I love the concept of the Farmbot and I am excited to become part of the community. I had a thought recently of other configurations for Farmbot. The major one that came to mind is a circular or rotary Farmbot. I see these all the time in actual farms. I propose that a circular Farmbot would consist of following features:
Able to rotate through an angle up to 360 degrees. I would not recommend totally continuous rotation as that would require rotary unions for water, vacuum, and electrical. 360 degrees of rotation would be easily achieved with a little bit of extra slack in the cables and would not require a cable carrier for the center post.
Leveled circular garden bed. This would likely be the trickiest part of the circular configuration. Curved bed forms could be used to get a nicely curved edge. Or a large section of corrugated pipe say 10 feet in diameter. (Ideas and suggestions would be appreciated on this one)
The theta (rotation) drive would be mounted on the drive wheel and would not require a timing belt, pulley and tensioner.
Magnetic encoders could also be used on the r and z axis which would cut at least $100 off the cost. Also magnetic encoders are better at environmental extremes and are not effected by dust or moisture.
I will work up some CAD and an updated bill of materials to show the proposed difference.
Any thoughts or suggestions are welcomed. I will keep everyone posted if/when I make progress.
This is a great idea, and we are actually going to be sponsoring a student engineering team to work on a circular bot design starting at the end of the month (assuming we get the interest!). Even if that doesn’t happen we plan on pursuing this idea next year, but it would be awesome to see you get a jump start on some designs and testing!
In that light, it might be prudent to use a minimum endstop, and limit travel to 185 deg. The reason for the half-circle plus a bit more, is because you have reflective symmetry across the X axis.
The only problem I see you could have, is that your quality of measurement gets progressively worse the further away from the center of the circle. We have similar problems with polar 3d printers.
Another idea, is to use visual odometry instead of opto-encoders on steppers. You already will have a camera; pipelining it through a travel/distance system would not be hard. You already know mostly where you are going (or should!)
I see your point about reflective symmetry. However the point of the circular design it to cover the largest area at the lowest cost so I would configure the arm to be the radius of the circle not the diameter.
You also are correct that the accuracy falls off as the radius gets larger. The key to this is having greater resolution in the sensor than you can mechanically achieve in the arm. The links I posted would give you 4096 pulses per pole and 160 poles around the ring. Divide that by say the circumference created by a 3 meter arm and you can achieve a resolution of 28 microns. I do not expect that a 3 meter arm will be straight to 28 micron so your resolution is much greater than your mechanical stiffness.
This is all with a sensor that costs $15 and a ring magnet that costs $3.
Just thought of another configuration that would make the circular or pivot farmbot interesting. If you wanted to set it up for 180 degrees and make a pivot along a fence you wouldn’t need another post to support the pivot. The fence post would serve as the pivot.
That’s not, by chance, a team at California State University, Northridge is it? If not, I’d like to talk to you about this. We are starting a project for a polar system right now.
Circular design is a really good idea. If you build an rotary arm of lets say 3 meter with gripper and wheels on the outside and design it as a modular system u can attach another 3 meter arm with gripper and drive-system on the outside to enlarge the operation area by potential times. So a single module can work on 28m², a double module operates on 113m² and a dribble module operates on 254m². The sensoring is also no problem if you resolut with the the sensor of the drivesystem on the outside. Also by building a rotary arm you can provide solar power and watersource from the center.
I want to build my first Farmbot with the circular/polar design as it reduces a lot of cost and complexity. I think it could possibly be reduced more in the future, but I’m too new to this to say for sure. My initial gut estimate said the cost should be reduced to $1500 and they achieved $1800 which was very impressive. We will ultimately have to drive the price point down progressively further in order to increase adoption and I know everyone here understands that.
I have a question for the Farmbot leadership. Are the digital assets from the fantastically detailed CalPoly capstone report able to be open sourced? These would be very helpful in driving this concept forward as a community. While the circular version cannot really compete with the Genesis XL in raw area, I think on a $/m^2 metric it could be quite a bit more efficient. I love both designs and really appreciate this movement.
Btw for those that haven’t seen it, here’s the link to their awesome final report:
I, too, am fascinated by the concept of a center pivot Farmbot. Center-pivot irrigation is a huge industry in my home state, Kansas, as well as many other regions across the Great Plains. Some of them can be hundreds of feet long, with several pairs of wheels supporting the movement. Some of them are driven wheels, while others are idle. It seems to me that installing Farmbot technology on conventional center-pivot hardware could decrease the time to a useful proof-of-concept. Also, developing the prototype at full-scale has many advantages. One advantage, for example, has to do with scaling. When you increase the radius of the Round Farmbot by 10%, you actually get more than 10% increased arable land.
I know of at least one challenge. Center-pivot irrigation generally relies on water pressure to provide the mechanical energy necessary to move the apparatus. In other words, there are usually few motors, if any, but sometimes there’s a pump out at the edge of the field that pressurizes the water supply for several irrigation posts.
Another challenge might be, a full-scale center pivot machine will have a lot of ground to cover, and will therefore need to move much faster than the slow rotational speed of a typical irrigation apparatus. There may be some design and engineering challenges to accommodate the greater travel speed of the critical implements, which the rectangular Farmbot has already encountered.
This is exactly what led me to become part of this community…I love the idea! Rotational Agriculture Robot (ROTAGROBOT). The idea here is that the media bed rotates while the farmbot retains its grid operation. This would enable an unlimited scale for commercial application! Thank you for starting this thread!!