Wednesday, 4 November 2009


UPDATE: upon reflection, I could actually just hack an Classic 4 prototype from work to do all of the storage (it has some flash memory) and message writing (using keypads) and sending (would be done over serial link to the arduino / radio transceiver rig)! Note: there is also a wifi link for internal communications.
Another idea about shorter range personal messaging over the 805.14 link.

This is a slightly edited re-post of a comment I left on Kal's Blog:

What do we understand by the term: 'off grid'? I reckon it means:
-no power grid
-no mobile phone service/signal
-no telephone service
-no internet

I think we should have a backbone of communication that does not rely on Internet i.e. a radio based network so we could, say, send data over radio links. I suppose this would be very useful if the internet or mobile phone networks went down or just generally for fun!


It would send things like SMS messages over standard radio links using minimal power. So the first thing to say is that means we probably wont use a desktop/laptop because they are power guzzling beasts.

I reckon the minimum hardware needed to do the job is as follows.

Core Specification:
1 x arduino
1 x storage card e.g. SD
1 x radio transceiver
1 x hacked ps2 keyboard
1 x small LCD screen
1 x solar panel
1 x battery bank

this would give the basic hardware platform to send and receive and store messages completely 'off grid'.

optional specification:
1 x ethernet shield for arduino

this could allow a full webserver to be run on the system. Now that would be interesting!

There are several conflicting goals when sending data over a radio link like this: reliability, processing, modulation techniques, security and efficiency etc.

The ideal outcome would be twofish / blowfish / some equally insane encryption on a 100% reliable link which consumes the least amount of power possible. Well this is just not going to happen- firstly, it is virtually impossible and would take an absolute genius if not a team of geniuses to pull off. So I think we should have a more modest initial goal:

Get the link working (unsecured, not optimised for low power)

Later development would consider the following issues:

Issue 1:
The system is powered on all the time?

Solution 1:
Scheduled communication.

Issue 2:
Time synchronisation between transceivers.

Solution 2:
Send SYNC packets between the transceivers.
Most watch crystals drift at 20us/second so we could use that as a starting point to work out the optimal sync period.