Shop OBEX P1 Docs P2 Docs Learn Events
Help needed for choosing appropriate sensor — Parallax Forums

Help needed for choosing appropriate sensor

nevizammnevizamm Posts: 11
edited 2014-05-26 22:48 in General Discussion
Hi guys,

Can anyone please suggest possible sensors which I can make use of to measure the dimensions of Playmobil plastic parts such as hats, caps and wigs the fastest way possible? I was thinking about vision cameras but I don't know if it is the optimal solution.

Thanks for your time.

Regards,
Neville.

Comments

  • GadgetmanGadgetman Posts: 2,436
    edited 2014-05-26 04:29
    I thought Playmobil parts all have standardised sizes?

    Exactly what are you trying to do?
  • nevizammnevizamm Posts: 11
    edited 2014-05-26 04:33
    Yes they do but I need a sensor to make sure that these parts are within the specified tolerances.

    Thanks for your help :)
  • T ChapT Chap Posts: 4,223
    edited 2014-05-26 05:17
    How do you want to measure them? Are they randomly placed on a conveyor? Can you physically place them on a jig to hold them in a fixed position? You have to first define what your options are for the physical manipulation, that will affect what options exist for the actual measuring method.
  • nevizammnevizamm Posts: 11
    edited 2014-05-26 07:02
    What I am actually going to measure is the inside diameter of the wigs/hats and make sure that these are within tolerance. Lets assume that a jig is already available. The measurement of the diameter is currently being done manually using gauges. So I need to come up the best possible measurement sensor for this application.

    Thanks for your reply.
  • T ChapT Chap Posts: 4,223
    edited 2014-05-26 07:58
    More info please. Can the sensor use a method of contacting the material(is the material stiff enough to receive some pressure from a switch etc)? If not, is the material conductive or able to be used as a dielectric(so that a sensor only lighting contacting the material can be used). Otherwise, some type of optical method may be used. For example, if the material is reflective of IR light, then an IR LCD could be moved along a path, with the LED masked off so that only a tiny hole remains for the light to pass through. The light(or pulsed light) reflects off of the surface and returns to a sensor(ie 38k detector). When the light stops returning, you have reached a point that can be measured by some means, perhaps just counting the steps of a stepper controlling the sensor's motion will provide the distance, or adding an encoder to the stepper will do the same thing. However, this scheme assumes there is a fall off point of reflection, and that there is a consistent reflection prior to the fall off.

    Without more details, including a drawing or photo, any attempts to offer advice is just speculation. You never stated what the tolerance is. Since you are already manually handling the hat to put it on a jig, maybe the jig itself is the crude measuring device alone. The hat moves down a graduating center section, when the hat no longer moves downwards, you look at marks on the jig to see where it stopped.
  • nevizammnevizamm Posts: 11
    edited 2014-05-26 08:08
    Yes after being injection moulded the parts are relatively stiff so a method of contact can also be used. The diameter tolerance is typically between 15.05mm and 15.15mm. Regarding the infrared light, I believe that since the parts can be black, the IR LCD cannot be used. I think that a probe can also be used, even though the speed of measurement is not as high as that of a camera. (The main parameters which I am considering are cost, speed and accuracy).

    Hope this helps.
  • nevizammnevizamm Posts: 11
    edited 2014-05-26 08:24
    The tolerance is 0.1mm (diameter of the hat should be between 15.05mm and 15.15mm). My mistake I though I specified it. These are two of the parts which I was referring two.

    IMG_20140526_171007.jpg
    IMG_20140526_171007.jpg
    1024 x 768 - 49K
  • T ChapT Chap Posts: 4,223
    edited 2014-05-26 08:44
    Injection molding has the problem of shrinkage, and I found that slowing down the time from inject to eject made for more consistent results. The other problem is distortion of the object during cooling. Are you trying to measure a single direction or X and Y? .1mm is a tight tolerance for a toy.
  • nevizammnevizamm Posts: 11
    edited 2014-05-26 08:46
    No its just single direction measurement. Yes 0.1mm is tight but unless its not tight, the wigs/hats will easily fall off the head of the toy.
  • ercoerco Posts: 20,256
    edited 2014-05-26 08:52
    0.1mm is tight tolerance? Not by Lego's standards! They have the best precision on the planet for plastic toys. From http://en.wikipedia.org/wiki/Lego:

    Each Lego piece must be manufactured to an exacting degree of precision. When two pieces are engaged they must fit firmly, yet be easily disassembled. The machines that make Lego bricks have tolerances as small as 2 micrometres.
  • T ChapT Chap Posts: 4,223
    edited 2014-05-26 09:09
    http://www.keyence.com/products/measure/contact-distance-lvdt/gt2/index.jsp

    http://www.itargetsensors.com/html_products/spring-load-LVDT-position-sensor-62.html

    There are various precision QC and automation sensors on the market. In your case, you likely need to place the hat on a jib, then enter the hat opening from some angle with the sensor in one direction, this will require some type of linear actuator for an automatic process. Since you are not specifically mentioning the Propeller, I will suggest that the Propeller would be an easy platform to set this up with.

    1. Move the actuator to the plastic surface
    2. Read the sensor position for change.
    3. On a change( contact ), stop the actuator, reverse the actuator until the sensor is back to zero,
    4. Take a reading of the actuator position.
    5. Repeat with separate actuator and sensor for the other side of the part.

    This is pure speculation on my part as to the best way to do it. :)

    http://www.amazon.com/Measurement-Tolerances-Manufacturing-Engineering-Processing/dp/0824701631
  • nevizammnevizamm Posts: 11
    edited 2014-05-26 09:26
    Thanks for your help. All of these suggestions seem to be feasible. If any other suggestions come to mind please do let me know.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2014-05-26 09:30
    How many parts per minute do you have to inspect? This may help to decide between contact vs. non-contact methods.

    -Phil
  • T ChapT Chap Posts: 4,223
    edited 2014-05-26 09:34
    BTW, as you mentioned that you want to datalog the results, that is a simple matter with the Propeller either by storing to SD card or sending the result to a PC. If you have a dedicated PC for this project, then I would make a GUI for the PC and have it talk to the Prop. The Prop would then drive the steppers for each actuator, keep track of the encoder on the steppers, read the sensor, report back to PC. For precision measuring with a linear actuator, this will require an actuator with ball screws for long term precision(less wear) with some regular calibration against a fixed dimension(an aluminum hat precision ground?). Sounds like a fun project to tackle.
  • nevizammnevizamm Posts: 11
    edited 2014-05-26 09:38
    Tell you the truth this is just a vague proposal for a project and I have not made any calculations yet. What I need to see is how many pars are being inspected manually and try to reduce it as much as possible.
  • dredre Posts: 106
    edited 2014-05-26 13:38
    Hi,

    Would this firm's product line be of interest to you?

    http://www.laserlinc.com/

    cheers, David
  • nevizammnevizamm Posts: 11
    edited 2014-05-26 22:48
    This is actually very interesting. Thanks David.
Sign In or Register to comment.