Shop OBEX P1 Docs P2 Docs Learn Events
Gum separation project — Parallax Forums

Gum separation project

NikosGNikosG Posts: 705
edited 2012-12-04 06:35 in Robotics
Hi everybody,

One week ago, a friend asked me to help him in order to complete a machine that will be able to separate Mastic Gum He sent me some clear sample of gum and some gum mixing with dirty in order to experiment with the separation.
The gum sample has some particularities. It is like a crystal and sometimes is almost translucent. There is a huge and unpredictable variety of sizes and shapes of every piece of gum with different surfaces. This fact has as result a variety of different measurements, even for the same sample, if you change a bit the sample's and the sensor's positions. Using the ColorPAL Sensor
In conjunction with a basic stamp board I had almost 70% success as I tried to recognize the clean and the dirty Gum.

attachment.php?attachmentid=97460&d=1354439631

I wonder if the TCS3200-DB Color Sensor could give me better measurements
Probably I will need more than one color sensors in order to complete the machine and a Propeller board will be a better choice than a Basic stamp board, as it has more I/O pins.
Have you seen spin code for ColorPAL or for the TCS3200-DB Color Sensor ?
gum_project.jpg
1024 x 680 - 111K

Comments

  • PJAllenPJAllen Banned Posts: 5,065
    edited 2012-12-02 05:37
    If you're looking for grit, then look with/for infrared.
  • NikosGNikosG Posts: 705
    edited 2012-12-02 07:46
    PJ Allen wrote: »
    If you're looking for grit, then look with/for infrared.

    Hi PJ, you mean the IR Transmitter Assembly Kit or something else?
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2012-12-02 08:56
    I meant, basically, that you consider more than visible light.
    The results from infrared reflections/absorption may prove revealing.

    I once worked at a company that made sorting machines (grains, nuts, rice, beans, etc.)
  • ercoerco Posts: 20,255
    edited 2012-12-02 09:27
    PJ Allen wrote: »
    I once worked at a company that made sorting machines (grains, nuts, rice, beans, etc.)

    Ahah, it IS you! I've been tracking you down for 17 years. You still owe me 20 bucks and a bag of rice...
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2012-12-02 09:42
    It looks like your friend has given you the most clear-cut set of samples to separate. It will get harder, believe me. What happens, for example, with a piece of gum that's 90% clear but with a few dirty bits in it? Also, getting something to work in a benign lab environment doesn't always mean it will work well in production. I'm not trying to discourage you -- far from it. But I did this kind of work for years using linescan sensors like the TSL1401, and every job came with a different set of challenges. I think the most valuable lesson I learned from all of it was how to tell which projects to decline. :)

    Now, for the technical part of the question. I agree with PJ that IR illumination can be a valuable tool. You might also try backlighting, since the "good stuff" is somewhat transparent. This will be tricky, though, because refraction can make parts of irregular clear objects look opaque. (BTW, backlighting will not work with the ColorPAL. It will work with the TCS3200 and TSL1401, though.)

    -Phil
  • NikosGNikosG Posts: 705
    edited 2012-12-02 12:20
    Hi guys,

    Nice to hear you again through the forum!!! (I was off-line for a long time)
    Recently I took a small a loan in order to buy a 3d printer. So I must find ways to raise money in order to pay back my printer. So I can’t refuse the offer for the “Gum sorter” construction. I’ll take the risk. In the worse case someone will try to find me in order to take his money back (as said Erco :)). This is very common in Greece these days…. :)

    So, Phil what you suggest between TCS3200 and TSL1401?

    Can I find Spin code for these sensors?
  • ElectricAyeElectricAye Posts: 4,561
    edited 2012-12-02 13:05
    Just thinking out loud here: you might also have a look at the fluorescence of this material under UV. A lot of aromatic compounds will glow under UV, and that might provide you with another bit of data on which to make your decisions.
  • ercoerco Posts: 20,255
    edited 2012-12-02 16:00
    Many particularly useful and insightful suggestions in this short thread already (except for erco's usual rubbish). THIS is what Christmas and the forum are all about.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2012-12-03 12:18
    Nikos,

    Here's how I would go about it if I had to say yes to the project:

    attachment.php?attachmentid=97489&d=1354565908

    The gum pieces would be singulated (forced to go single-file) on a conveyor. (This, BTW, is a project by itself.) One-by-one, they would drop off the end in front of a curved, backlit diffuser. (The reason for the the curvature is that you want to eliminate as much shadowing from refraction as possible, so that any shadows are due to dirt and other inclusions.) As the pieces drop past the TSL1401, it looks for dark spots in the field of view. If it sees any whose size and darkness meet a certain threshold, it will send a timed blast of air to divert the piece from the stream of "good" pieces into a reject bin.
    gum_sorter.gif
    600 x 631 - 7K
  • ElectricAyeElectricAye Posts: 4,561
    edited 2012-12-03 13:29
    Phil's solution looks great. I am particularly impressed by his use of the curved diffuser.

    Any idea what the dirt actually is? Is it soil or dead bugs or what? I wonder if dirty pieces would absorb heat faster than clear pieces. Could you pass them briefly under a mild heat lamp, then read their temperatures and/or thermal emissions as they pass into another section of the process? This would require a sensor different from the "normal" (near) IR you might get from an LED IR emitter.

    I suppose if you were really desperate, you could have the gums immersed in a fluid that matches their index of refraction so the gum itself seems to disappear and you then see only the dirt. But you'd need a fluid that doesn't dissolve the gum and is easy to clean off at the end. Probably not practical, but just thought I'd toss that out there. Mineral oil might come close, but maybe it would make a mess of everything, too.
  • NikosGNikosG Posts: 705
    edited 2012-12-04 05:58
    Phil ,

    For another time I’m curtsy! You are a genius man!
    Look what exactly my potential customer had asked from me some days ago! (I send you his original mail with black and the translation with blue notes).

    attachment.php?attachmentid=97519&d=1354629186

    I think that your idea is exactly what Mr. Pantelis (my customer) wants! I will talk him for you and for your brilliant design!
    Phil, I can build the curved backlit diffuser, using my 3d printer and start the system’s construction, connecting sensors electronics etc. I can also build mechanical parts with aluminum etc. However I would feel more comfortable if I knew that I could have your help. Do you agree to help me and share the profit from this project? Please say “YES” I ask you NO MONEY only your advice and some lines of Spin Code :) .
    In every case I think that I must order a TSL1401 sensor…..

    ElectricAye,
    Regarding the sample and the dirty on in, I’ve just received an envelope including gum sample from Mr. Pantelis. Here is how it looks like…

    attachment.php?attachmentid=97520&d=1354629198


    The dirty is mostly small pieces of tree skin, leaf mold, and phyllome and stuck sand on gum’s body.
    Pantelis_Gum_machine.jpg
    759 x 620 - 99K
    Gum_sample.jpg
    1024 x 777 - 63K
  • ElectricAyeElectricAye Posts: 4,561
    edited 2012-12-04 06:04
    Nikos,

    I'm just curious. Is there a reason your customer does not simply dissolve the gum in a solvent like water, ethanol, etc. and strain out the dirt, then evaporate the solvent and have a purified gum? Is he selling this stuff as unprocessed, all-natural or something? Or does processing it mess it up somehow?
  • NikosGNikosG Posts: 705
    edited 2012-12-04 06:35
    ElectricAye wrote: »
    Nikos,

    I'm just curious. Is there a reason your customer does not simply dissolve the gum in a solvent like water, ethanol, etc. and strain out the dirt, then evaporate the solvent and have a purified gum? Is he selling this stuff as unprocessed, all-natural or something? Or does processing it mess it up somehow?

    I don't know why they doesn't dissolve the gum....
    But I suspect that the natural unprocessed gum is more valuable and has unique properties.
    As they say "Chios Gum" is unique and they prefere to buy it on its initial "natural form" and then they process it.
    For centuries the separetion (clean and dirty) of this unique product use to be a very trying work has been done from old women (island's habitants)
    attachment.php?attachmentid=97521&d=1354636031
    traditional_gum_separetion.jpg
    805 x 1002 - 389K
Sign In or Register to comment.