Short range wireless
Two’s company, three’s a...
Not so obvious choices The obvious short cut – adopting Bluetooth or ZigBee because they are industry standards, without considering other options – is flawed. In some applications, such as the after-market for mobile phone peripherals, the designer has no choice but to use a standard. But Bluetooth and ZigBee are resource- heavy and can be expensive to implement: in many applications, a thorough evaluation of the market will uncover alternatives that could result in better end-product performance, lower cost, or even both.
Does the market really need so many short-range, low data-rate RF
components? Ulf Ziemann examines an increasingly complex market
I
t is hard to think of a more crowded, confusing and contradictory component market segment in electronics than short- range, low data-rate RF. This market has more than once managed to complete the long, involved process of establishing an industry-wide standard (most obviously, Bluetooth and ZigBee). But in practice the area is far from standardised or commoditised, and a forest of proprietary technologies seems to spring up almost as soon as the standardisers think they have cleared the field.
If only short-range RF
communication technology for industrial and embedded developers was as uncomplicated as its wired counterpart! For fixed-link systems, the situation is simple, easy to understand and, in a measured and rational way, responsive to different customer needs. If you need high data rates, look no further than Ethernet, which is widely deployed, easily understood and supported by a vigorously competitive component market, which keeps bill-of-materials costs low.
If you do not need such elevated data rates, but value a high degree of robustness and reliability, the
12 July/August 2010
Foundation fieldbus, LIN, CAN or Flexray bus standards are ideal: choose your data rate and feature set and take your pick. And, as with Ethernet, the disciplined approach to standardisation and interoperability ensures that component prices are competitive.
Even in the field of RF the
industry has managed to successfully implement a uniform, standard approach when it comes to high data-rate applications – it is hard to think of more successful examples of industry standardisation than WiFi and GSM/3GPP.
So on the surface, it appears immensely frustrating that the field of short-range, low data-rate RF – a niche application, when all is said and done – is served by so many competing approaches and technologies, and has achieved so little genuine standardisation. If the whole industry could agree to use Bluetooth for short-range RF, just imagine what this would do for component costs, security of supply and the pool of skilled developers! It is only when you study the details of your application, and attempt to match a short-range RF IC to your application’s requirements, that you begin to understand why such a confusing mass of choices confronts you. Quite simply, there are so many different choices because there are so many different kinds of short-range RF applications.
Components in Electronics
Imagine, for instance, that your product spec calls for a wireless link providing a real maximum data rate of 1Mb/s and a minimum range of 10m. This is right in the sweet spot for Bluetooth and ZigBee. Why would you look any further, given the multi-vendor support for both technologies and their proven performance in hundreds of deployments? This is true – but on investigation you find that Bluetooth’s large protocol stack needs huge RAM resources: at least 64kB of Flash for a full
implementation of Bluetooth. In embedded terms, this is an enormous memory footprint, which is both expensive and wasteful of power.
Happy to avoid the high costs of Bluetooth certification, you turn your attention to ZigBee, which has a reputation for being cheaper to implement and more power-efficient than Bluetooth. This reputation is true – up to a point. ZigBee was indeed designed to be used in low- power mesh networks – the mesh topology providing for a greatly extended range compared to simpler star topologies.
But as with Bluetooth, there is a price to pay, literally and technically. The literal price is the cost of admission into the ZigBee ‘club’: membership of the ZigBee Alliance is a considerable lump sum to pay during the development phase of a project, before you are assured of your product’s success (and therefore of your ability to pay back the cost). And the technical cost is for the hardware resources needed to support the complex software that
implements mesh networking – in full-function versions of ZigBee, as much as 46kB of Flash. There are short-range RF technologies that operate on far smaller hardware resources, so you have to ask whether you really require the benefits – range, interoperability, multi-vendor support – that ZigBee offers. Bear in mind that the long range requires the implementation of mesh networking, and this in turn requires a router that is always on – it cannot be battery-powered. Can your application accept this? Very quickly the choice of a proven, industry standard that looked obvious at first has become no longer so obvious. And other options are readily available.
Alternatives Take MiWi, a proprietary technology from microcontroller house Microchip that supports mesh networking. MiWi is derived from ZigBee, and like ZigBee it supports mesh networking. But because the source code for MiWi is available free from Microchip, there are no up- front costs to bear, and there is no requirement to pay for membership of the ZigBee Alliance.
But does it make sense to adopt a technology that is similar to ZigBee, offers similar functions including mesh networking, with a similar hardware overhead, but that does not benefit from the multi-vendor support and interoperability that ZigBee offers? It’s also worth considering that MiWi does not implement channel-hopping, which is a technique other technologies use to avoid interference in the cluttered spectrum common in industrial and residential applications. If MiWi is a sort of halfway house, other options offer dramatically lower overheads than ZigBee or Bluetooth. CyFi, a proprietary technology from Cypress Semiconductor, is one. Because it implements a simpler star networking topology rather than ZigBee’s mesh network, the protocol layer is far smaller and simpler. This results in a smaller hardware overhead – typically an 8kB Flash requirement at the hub and 5kB at each node. In addition, CyFi
www.cieonline.co.uk
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44