Two very successful days in a techie kind of way… and sorry, this one is just for techies…
One of my big bug-bears with the Ethernet/radio interface for the home controller has always been the cheap short range radios as you know if you’ve followed my blogs. The NRF24L01 boards are just crap – ok, they work but only short range – and the RF24NETWORK network is great for what it is – but it’s not a true net and there is no chance of making into a true network – you have to have a master and spider-type network with hopping nodes. Their numbers have to be predefined as the hopping sequence is limited. In addition the radios can’t send and receive at the same time so the master really has to keep control at all times of the communication sequence… AND THAT’s FINE – I have it off to an art – but if only there was something better.
Well there is – the Atmel chip range include a pair with radio built in and they’ve developed a lightweight network – but a true network where any node can step in as a relay. Problems? Getting the things, having the software, defining the interface. Well getting them was the first thing we solved, thanks to my pal buying an oven to solder the chips to the boards (impossible with a soldering iron as the pads are underneath the chip), the software we now have working, due to a sensible approach… and the interface is now cracked.
You can get samples out of Atmel and that’s just what my pal Aidan did – well both of us, allowing him to make a series of test boards up. Until this weekend that’s about as far as it went – but yesterday we managed to get the network demos up and running and there is no doubt about it, the extendable network WORKS – we made the mistake of starting with on-board chip aerials which are pretty useless but we know from the demo boards that Atmel produce that they CAN have a decent range with the right stub aerial.
We have SCOURED the web and it is full of failed projects and promises due to people going down the wrong path either hardware or software wise in producing low cost radio networks. One set of guys in Canada thought they had the solution, a little radio board with USB on it – they needed 20k funding – and only got 6k – the project stopped this time last year and even their website has disappeared. WHY? Heaven knows but I suspect the USB interface – why would you want to fasten these to your PC when you already have WIFI…. surely a simple low cost interface for hardware projects using PIC, ATMEL or other chips is the answer, something compatible with he many cheap radio boards out there but considerably better? Well we solved that one today – the SPI interface – after a several blind alleys – we got it working. Making a stand-alone radio board avoids the hurdles of making the boards work with a particular chip. The final hurdle – antenna – the chip antenna designs are useless – few feet range – so we’re going to make 2 boards, one with on-board antenna and the other with a short stub aerial – either way we expect up to 30metres range. With 2 boards you have 60 metre range etc. etc… a true network.
We tested the network by having two boards talk to each other and moving them further and further apart until they were out of range. Merely introducing a third, arbitrarily numbered board in between the two, re-instated the link – we went up to 4 boards two of which simply sat there on the floor of a large house, battery powered and we managed to triple the working distance of our boards.
Nothing more will happen before June as we need some boards with proper aerials to test and I have holidays but we’ve cracked the back of this one. An excellent way to spend the weekend.