[python-uk] Tell us what you did with Python this year....
python at rotwang.co.uk
python at rotwang.co.uk
Mon Dec 20 20:32:09 CET 2010
> Also I've been playing with Arduino, although that's so
> straightforward to use Python does everything I need it to. I've
> fiddled with ZigBee networking, which is quite good fun and I've got
> some basic stuff for that in Python.
I went to a talk by you (on TurboGears?) in 2007 just before I moved out
of York. I hope things are going well. I was working on a mad project at
the time to analyse rail journeys to determine which single staged fares
could save money. The UK rail system has the most bizarre fares
structure. You can exploit it by breaking the route into sections and
buying tickets for each step. On some routes you can regularly save 40%
or more on the fare. We pulled the plug as we realised that we would run
into legal problems. I should have open sourced it at the time. We used
Beautiful Soup to gather the data and used a distributed network using
Twisted (we never used the word botnet) to disguise the source of the
requests. The code worked fine but it was not a good project to pursue.
I did some Zigbee work in 2006 for a company in Cambridge, Alert Me.
They were using it for home security systems, but seem to have shifted
more to green energy monitoring. They used Python a lot too, on embedded
Linux. The Zigbee specs were all in flux back then.
The Arduino is a lovely little board. It is what I was using for motor
control for my solar concentrator. I used Python and PyEphem to track
the sun, or rather reflect the sun onto a target. I couldn't find a
partner or funding so I had to get a proper job. The electronics is all
very simple, but I'm not a Mech Eng.
The idea is simple and not new. An array of mirrors to focus sunlight
onto a single spot. Static panels only get a cosine of the available
sunlight, to they lose out in morning and evening. If you track the sun
with a mirror you effectively halve the angle to get useful energy all
day. Mirrors are much cheaper than solar panels. Most work has been done
on big installations with expensive mirrors, massive motors, big
concrete mounts. My approach was to use small cheap lightweight mirrors,
more of them, each controlled by a micro, fixed to existing structures
(eg. a wall). The prototype used the Arduino. As long as you know lon /
lat and your current time you can calculate the Sun's position in the
local sky. The difficult part is calibrating the mirrors. I was working
on that when I ran out of money and had to get a "proper job".
A single low powered processor board (something like the beagleboard)
running linux talking to multiple cheap microcontrollers (eg. Arduino)
would be able to control a sizeable array with very low power consumption.
I hoped to target small installations in places with no fossil fuels,
plenty of clear sky days and poor power infrastructure. Alternatively
people with a property portfolio, eg. housing associations or developers.
This guy is building really big installations and has very good funding :
I wanted to do the same thing on a small scale. It could make a big
difference to a lot of communities.
More information about the python-uk