Initially i’d hoped to dispense standard, roasted monkey nuts as the squirrels love them! I’ve built many variations of the dispensing mechanism out of foam card however it was proving to be really tricky with the irregularly shaped nuts and their friction.
I’ve tried all sorts of mechanism from the simple opening and shutting of a spring loaded door to the moving of gravity fed nuts using a cut out like the board game downfall.
The downfall was my favorite but very problematic with the monkey nuts so I changed to using shelled nuts which were much easier to dispense.
Above all the design needs to be reliable so I have constructed many prototypes with the most reliable by far being a simple trap door – I don’t mind it dispensing too many nuts at once however I don’t want it to clog up and this design proved to work well.
The whole thing needed to be motorised and the foam card just wasn’t strong enough so I reverted to creating the mechanism from Lego at this stage. This was an easy decision as I have tons of it ahnging about with the only downfall ( 😉 ) being that I didn’t want to conform to the lego standard dimensions however for a prototype it would be fine.
The hardware for the project can be broken down in to 2 main sections – the electronic control of the various sensors and the mechanism to dispense the nuts.
A) Electronic control.
Having played around a bit with breadboard I’ve worked out what I’d like to have for the various inputs and outputs:
1. Motor to drive the nut delivery system behind a reed switch or relay.
2. Microswitch to tell the pi where the motor is – to stop it chucking out tons of nuts or none at all!
3. Motion detector – I’ve bough one from Tandy which works pretty well .Not sure of the use for it yet but I:m going to add it in should it be necessary in the future.
4. Switch to trigger if the nuts are deployed. (possibly an IR breaker beam rather than a switch to avoid complex mechanical issues).
5. LED. A light that I can configure to show different colours. I’ve chosen a RGB LED so that it can be positioned in one place with the colours showing the status – I’d considered several different coloured ones but was concerned that the wildlife would notice the position of the light (i/e 3rd from left) rather than the colour = possibly a new experiment, though.
I wanted something that was easy to connect to the pi and so have come up with a veroboard circuit that connects to the pi via a ribbon cable.
This has the various connectors in place to make it fairly easy to connect up the swiitches etc and the ribbon cable to ensure that they are alwas connected to the correct pins on the pi.
The second piece of veroboard has the connection for the lego motor – these have special connectors and I decided to use an old battery box as a doner for the plug. I then soldered this on to some board and glued it down having worked out the pin arrangement from here: http://www.philohome.com/pf/pf.htm
It all works pretty well however the wires are inclined to come off the pins at the slightest vibration! Also I’ve had trouble passing 9v through the opto isolator as it sticks at anything over 3v – I’ve since got some relays that may prove better.
Having made the decisions about the design of the feeder I needed to begin getting the basics of the pi set up together.
AS I’m going to be using RISC OS there are a couple of hurdles to overcome that aren’t present in linux.
The shopping list for the pi setup is:
1 – Remote access for running the program (or monitoring if autonomous).
2 – Wireless networking.
3 – Sending or recording video.
4 – Get it working from battery power.
Before I get to running wirelessly I thought I’d begin with accessing the pi from a remote computer using its current remote connection – For this I used the VNC server by Adrian Lees – http://adrianl.drobe.co.uk/
He describes this as a shaky alpha release but, while a bit slow, it is perfect for running my basic program.
I accessed it from the laptop using tightVNC which provided a good & stable connection. The pi is set up to display on my tv using hdmi and I didn’t want to fiddle with the screen resolution, so I wasn’t surprised when it didn’t like running at 256 colours.