possible issues.

    This is not an entirely enclosed chamber. It has a clear section to allow viewing. As such I wonder if whatever optical sensor you use is sensitive enough to ambient light to trigger a false positive. Is there something I could do to mitigate this, besides enclosing the chamber/room?

    The largest hole is 27,5 mm/1.10 in across. Does a simple LED have enough intensity to reliably be detected by an optical sensor at this distance?
     

    Attached Files:

    • A.PNG
      A.PNG
      File size:
      315.7 KB
      Views:
      4
  11. Harald Kapp

    Harald Kapp Moderator Moderator

    9,219
    1,856
    Nov 17, 2011
    I don't see why the device linked by Extenet should not work in a horizontal position.
    The coins fall vertically in your image. They are oriented horizontally, but that is not the same as your post #5 made us believe.
    Why the need for "no slowdown"? According to your image the coins fall into a box. The orientation of the fall is seemingly irrelevant. You could build each opening in such a way that the coun's orientation is changed from horizontal to vertical after passing the hole (in a kind of a funnel) and then pass the photoelectric sensor. The speed of the distributing wheel can be adjusted to account for the small delay incurred by thie mechanism.
    That is not a major issue with the kind of sensor proposed. The coin will interrupt the light beam from the emitter (LED) to the receiver (Phototransistor, LDR, whatever). Ambiengt light will increase the light falling onto the sensor, which will not trigger ther sensor. As long as the coin interrupt suffcicient light, be it from the emitter or from an ambient source, the sensor will register the coin.
    With a setup as proopsed where the coin passes a narrow slot between transmitter and receiver. the distance to the hole or the diameter of the hole (or the coin for taht matter) are irrelevant as long as the coin is bigger than the light beam from the emitter, whci it is in this case.
     
    hevans1944 likes this.
  12. BobK

    BobK

    7,607
    1,648
    Jan 5, 2010
    An LED and a phototransistor would be my choice. IR if the light would be visible through the window and you don't want that. If I understand the geometry of this thing, the LED / phototransistor would see the coin edge on, maybe as little as 1 mm. In this case, the detector will probably need to be shielded by a narrow slot to guarantee that it is completely blocked by the falling coin.

    And yes, an LED can be detected by a phototransistor at well over your 27.5 mm. The detector should be shielded from any other direct light, so it it only illuminated strongly by the LED itself.

    Bob
     
    hevans1944 likes this.
  13. Externet

    Externet

    711
    145
    Aug 24, 2009
    Not that matters much for detection; but with such emphasis in 'horizontal slot' , clarify :
    if the slot is horizontal on a vertical panel (slide-in/push-in)
    or the slot is horizontal on a horizontal panel (drop-in) Seen from above ( - ) or ( | )
    The picture that did not post at #3 before :
    ----> http://www.vinland.com/Images/Opto-Interrupter.gif
     
    hevans1944 likes this.
  14. Inverness

    Inverness

    12
    0
    Oct 30, 2018
    In either case, I don't want to add a slowdown (in the distributing wheel) because no matter how insignificant it is in itself, it adds up. I probably have to build a funnel or some kind of catcher. As for something flipping the coin, it adds a level of complexity I don't want to deal with, possibility of jams and so on. I have an idea for a solution and I'm gonna try to make a model for it.
     
  15. BobK

    BobK

    7,607
    1,648
    Jan 5, 2010
    The light pressure from the LED will not slow it down.:)

    Bob
     
    Harald Kapp likes this.
  16. hevans1944

    hevans1944 Hop - AC8NS

    4,214
    1,994
    Jun 21, 2012
    @Inverness: Thank you for your clarification! The 3D CAD concept drawing is also appreciated. I will try to answer the questions you have posed so far, but I am sure there will be other questions and answers brought by you, as well as by other responders. That is the nature of a dialog and the way this forum works best IMO.

    When you said the coin slot was horizontal, as opposed to a vertical orientation, I envisioned the coin slot on a typical vending machine (which is usually in a vertical orientation) being rotated ninety degrees to make the coin slot horizontal but still located in a vertical plane. We now know from your description and 3D CAD image that the "coin slot" is horizontal and located in a horizontal plane. The difference may or may not be important to how the coin is detected.

    The optical-interrupter type of sensor that @Externet linked to will work in any orientation, but the opto-electronic components (consisting of an infrared LED and an infrared-responsive photodetector) would probably have to be physically separated from the associated circuitry to provide adequate clearance for the coin drop. Such mechanical separation is shown in @Externet's image in post #13, but this can be carried much further by cutting the plastic joining the two-wire emitter and the three-wire sensor, thus allowing further physical separation of these two opto-electronic components. Or a completely custom configuration of light emitter and photodetector can be made.

    Some edge-guidance of the coin would probably be required, to ensure that one edge passes between the infrared LED and the photodetector. OTOH, if the coin is to be "viewed" edge-on by the opto-electronic device as it falls through an appropriately sized hole (NOT a slot!), then clearly the separation between the LED light emitter and the photodetector must be larger than the coin diameter. Why you would want to place such an artificial restriction on coin detection is a mystery, at least to me, but it can be done.

    How is this accomplished? Does someone move the coin, laying horizontally, with a finger until the coin is over a hole? Or is the coin held between forefinger and thumb and dropped through a rectangular horizontal slot of appropriate length and width, whereupon the coin lands on a rotating horizontal plate which carries it toward various diameter holes that will presumably somehow select coins of different diameters? How is that accomplished, exactly?

    Does your reference to "one slot for each coin" refer to the various sized holes shown in your 3D CAD drawing or does it refer to various sized, multiple, horizontal rectangular slots that are not shown on the drawing? Exactly how does a coin move from the pocket of its owner to end up on the rotating plate? How does the spinning plate center and deposit a coin over an appropriately sized hole? Are you aware of how difficult it is to coaxial align a round coin with only a slightly larger hole so the coin will drop through the hole?

    It would be much easier to measure the coin diameter with a frame grabber attached to a digitized CCD camera while the coin is resting on the spinning plate. Software then identifies and sorts the coins by size, dropping each sized coin into the proper bin, perhaps by opening a trap-door in the spinning plate or by deflecting the coin off the plate and into a chute over the appropriate bin with a solenoid actuator. Several hundred coins per minute could be sorted and counted in this manner.

    Are we re-inventing a coin-sorting mechanism here? That sounds more like a mechanical design problem with electronics just added on at the end to tally the coins that fall through holes. Good luck with that! Wiser heads than you will find here have spent decades refining coin-sorting technology for vending machines. The slot machines in Las Vegas (and elsewhere in the world) couldn't survive without it! Not only do the mechanisms sort coins by value, they also detect many counterfeit coins, which your machine will not do of course. And many will detect "bent" coins that could jam the coin sorting mechanism, and such coin sorters will have a reject lever to remove the "bad" coin and return it to its owner... usually... and if it isn't too bent out of shape.

    But it appears (at least to me) that you are trying to create a desktop novelty with no practical value except for the "Gee whiz" exclamations it may evoke. So, I will take a short divergence here to try to explain how opto-electronic devices work. I looked for such a discussion in our resources section, but found nothing. Perhaps I or some other member of EP will attempt to write about it in some depth later. You can also visit this website for a basic overview of semiconductor optical detectors.

    The use of light, in particular collimated beams of light, to detect the presence or absence of physical objects is almost as old as electronics. Originally, light for this purpose was produced by incandescent light bulbs and was therefore a broad-spectrum "white" light. A simple lens collected the light and directed it as a beam, similar to the beam produced by a flashlight. A photodetector received the light beam, again through a simple lens, and focused the beam onto the sensitive area of the photodetector. As long as the light path was uninterrupted, the photodetector could "see" the light source and signal the presence of same. If the light path were interrupted by an opaque object, the photodetector would no longer "see" the light source and would signal that an obstruction had occurred.

    That is essentially all there is to optical interrupter types of detectors. Over the years the technology has changed to offer simpler, cheaper, brighter, or more reliable light sources such as LEDs and lasers as well as a wider range of photodetectors. There are also variations on how the light beam is interrupted: a co-located photodetector with an LED light source can be used with a retro-reflector (or corner-cube reflector) to detect light path interruptions over longer path lengths. The advantage is all the electronics is in one location, and separate wiring for the photodetector and light source is not required. Also, the range is easily extended with plane mirrors.

    Early photodetectors were cadmium sulfide (CdS) light-dependent resistor (LDR) cells. These are still dirt cheap, very sensitive to changes in light intensity, and simple to make, but they have many disadvantages. I won't go into that here, but LDRs are generally NOT acceptable when fast response to changing light conditions is required. Here is a datasheet for a typical CdS photocell.

    Vacuum photo-diodes were next used, particularly ubiquitously for motion picture sound reproduction. Later developments along those lines include photomultiplier tubes, which are still used today when extremely low light levels must be detected with very fast response times. An excellent description of the huge variety available can be found in this Hamamatsu PDF manual.

    The real breakthrough technology came with the development of the bipolar junction transistor or BJT. It was soon discovered that PN junctions respond to light, and that transistor action would amplify the effect to useful signal levels. Phototransistors, photodiodes, and photo-triggered SCRs (solid state rectifiers) soon followed. All of these semiconductor devices respond to visible as well as near-infrared wavelengths, their response peaking near 900 nanometers wavelength in the near infrared. If infrared response is the ONLY desired response, a device may be packaged in what appears to be opaque plastic, but the plastic actually passes near-infrared wavelengths quite well while blocking visible light. This helps prevent, but does not eliminate, false positives caused by other nearby illumination sources.

    Opto-electronic devices require some understanding of physical optics to be successfully applied. The illumination field and its intensity must be specified or determined. The required sensitivity of the photodetector and its range of wavelength response must be specified or determined. Most folks involved in electronics do not have the knowledge, experience, or skills necessary to specify or design opto-electronic solutions, but today that isn't strictly necessary. There are plenty of commercial off-the-shelf (COTS) solutions available for virtually any opto-electronic sensing application. The problem is finding something suitable for your particular problem. If you are an industrial customer, vendors and manufacturer's reps will usually be more than willing to help you select one of their products as the correct "solution," but winnowing the chaff from the wheat can be very time consuming. Fortunately there is Google to help with this. Use it aggressively and often to narrow down your search for possible solutions before turning here for advice and help. Keep an open mind: there are many alternative solutions to every solvable problem.

    Getting back to answering your questions...

    Hobbyists use LDRs for many reasons, chief of which is probably cost. The resistance range as a function of light intensity is huge, being a few hundred ohms with strong lighting to several million ohms when placed in totally dark conditions. It can take several hours for an LDR to reach it's maximum dark resistance but only a few milliseconds for the resistance to drop to the minimum resistance when fully illuminated. The resistance versus illumination characteristic is highly non-linear.

    The next most popular optical sensor is the photo-diode. These can range from a device packaged much like an LED to a simple silicon solar cell to a sophisticated PIN photo-diode, the latter being usually operated in reverse-biased avalanche mode from a current-limited source to provide nano-second or pico-second response times to pulsed light sources.

    Your application is probably best served by a common silicon photo-diode packaged like an LED. Photo-diodes have two modes of operation: voltage generation mode and reverse-biased current mode. The current mode is very linear in its response to light intensity whereas the voltage generation mode is simpler to configure and use, especially for on/off type detection applications.

    Phototransistors are also available, consisting of an ordinary NPN transistor with the emitter-base and base-collector junctions deliberately exposed to external illumination by means of transparent packaging at the wavelengths of interest. Some packaged opto-isolators use a phototransistor to improve their sensitivity.

    This is a common problem. There are several different approaches that may yield a viable solution. The first, and easiest, solution is to narrow the optical bandwidth using light emitters producing a specific narrow range of wavelengths, followed by optical bandwidth filters in front of the photodetector. LEDs produce light having a narrow range of wavelengths, as of course do lasers. Red and near-infrared are common emitter wavelengths that are used and colored plastic filters limit the bandwidth of optical radiation reaching the photodetector. Narrowing the optical bandwidth is often effective because ambient light sources are usually broadband in nature. By limiting the range of detectable wavelengths, the photodetector becomes less sensitive to interference from light sources outside that range.

    The most common way to eliminate responses (false positives) from external, ambient, light sources is to modulate the active emitter, usually by turning it on and off rapidly with a square-wave modulation signal, a process called light chopping. The photodetector is equipped with a synchronous de-modulator that responds only to the chopped illumination. Chopping frequencies of several kilohertz are possible with inexpensive emitters and photodetectors, yielding rapid response times in the millisecond range. A block diagram and more information can be found at this website.
     
    Last edited: Jan 14, 2019
    Harald Kapp likes this.
  17. Externet

    Externet

    711
    145
    Aug 24, 2009
    Yesterday saw this gadget at Goodwill for $1. Thought of this thread...
    [​IMG]
    This counts money, not coins. I suppose by the amount of opening the coin pushes at the insertion slot width assigns a value. Hackable to just count number of coins, (if that is the intention) I suppose.

    Hevans : On your post #16, perhaps 'plane' should say panel ? The above picture has a horizontal slot on a horizontal panel.
     
  18. Inverness

    Inverness

    12
    0
    Oct 30, 2018
    This is the idea I came up with. Its purpose is to bulk count coins. I'm working with what I resonably have avaliable, which is a laser cutter. I probably could come up with a design that flips the coin to upright position, but in my opinion it requires more complexity and may suffer jams or failure more easily.

    While this is still a work in progress, the idea is that you drop a bunch of coins into the reservoir seen on top of the box in the picture, where a servo hidden in the middle spins the top layer (the distributing plate)

    It does indeed refer to various sizes holes for each coin. I understand the difficulties with alignment, and tried to design this with that in mind. The idea is that dragging the coin along the wodden surface will align the coin in the middle of the hole it "travels in".

    While the CCD camera idea is something I take note of and indeed a valid solution, its complex and as you can probably note, I am not at the level yet.

    That is 100% what I'm doing but to me it teaches electronics, CAD and design.

    Indeed, but these are not US, or any other common coins like GPB or EUR. And you have to put them in one at a time... I'm also doing this for learning/fun. Though the point isn't only the amount of coins, but their value, combined value, etc.

    Sorry regarding the confusion of slots/holes and horizontal. In any case, if there are suggestions on how to design this better they are more than welcome. I think I'll design a ramp for the coins to fall onto. Then I can add a detector in the ramp, and a light source in some way. This should also completely block the light to the sensor, meaning false positives should be eliminated. Thank you Hevans for an excellent explaination on the types of sensors.
     
  19. Inverness

    Inverness

    12
    0
    Oct 30, 2018
    Well, I know that much! :p
     
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

-