584,833 active members*
5,419 visitors online*
Register for free
Login
IndustryArena Forum > Hobby Projects > RC Robotics and Autonomous Robots > Big Robot for Feeding Lots of Fish
Results 1 to 9 of 9
  1. #1
    Join Date
    Apr 2011
    Posts
    73

    Big Robot for Feeding Lots of Fish

    I am helping a friend with ideas on building a robotic feeder for growing fish. This robot will serve many containers that will be laid out around the four walls of a newly constructed room. So far, I think that the plan is that each shelf will be several containers deep and several containers wide and there will be several shelves on each rack. The racks will be placed side by side mostly all the way around the room.

    I’m starting this thread for him because I built a robot with help from this community already and am comfortable with this space. I was a complete novice when I started but have not grown enough to give him the input that he will need for this project.

    He may chime in with more details, goals, corrections, constraints and preferences.

    “He” (sorry, I may use “I” on occasion) hopes to build a robot that is battery powered so that it can cycle around the room continuously, for a day or two, without crossing up the wires or recharging.

    Attachment 269764 Attachment 269766

    This is a picture of a robot with a large base that can house a large battery and other equipment. It could be bigger if needed. Let’s say that the base would move in the X direction although it would actually go in a square pattern with rounded corners. I am not sure, at this point, how much vertical rise, if any, will be implemented. Let’s say that the Z direction would be for a vertical boom that would also have a horizontal boom which we can call the Y direction. I included a picture of a vertical boom just for reference but may take a vastly different look as we get your input.

    The Y axis is necessary so that a feeding array, that will be attached to it, can move in over the fish bins and then come out again before it moves on to the other shelve without hitting the last one or on to other rack without hitting the shelf standards. As I understand it, this array would have several units like grease guns that would dispense fish food in the form of a paste. They would need to be cooled with liquid tubing or individual Peltier thermoelectric cooling chips so that the food will not spoil during a day or more of continuous operation. This array might have several layers so that more than one shelf could be serviced with each extension and retraction.

    The base will need to follow a track or line of some sort, around the inside of all of these racks. At this point we hope to design and build a self-powered base that will run along the cement floor with regular drive wheels. We also hope that the guide wheels will be controlled, as I said, by a physical track or rail system …or by a sensor based system that is close to being “off the shelf”. Most hobby sensor guided robots turn very small wheels but the weight of the base will be high and the steering will have to have a little more power to turn the wheels.

    We were thinking about using an Arduino as the basic brain but the code could get ridiculous real quickly with so much stuff going on. For this reason, we thought of using a peg system that contacts or other type of sensors could “see”.

    After writing some code, it would say something like: “Feed” which is, extend in Y, dispense the food and retract. Then as the base roles in the X direction, it moves along the track and around the room and the brain could simply respond to pegs or some other marker that says:


    1. You have arrived at a new column/rack of shelves so stop and begin “Feed”ing the bottom set of bins and then move up.
    2. You have arrived at this new set of bins so “Feed” and move up again.
    3. You have arrived at the top set of bins so “Feed” and go back down. Then move to next rack of shelves.


    When the user has to take shelves or rack out of service for cleaning or repair, that person could simply remove or lay down the appropriate pegs and the robot would simply skip and move on.


    There is a lot of detail that needs to be added but I hope that this will paint a good enough picture. We need all of the help that we can get so we would love to get your input on any part of this project.


    Here’s thanking you in advance.

  2. #2
    Join Date
    Feb 2015
    Posts
    3

    Re: Big Robot for Feeding Lots of Fish

    Hey folks, I am a novice to automation and thank you in advance for your help. I have been talking with herring_fish about an automated fish feeder.

    The 20' x 20' square room has horizontal perimeter shelves on three levels: knee height, waist height and neck height. To just have a two axis system, I don't mind the feeder disbursing to all three horizontal planes at the same time. Maybe the entire feeder array slides/throws the same distance along the Y axis. I assume the X axis would be the movement of the base around the room and the Y axis is the horizontal throw as the feed guns move about a meter from their "home/all clear of obstacles and ready to move around the room" operation to their "moved horizontally over the shelves into food disbursing position" operation. After food dispersal, the Y axis retracts horizontally about a meter from feeding position over the shelves back to it's home position/clear of obstacles. The the base would move clockwise around the room about a meter on an X axis, come to a stop, and repeat the Y axis throw again.

    On the other hand, in a square room there are two walls running north/south and two walls running east/west, so that's two axis of movement around the room. I get confused sometimes. I'd like to put the feeder on a track so it moves clockwise around the room forever. I'd like to use a couple 12vdc car batteries for power. Battery charging will be manual, otherwise it probably complicates things too much. Each feeding station has room on the sides for guide tracks on the Y axis if needed.

    Alternatively, the feeder could feed one shelf at a time. It would move a meter on the Y axis to feed the bottom shelf, disburse food, then withdraw. Then lift vertically a couple feet, move into place a meter to disburse to the middle shelf, then withdraw. Then lift vertically a third time a couple feet, move a meter on the Y axis to feed the top shelf, then withdraw. With that section fed, the base would move along the X axis to the next station, the Y axis would either return to a home position at the bottom shelf or stay in it's last position at the top shelf, and repeat the cycle.

    My preference would be to build a three tier feeder, disburse to all 3 shelves simultaneously, and get rid of the vertical movement altogether. I am guessing that this would be a simpler and more robust method. I have plenty of grease guns.

    Here are two other details.

    One: For simplicity, I'm thinking of bolting a horizontal C channel (like a garage door track) on the shelves so a dumb feeder simply moves along the rails to stop at each station, working clockwise around the room forever. Instead of a robot or motorized base unit moving around the room perimeter, a belt/cable/chain moves the feeder clockwise to its next feeding station. At each station I envision proximity sensors or pegs that stop the feeder for a few minutes until the Y axis moves into place, distributes food, and then retracts back to the "all clear of obstacles/home base/ready to move to next station".

    Two: Each shelf will have up to 8 fish tanks. Three shelves of up to 8 tanks means up to 24 tanks grease guns distributing food at each station. I envision the triggers being pulled by stepper motors, probably 24 stepper motors for individual control. However if some of the fish tanks are empty I don't want to disburse food and pollute the water. So I'm thinking that 8 optical or other sensors/indicators on the outside of each shelf are read by the auto feeder and dictate how many stepper motors/grease gun triggers get activated. For example, the bottom shelf may have 5 tanks with fish. The other 3 indicators/sensors are removed or set into a neutral position and when the 8 bottom grease guns move into feeding position, only 5 triggers get pulled. Perhaps on the middle shelf perhaps 7 fish tanks are occupied and 7 indicators/sensors tell the feeder to pull the trigger on 7 grease guns. And on the top shelf perhaps 3 tanks are occupied and only three sensors are active.

    I hope you get the picture. Because the fish tank occupants may change on a daily basis, it would be nice to have color or proximity sensors that are easily changed.

    My wording probably is not as clear as it could be. Please let me know what details can be clarified.

    Thanks
    David

  3. #3
    Join Date
    Dec 2007
    Posts
    2134

    Re: Big Robot for Feeding Lots of Fish

    Seems like it's just stateful logic being applied, so it sounds to me an ideal application for a second hand PLC and some ladder logic.

    cheers, Ian
    It's rumoured that everytime someone buys a TB6560 based board, an engineer cries!

  4. #4
    Join Date
    Feb 2015
    Posts
    3

    Re: Big Robot for Feeding Lots of Fish

    I'm not familiar with ladder logic so I'll look it up. Do you think a PLC does the job more easily than arduino? I'm a programming simpleton, so the easiest way to program is probably the best. The reason I like a peg or proximity switch system is because distances might be an inch off from one station to the next. One station might be 39" from the previous, but 40" to the next. A stepper motor that turns, say, 2000 steps would not be as accurate as a motor that activates until the next two sensors line up.

  5. #5
    Join Date
    Apr 2011
    Posts
    73

    Re: Big Robot for Feeding Lots of Fish

    As I understand it, ladder logic controllers would do a great job as well. It has been the standard for industry many many years. I took the course but didn’t practice it because my job didn’t involve using it at the time.

    On the other hand, I had other language based macro level projects over the years so I kept my hand in the text based programming side of the game. For that reason, it was comparatively easy for me to convert over to the pruned down “C” language that the Arduino uses
    .
    That is why I suggested an Arduino based system. It is a platform that has filled out quite a bit lately. There are tons of hobby level based solutions and Arduino is starting to move up to penetrate the industrial market. At the same time latter logic based systems are starting to come down into the hobby market.

    The bottom line for me is that when you are a hammer, everything looks like a nail so that was my suggestion but latter logic is also a great way to go as well.

    The peg idea will work with either style system.

  6. #6
    Join Date
    Dec 2007
    Posts
    2134

    Re: Big Robot for Feeding Lots of Fish

    Quote Originally Posted by fishmandave123 View Post
    I'm not familiar with ladder logic so I'll look it up. Do you think a PLC does the job more easily than arduino? I'm a programming simpleton, so the easiest way to program is probably the best. The reason I like a peg or proximity switch system is because distances might be an inch off from one station to the next. One station might be 39" from the previous, but 40" to the next. A stepper motor that turns, say, 2000 steps would not be as accurate as a motor that activates until the next two sensors line up.
    I don't know about better, it entirely depends on the application, skills of the engineers, and the degree of difficulty. But this is what PLC's were invented for, and they can be bought extremely cheap now, less than $50 I've seen on ebay.

    They rely simply on inputs and triggers, performing pre-determined actions, move until hit sensor or limit switch, activate relay to drop food, pause for set time, move until hit next sensor or limit switch, rinse and repeat.

    Regardless of the hardware used, in the end it's still just a stateful system being deployed.

    If you can build this logic into an arduino, then it's horses for courses, whatever gets the job done best for the level of your ability is what's required. Would be cool to implement a PLC in arduino though, and I suspect it's already been done a number of times to make the job easier.


    cheers, Ian
    It's rumoured that everytime someone buys a TB6560 based board, an engineer cries!

  7. #7
    Join Date
    Jul 2007
    Posts
    1602

    Re: Big Robot for Feeding Lots of Fish

    This sounds really overcomplicated. Why not run lines to each tank from a central point say a modified refrigerator, pump the glop into as many lines as needed for the day (peristaltic pumps meter well and are easy to make) and use a water flush to deliver it and clear the lines.

    bob

  8. #8
    Join Date
    Feb 2015
    Posts
    3

    Re: Big Robot for Feeding Lots of Fish

    Bob, you have a very good point. The challenge is food viscosity. Fish food should be a pasta-like consistency, pumped food should be more watery. Watery food would overwhelm the biological filtration system. Ideal food is a blob of paste with enough gelatin which, if uneaten, holds together long enough to get caught on a filter screen and removed. Think of the home made pasta that comes out of the extruder, or think of a thin bead of toothpaste. This is a food consistency that sticks to itself and hold together.

    Liquid food won't do that. But your point is well taken. If a peri pump can push thick paste through 15'-20' of line, that would be an easier way to go.

    The other item is cleaning. Fish rooms are very humid and warm. Bacteria will build up quickly in wet food. The grease gun works on a one liter plastic bag which can be exchanged every couple of days while the feeding machine gets cleaned. I'd prefer to clean water lines by clearing with water and then chlorox, but the feeding tubes would have to be diverted to a drain system so none of that ended up in the fish tank system. That can be done, water line cleaning is not hard.

    Pumping thick paste through a long tube is the challenge. But it is a point worth thinking hard about. How would you make powerful peristaltic pumps?

  9. #9
    Join Date
    Jul 2007
    Posts
    1602

    Re: Big Robot for Feeding Lots of Fish

    I didn't realize that the food was that viscous. I was thinking more a thick custard consistency. A design like the one in the animation here: Peristaltic pump - Wikipedia, the free encyclopedia might work well. However your grease gun type device can easily step in.

    How hard would it be to push the food through the tube? If you put a tight slug in a tube, can you move it throught the tube with reasonable air or water pressure? If you could push the slug through with water, the tubing would largely be self cleaning. It would be a matter of determining what length slug can get through without having to apply so much force that the food gets eroded away. Maybe the delivery tube could be a slightly bigger diameter than the slug and then only needs a bit of help to ensure it doesn't plug up.

    I googled peristaltic pumps for fish feeding and got quite a few hits though most seem to be using it for less viscous feed.

    bob

Similar Threads

  1. 3D Fish
    By Wildsign in forum MetalWork Discussion
    Replies: 1
    Last Post: 05-29-2013, 07:39 PM
  2. Replies: 5
    Last Post: 07-03-2012, 08:21 AM
  3. Fish cut in a quarter.
    By EDM_Fixer in forum Machine Created Art
    Replies: 2
    Last Post: 11-06-2010, 02:36 AM
  4. fish plate
    By woodman08 in forum Gorilla CNC Machines
    Replies: 1
    Last Post: 08-22-2009, 06:38 PM
  5. Looking for Robot Auto Palletising (Panarobo and or STORK Robot)
    By rsarwanto in forum European Club House
    Replies: 0
    Last Post: 09-12-2008, 12:22 PM

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •