August 2013
The system has been relocated. The original Sandhills system was first installed a few years ago on a hacked Seagate Dockstar. It was originally just an experiment in running a computer off-grid and 3G connected. Since then it has evolved into a Raspberry Pi powered system, and includes various battery and environmental sensors. On the 17th August, it moved to a completely new location. It now sits attached to the new shed up the road.

It’s not finished yet. On the immediate to-do list is to finally implement the 802.11s mesh. The Sandhills system will act as the Mesh-Portal-Point (802.11s speak for the Internet gateway), as well as continuing to do all the chores it does now.

May 2013
Things have been stable with the remote monitoring system at Sandhills lately. This has led me to pursue the next challenge on the list - providing wireless coverage across the whole property.
I'm going to be trialling some simple 802.11s nodes. Have a look at the new link (802.11s) above for more details.

January 2013

It's been a bumpy ride with the Raspberry Pi. Until recently I've had trouble getting up-times anywhere near as long as the old Dockstar could achieve. I also managed to corrupt a couple of SD cards (SD cards are the "disk drives" for the Pi). I think these problems are behind me now. With every new kernel and operating system update (dozens since I started using the Pi), it gets a little bit more stable.
There's nothing like the Pi on the market for anywhere near the price. It's almost the perfect computer (missing ADC only IMHO!) for the tasks I need it for. The Raspberry Pi foundation have done a great job. I know it was never their target audience, but I hear most of them have been sold to grown-up electronics and computer enthusiasts. No wonder! Can't wait for the model A to be released, with its even better power efficiency.
Check them out at raspberrypi.org.

You'll notice a new WEBCAM link. It's using one of the Dockstars running "motion" software. This system will be up and down as I play with it.


It was around this time last year that I heard about a small, cheap ARM based computer being brought on to the market. This board was designed for low-power operation, and have all sorts of digital I/O on it. Well, I ordered one in early this year (the day they were released), and it finally arrived around mid-year and I slowly worked on making it the new computer for the block. As a result, the sensor data is different. Gone are the ADC and 1-wire capabilities of a PICAXE microcontroller, and in their place is a SPI-bus connected MCP3201 ADC chip, and a direct conenction Dallas 18S20 temp sensor. I now get 12-bit voltage measurements, and temperature to three decimal places!
It works well. I now have three Raspberry Pi's (that's what these new $38 computers are called), so lots of new projects are on the cards.

Rainfall figures are now plotted on a graph. I've been recording them for ages (well, September), but only just got around to producing a plot.

The rain sensor is operational. Currently only its raw data is reported. It's the last value in the raw data.
For every count up from the previous value it's an extra 0.508mm of rain that has fallen. For instance, if you see that the last number in the raw data has gone from 22 to 23, then you know that 0.508mm of rain has fallen at Sandhills. Since 0.508 is near enough to half a millimeter, just think of it as a count of two equals one millimeter of rainfall. The count keeps going up, so you need to look at the delta between the current value and the previous value to work out how much rain has fallen.

I'm currently testing remote sensors. I have a remote temperature sensor running in the shed at the moment. It's the third value reported in the raw text output.
I recently purchased a rain sensor from Jaycar that is meant to be used in conjunction with one of their weather stations. I took a gamble and bought it anyway. Well, it looks like it will be very easy to turn in to a remote rain sensor. It's a "tipping bucket" design, and uses a magnetic reed switch for detecting when the bucket has been tipped. I'll hopefully have rainfall measurements running soon!

JULY 2011

The microcontroller part of the system has just been totally re-done. The "prototype" version is gone, and the new one based on a PICAXE18M2 and a PICAXE08M has gone in. Yes, it still uses Veroboard, but it is much neater than the old one.
Now it can make use of remote sensors.
Major web-page redesign.

MARCH 2011
Sandhills has just had about 50 Amp/hours of extra battery capacity added. You'll see evidence of this by looking at the battery voltage graphs. The voltage is higher during the nights.
I have just constructed a new solar charge controller (it is the one published in the Feb 2011 edition of Silicon Chip magazine). It does MPPT so I'm hoping for some excellent results and some lead-acid batteries that are feeling well looked after!

The 3G connection is much more reliable now. The new 3G Yagi is great, and was well worth the construction effort.
Have a look at the Yagi here.
The 802.11 side still locks up after a few weeks, but I can live with that for now.
If you notice that the raw temperature text shows a number above 128, that is because the microcontroller has set the sign bit. Any number above 128 is a value below zero. So, 129 is -1. A value of 130 is -2, etc. The microcontroller uses 8-bits per temperature reading. 7 bits for the absolute value, and the 8th bit to indicate the sign. Wikipedia has a description.