2005 Solar Sheep
Solar Sheep
Solar Sheep
Aim of Project: To develop a solar powered lawn mower robot.
Team Members: Alain Yee & Joe Rigg.
Update 17/5/06 !!! Final Design!!!
Position Sensing (Electronics) - Alain
__________________________________________________
Weekly Goals:
Week 1 - Build a moving chasis.
Week 2 (Alain) - Learning to programme the microchip using Wiz-C.
Week 2 (Joe) - Plant grass. Research cutting mechanisms.
Week 3 (Alain)- Build powerboards & programme microchip.
Week 3 (Joe) - Develop testing rig for investigating cutting mechanism efficiency.
Week 4 - Review and iterate lawn mower designs.
Week 5, 6 ( Joe) - Develop means for testing for grass
Week 7 (Joe) - Test for grass
Week 6, 7 (Alain) - Solar panel + power storage
Week 8 (Joe) - Research cutting mechanisms
Week 8, 9, 10, 11 (Alain) - Optical GPS
Week 9, 10, 11 (Joe) - Try out different mechanisms and practicalities
Week 12 (Joe) - Power considerations for cutting mechanisms
Week 13 (Joe) - Attach cutting mechanism to chasis
Week 14, 15 ( Joe) - Refinement Week 12, 13, 14, 15 (Alain) - Control system to optimise cutting strategy
Week 16 - Write up Summary
Week 1: - Managed to put together motor and gearbox in an effort to get a better understanding of the project ahead. - Worked on chasis design. First prototype, aluminium chasis. Final model, molded plastic.
Week 2 (Joe):- The grass has been planted, and already signs of life! (Photos to come...) The cutting mechanism research threw up some surprises; namely that the use of lasers to cut grass has already been achieved by German company Wolf Garten. Sadly, with a power requirement of 6-10 kilowatts it will not be featured in this project. It seems likely that a more conventional approach will be adopted. Whilst a rotary mechanism is probably the simplest, various blade configurations are available, making analysis more complex. Variables include blade material, number of blades, blade shape, as well as how the blade is attached to the hub. The plan for next week is to design a testing rig that will allow us to investigate these parameters.
Week 2 (Alain):- This week was spent learning to programme the microchip. Unfortunately half the week was wasted learning to code in assembly language. Latter half of the week was spent going through the \'correct\' literature; Wiz-C tutorials and guides which turned out to be quite fruitful. Succesfully programmed the microchip to act as a key logger and an LCD clock. Hopefully by the end of the next week, the solar sheep will be able to move according to the sequence programmed on the microchip.
Update:- Completed building of a DC-DC bridge to drive the motor forwards and backwards. Also completed the circuit board to power the microchip which will be brains behind the sheep.
Week 3:- A large part of the week was spent building a rig for "high speed photography" of a blade of grass being cut. This experiment is done to help determine how grass is cut under different conditions and blades. As we did not have any high speed cameras, we resorted to using a strobe to flash at the instance when a blade of grass is cut while having a normal digital camera\'s shutter opened. This was achieved using an LED and (infrared?)detector which was positioned close to the grass being cut. So as the blade passes the diode and receiver, it cuts the light beam (and grass) which triggers the strobe to flash. Performance of the set up was satisfactory until the diodes burnt out as we were putting too much current through it.
For next week, it was suggested that we replace the LED(infrared?)(the detector works) and replace the strobe with an ultra bright LED which has a shorter response time hence a higher flashing frequency.
Week 3 (Alain):- The DC-DC bridge which was built last week could not supply the motors at the desired output voltage with sufficient current. Extra transistors were added to allow a larger output current without frying the transistors.
Programming of the PIC (microchip) took place this week. It involved the flashing of a strobe when a light beam is broken.
Weeks 3 & 4 (Joe):- As Alain mentioned, the task for these weeks was to develop a test rig which allowed us to take photographs of blades of grass as they get cut. Specifically, my task was to design and build the rig, starting with only the cutting motor. The biggest areas of consideration were making the test rig flexible enough to accomodate different blade configurations, grass length and motor speeds, as well as fulfilling safety criteria alonger the way. Solutions to all of these problems are illustrated below. Particularly of interest is the blade mount "boss". This was designed to accomodate up to three blades in two configurations. One in which the blades a fixed relative to the boss, and the other in which the blades are free to pivot at the boss resulting in a flailing action. Once the rig was built, the next phase was to incorporate Alain\'s light gate trigger circuit. Again, flexibility was the key with the gate width critical in order to get a strong trigger. Once this was all completed we were in a position to attempt taking photos. The trial photos are shown below, and, apart from being slightly out of focus, are very pleasing. They confirm what we suspected about the way grass is cut. If we treat a blade of grass as a simple cantilever then in the fundemental mode of vibration the cutting blade will simply knock the grass over and not cut it. We need to cause the grass to behave according to a higher mode of vibration. In the first photo below the grass was successfully cut. If you look closely, at the moment captured below the blade of grass appears to be in the fourth mode of vibration. This raises the question, can we save energy by cutting the grass in the third, or even second, mode? This will be the line of investigation for the coming weeks.
Week 4 (Alain):- Sorted out all niggling problems with the DC bridge which drives the motor. PIC was programmed to perform to drive the lawn mower forward and backwards for 3 seconds each via pulse width modulation. It was found that if a high voltage (7V and above) was used, no current would flow through the motor. I suspect that the outputs of the chip was creating a mini short circuit. At lower voltages (6V~ish), the circuit worked as fine except for the transistors directly connected to the power supply getting a little bit warm.
Week 5 (Alain):- Orders were put out for motors & gearboxes which hopefully will be used in the final design. Feasibility tests were carried out on perspex rods of 12mm and 10mm diameter to be used as lenses for the optical GPS system. They were tested to see how a beam of light would emerge through the rod and the number of sensors which will be needed to cover a 90 degrees lateral shift of the light source.
Week 6 (Alain): Came up with an alternative to using a PLL chip. Each detector is connected to an op-amp (amplifier stage). This is then connected to an 8 channel multiplexer which will be used as a switch to select the output between detectors. This then leads on to a switch which multiplies the output signal from the multiplexer by a square wave. It will have 2 outputs, 1 with the signal multiplied by a square wave and the other, multiplied by a square wave 90 degrees out of face. The 2 outputs are then low passed filtered and fed into the analogue input of the microcontroller where its magnitude is measured and used to determine position.
Week 7 (Alain): Constructed and tested the amplifier circuit. It consists of an op amp which is used as a current to voltage amplifier. One resistor of 100kOhms and one capacitor of 2.2pF was used. Constructed and (simply) tested the switching circuit without any faults. Putting the amplifier and switching circuits together resulted in disaster. The expected result was not achieved. Worst still, the amplifier circuit would not work on its own anymore. Moved on to programming the microprocessor to produce square waves as PWM for the motors and for coherent detection in sensing its position. Concluded that a lower frequency oscillator crystal should be used as the lowest frequency square wave signal that could be generated was 1.3kHz. A lower frequency square wave means lower frequencies can be used for beacons in position sensing, hence less noise and more gain from the op amps.