Requirements

Sensors

The design will have a few types of sensor:

• Bump Sensors
• RF sensors to detect a perimeter wire.
• Tilt/lift shutoff sensors.
• Some kind of sensor to go home to charge.

Power

The robot will run on a 12V supply. This may be NiMH cells of lead acid. The mower will be large enough to accommodate either to cater for various motor swap outs that may occur.

Charging

The robot will be large enough to support a largish solar panel which would make the charging of lead-acid batteries feasible. It will also be stable enough to make a self docking charging station. At this early stage, I will keep my options open.

Cutting Width

The cut width must be at least 30 cm to mow the lawn in a reasonable time.

Cutting Blades

Lawna started off as a project with the kids. As such, I was ever mindful of having to make it safe for little feet and fingers. Now they are older, I feel I am able to start looking at sharp cutting blades that utilize the width of the mower. So this mower will have sharp blades.

Cutting Motor

I found with Lawna that having a very fast rotating blade gave good cutting results. This came at a price of increased power drain. Using a slower motor that had been geared down gave good torque with lower power consumption, but poorer cutting performance. By using sharper blades, the cutting performance can be improved. However, the key difference with using a lower speed motor is noise. The high pitch whine of a fast motor I find particularly annoying, and if this thing is going to be wandering around the garden for hours, I want it to be unobtrusive.

Motors used: MFA 950D 4.5-15V with 6:1 reduction gearing

I want the mower to cut a 30 cm strip of grass. This can be achieved with one large blade or a few smaller blades. I am inclined to go for several smaller blades rather than one large one although I will need to perform some experiments first.

Locomotion

Lawna used two drive motors at the front and one castor at the rear. This enabled the mower to aggressively move into bushes and over paths before the eye-stalk feelers cause the mower to stop and then back out. The motors used worked admirably and I will use the same motors unless I find them struggling to move the extra weight. However, this mower is going to have the propulsion at the back (with a lot of the weight). The front will be lighter with two castors. The aim is to have an RF sensor that will detect the perimeter wire and back away without the need to punch into a bush. Yes, this means having a perimeter wire around the garden, but this simplifies things considerably. Once the mower is working and cutting the grass, I'll start working on V3 which will utilize some more intelligent software and sensors.

Motors used: MFA 950D, 4.5-15V with 148:1 reduction gearing

Lawna would sometimes (especially in wet conditions), wheel-spin when pushing through fallen leaves. Having extra weight over the drive wheels should give better traction so the battery should be located over the driving wheels, and this is where a lead-acid battery's weight may prove to be an advantage.

A note on software. When lawna hit an obstacle, it stopped immediately and/or changed into reverse. This gives a pretty standard robotic 'jerking' motion. In the new mower, I will make better use of the PWM on the microcontroller to slow-down and stop over several milliseconds, and similarly for the start. I believe this will be kinder on the motor gearboxes and prolong their life. It should also make the turning and reversing motion more gentle.

June 2014