A new motordriver arrived the other day. This is a 2-channel, dual H-bridge, driver able to produce up to 30A.

Martin has done some brilliant work on the BWF receiver. The signal is now clear enough to be recognized at a disctance of 20 cm. This should be enough. There is still potential for improvements though. The signal is interfered by what we believe is the PWM signal from the motor drivers. Some work has to be done on the motor drivers to remove that signal. Also the receiver is sensitive to disturbance from the Arduino card. We are thinking that a metal shiield around the coils should fix that issue.

Session 5 was held this evening. Great success! For quite some time we have been working on getting a new software architecture in place for the android app, trying to port the already working version . The new version is event driven, using state machine logic to determine how to react to sensor input. The code can be found in github. Here is a short filmclip of the new software running on the prototype. Yes, it looks pretty much the same as previous clip since no new functionality has been added.

Some hard work has been put into building a simulator of the mower. Using this simulator the whole team is now able to develop and test PurpleMow software in the comfort of their own home, in front of Let's Dance on Friday nights. Good work Henrik!

Managed yesterday to get range sensor data combined with logic. The machine is now randomly, and with some help, moving around our office space avoiding things or humans that happen to get in its' way.

The Raspberry Pi sold out in a couple of minutes last week so PurpleMow has to wait for the second batch. It now turns out that the first batch was badly manufactured.

Got some useful advice from mr Google where we found others experimenting with coils. The advice was to use something called "Effektdrossel". This was purchased at Elfa and mounted onto the BWF receiver. That did the trick! Now the signal is detected at a range of as far as 20 cm. All remaining to do now is to mount the new coils in the mower.

Third session. Goal for this session was to get sensor data and motor control working together. It ended up in us re-writing the complete Android app. A new architecture was drawn up. The main control will be done using a State Machine mechanism. The motor control and sensor data run in their own threads producing or consuming events from the State Machine.

The BWF sensor is now actually working but it detects the "fence" at a too close distance. A qualified guess is that the coils used as antennas are not optimal for the purpose. Need to do some googling and experimenting.

Another weekend has past, another weekend of soldering.This time Martin has worked on the receiver part of the Burried Wire Fence (BWF) sensor. As you may notice the coils have not been applied yet. These will somehow be mounted in the front corners of the mower and connected via wires to the receiver circuit.

Martin has put in some hard hours the past weekend and completed the sender part of the Burried Wire Fence (BWF) sensor.