farm robots
Main page for Farm Robots    Robots and the Environment       Robot Features       About the Author      
Invest In Or Donate to Farm Robot project 
This project has been suspended until I can find permanent work and can afford to carry on.
Parts and Expenses
Proposed Development Plan

Robot Related Links
Optical Encoder Pattern Generator
Incremental Optical Encoder info

The Box Robot testing platform project.
Construction started: April 2009
Inspite of not having enough money to build the first prototype farm robot, I took my lay off from my last job [Nov 2008] as an excuse to cash in my RRSPs that I had accrued with the former company, and I was able to purchase enough parts to build a very simple robot. This is not a farm robot, but it gives me the platform [simple robot] in order for me to learn how to use the electronics, and to start the farm robot software.  Some of the parts I purchased, or had already purchased, were 3 PC104+ embedded computers, Wireless LAN kit for PC104s from Sensoray, Frame grabber for PC104s, simple video camera board from ebay, AMP Bridge2000, Electronic Compass, GPS Unit from the RobotShop, PC104 Power Supply unit, and the cheapest simple two wheeled robot body frame [+ 2 balance wheels] that I could find - from Zagros Robotics. I chose to buy a partial frame in order to save time. There can be a lot of design considerations for a robot body, and I was more interested in moving ahead with the project as fast as possible then with determining which motors, gears, brackets, etc to use for a simple two wheeled robot. My only consideration was that it was big enough to carry all the parts I wanted, and thus carry enough weight. I chose the 16 inch by 16 inch Zagros frame with the gears for the slowest speed. I'm not interested in speed. And in fact I wanted it to be slow as a way to protect it from damage during testing. With this simple robot I will be able to develop the skills that could help me find a new job,  but more importantly start some of the key systems for the real farm robot. Another source of some of the parts was WinSystems. I chose some of the parts and companies based on their parts having sample software and/or drivers for the real time operating system QNX [aka Neutrino]. However, to start with I may not be able to use this RTOS.

Stage description for development of ROBOT/ROV [Remotely Operated underwater Vehicle] hybrid. (Note: ROV does not normally refer to out of water robots.)
Development of a Robot/ROV hybrid version of the robot to develop some of the systems to get the robot going and to have the basic concepts "made into a reality." [Most of the 1st year.]
Semi-autonomous version. Introduction of some of the automated logic/programming for the Robot/ROV to actually do some of the work on its own. [End of the 1st year.]
Autonomous version. Indefinite process towards a robot that can work on its own to keep the use of pesticides and herbicides to a minimum. Only intervention required is maintenance and repairs. [Continuous process - forever.]

Link to Box Robot Images and Structural Progress [bottom of this page]

Testing of the various parts on their own.

Month 1

Electrical Systems:  

Mechanical Systems: 

Month 2

Electrical Systems: 

Mechanical Systems: 

Month 3

Electrical Systems: 

Mechanical Systems: 

Month 13+

Continuation  into the future for new designs and features.

Box Robot - testing platform
Under carriage of box robot
The Box-Robot testing platform. This is a slightly modified Zagros Robotics body.
I added the perforated metal panel since it will allow me to add parts just about any
where I want. Plus it will act as a static discharge plate while I'm working on the
electronics. I guess I will have to add a grounding wire to it. The space between the
plate and base also allows me to fit my hand between them, to hold screws and
bolts in place, while I'm mounting parts.
[April 27, 2009]
Here you can see the motors and the optical encoder disk that I am planning to
mount to the system here. At the moment I'm waiting for one electrical component
for the test circuit for testing the optical encoders on one of my breadboards.
The optical encoder disk is actually a CD that I printed a black & white pattern on.
Its a 48 segment disk since this works out to about one pulse per centimetre. Here's
a handy page to generate optical encoder patterns, I did my own though before I
found this page. Busyhtail provides a variety of options for the pattern to be
generated. [Note there was a problem with the disk. See
Farm Robot - Box Robot testing platform
Farm Robot - Box Robot testing platform
There was not enough space between the wheel, the screw that mounts it to the
shaft and the gearbox to put the optical encoder disk between them. So, I had to
mount the disk on the outside. This will actually make it easier to mount the sensors
since I will have more room and can hang the circuit board down from the base.
Note the large black box in the middle of this robot base. It would be nice if Zagros
would offer a longer shaft for mounting the optical encoder disk between the wheel
and the gearbox. Also they provided only a single pair of sensores, but quadrature
encoding requires a pair per wheel. So, it would be better if they sent four of them.
[May 2, 2009]
The black box will hold a 12 volt gel cell rated at 7.2 AH [Ampere Hours]. Zagros
provided two - 4 D cell battery holders to provide 12 volts, but the cost of me
getting 8 rechargeable D cells, plus a new charger is not worth it.My current normal
battery chargers only hold C cells and smaller. Its cheaper to buy a 12 volt gel cell
and mount it to the bottom. Luckily I purchased the 16 inch by 16 inch version
which has enough space for a 12 volt gel cell underneath the body. The 12"x12"
would not have allowed me to fit such a battery under the body. Here I've mounted
a heavy plastic box to hold it. And the box fit between the balance wheels with about
1/4 inch to spare on each side of the box. I would recommend that Zagros add a 12
volt gel cell battery holder to their list of options. As having the battery here not only
makes for more space on the top of the main base, and thus more room for circuitry
[etc] it also lowers the center of mass for stability.
Stepper Motor Test Circuit

I ended up finding a Stepper Motor test circuit that allowed a 1 MOhm linear pot
to be used to control the speed of the stepper motor. I chose a stepper motor to
test with as I had a few lying around. Stripped out of old floppy drives. Finally I
had a use for them. :) This required the use of a ULN2003 chip. I found out more
about this device from Electronics Schematics for Hobbyists web site. But I found
the original test circuit from Basic Stepper Motor Driver circuit before I even knew
about this chip. [May 11, 2009] Click on the following LINK for an image of the
schematic for the stepper motor test circuit. This is a test circuit to control the motor
by using a dial, really a linear potentimeter. The actual control of a stepper motor
by a computer would not use this circuit. Instead you would use the ULN2003 chip
differently through your parallel port or some similar circuit.
I found a circuit to drive the photomicrosensor on one of Omron's pages about
photomicrosensors in general. This is a pdf and it should load into your browser
I was not able to get the SY310 to work properly, plus its leads are too short to
be used well on a circuit test board. I then tested the original QRB1134 [data sheet
courtesy of Seattle Robotics Society], which I don't like because it won't sit flat on
a circuit board due to its design. This also did not work. I then considered perhaps
the problem was the disk and paper I was testing. And sure enough it was, I had
never tested my finger, and yet my finger was working with the infra-red sensor.
Oddly enough it saw the white and black regions of the CD as having the same
infrared reflective properties, and even the clear plastic disk you get when you buy
a pack of CDs or DVDs as being just as reflective as even the white and gray
paper I was testing with. All these materials reflected infrared in the same way.
The solution turned out to be to re-paint the optical encoder pattern with Testor's
Flat Black paint. This worked very well, and both of the photosensors now
work properly. [May 14, 2009]
*Currently studying for a job interview.

Dragonfly Web Design          Contact Us                About the Author

1 1 1