Cost is the driving factor for much of what we do. Wally Schirra, the fifth American launched into space, was once asked what he thought about while sitting atop a rocket on the launchpad. Capt. Schirra replied, "This was all put together by the lowest bidder."
The cost of raw materials often is the key constituent to overall costs, and the basic fact is that raw resources are limited. Take copper used in cabling, for example (Fig. 1). While subject to the ups and downs of any market driven commodity, the price variation in just the last 5 years is evidence that we cannot afford to do things the way we have been doing them in the past. Things have to change.
Figure 1: Increases in raw material cost (example above: copper) are driving technology developers to offer lower cost systems that also provide additional features and benefits.
Consider the lighting requirements of a typical office that has overhead lighting and banks of switches located at an entrance/exit point. For each office to have has its own lighting and switches requires extensive wiring and partitioning electric usage across banks through circuit breakers. Distributed breaker boxes fed by high current feeds in a hierarchal configuration can simplify some of this otherwise redundant cabling, but no matter how you slice it, a lot of cabling is going to your switches and to your lights.
What’s more, you still need labor to run the wiring to the fixtures and install the switch boxes. Dimming adds complexity with this approach, as well. Dimming is a feature that may be required, especially in conference rooms and presentation rooms. What’s more, local codes may require thicker cables even if low voltage and low current control signals are used.
The bottom line is that in order to create a cost effective, controllable and automated lighting system, wireless communications will be the key building block going forward. Wireless technology permits flexible and energy efficient implementations, provides useful features and has robust status reporting, which is needed for any real energy management system.
Lots of choices
A few advances have enabled us to get to an inflection point in the use of wireless topologies for lighting. First, the somewhat standardized, international, unlicensed ISM bands bring a lot to the party. This article will focus on ISM bands based 2.4 GHz and 5 GHz RF, which allow components and wavelengths to be small enough to make very compact devices. Also important is the spread spectrum frequency hopping radio technology that allows several similar networks to (somewhat) coexist and operate independently, even though they may use different modulation and hopping schemes. For example, Bluetooth uses frequency hopping while ZigBee uses direct sequence. These should be able to function together in a shared environment, but there may be times that the protocol will need to detect and account for potential collisions.
In the 2.4 GHz and 5 GHz bands alone protocols competing for an overlapping market segment include Wi-Fi (802.11), Personal Area Networks (802.15), Bluetooth (802.15.1), and ZigBee (802.15.4) to name a few that have IEEE classifications. There is also ANT+, IOHomecontrol, 6LoWPAN, RF4CE, Wireless HART, WiMAX, WISA, Wireless USB, ISA100, and WIMI, to name a few. These are either open-sourced or proprietary architectures trying to get a foothold into this very lucrative market.
It is important to know that some protocols are not designed for automation and lighting control, so the field can be narrowed somewhat. Take Bluetooth, for example. As a low power streaming protocol, it is tethered to a host and has a relatively short range, so it is not a suitable candidate for lighting control – but Bluetooth Low-Energy is a different story. (See the TechZone article “Low-Energy Bluetooth Gets Personal”).
Narrowing the field
When it comes to determining your design approach, a few topologies can be distilled from the process. Since narrow-band, continuous-spectra techniques are not part of the ISM band usage, bus topologies such as Time Division Multiplexed shared medium are not viable. Ring topologies are not viable either since with wireless close neighbors do not necessarily bind you.
Overall, a master/slave client server network, or a peer-to-peer, point-to-point ad-hoc network will ultimately be decided upon. Both have advantages and drawbacks, and both may also call for the creation of specialty nodes as either access points or control nodes.
The ZigBee architecture, for instance, uses functional blocks called Coordinators, Routers, and End Devices (Fig. 2). ZigBee is one of the wireless protocols actively targeting remote control, building automation, and, importantly here, lighting control.
With ZigBee, there is always a single Coordinator node, which is the root of the network tree. It is this point that can bridge it to other networks as a sort of gateway point. Routers act as transit points along a mesh topology, which allows source and destination points to dynamically connect.
Figure 2: ZigBee architecture is a flexible mesh that uses routers and a single coordinator as the network root.
What is nice about ZigBee is that its comprehensive protocol stack creates and maintains routing tables that dynamically re route a connection if a commonly used node is busy or broken. An automated lighting technology will not survive if lights do not go on when you need them, so a robust control plane that is fault tolerant is something to keep in mind.
Wi-Fi, too, has a well-defined upper layer component called a mobile station (analogous to the STA in wired Ethernet) and an Access Point (AP) that can implement Basic Signaling Services (BSS) and provide a global presence on the Internet. Without an AP, an ad-hoc point- to- point type of network topology can be created and each node must implement Independent Basic Services (IBSS) and know each other’s SSID, in this case, a BSSID. This can allow for mesh-type topologies.
When an AP is present, data routes through it to all the STAs and the AP can provide services like broadcast, repeater, multicast, QoS, and network management. An AP can act as a layer 3 router and switch between devices in a client/server fashion (Fig. 3). It can also provide a cloud-based presence for control or reporting. Note that channeling, communications through the AP is less efficient than direct communication, adding complexity and cost to the requirements of the AP compared to the functionality of an individual node.
Figure 3: An IP network can permit an access point to be a control node, or a gateway point to the cloud, or both. This allows local only or global access and control while permitting interoperability of different but related devices. Here, for example, a curtain controller for a conference room could be part of the automated lighting solution for presentations and training.
Using TCP/IP as a native transport makes it rather transparent to connect a Wi-Fi network to the cloud. Another protocol that can benefit from this is 6LoWPAN, which is an IPv6 based wireless protocol with the “Internet of Things” philosophy as an underlying intent. It is designed to run on an 802.15.4 backbone, which is where ZigBee, WiHART, and ISA100 exist, but it uses IPV6 packets targeting low power, medium range, (10 meters typically) and fairly low bandwidth (100 – 250 Kbits/sec) applications. It is a lower power lesser range Wi-Fi for non-data- intensive applications.
This is rather well suited for lighting control, energy management, and automation solutions. Control packets are not data intensive and can be tokenized to use very little bandwidth, so less data means less power which, in turn, means longer life on limited power.
A unified or a modular approach?
With IP protocols, a simple switch or router can bridge everything to the cyber world. Every light element, every switch, every outlet, all with an IPv6 address and route path, controllable and query-able at any time any where in the world. This is the simplest architecture to implement, deploy, and use. However, it may not be the most cost-effective approach. Adding Wi-Fi to every fixture, keypad, sensor, or node takes more current during discovery and has more packet overhead, meaning possibly longer transmission times.
A dual protocol interface as an access point may be the answer (Fig. 4). This permits a simpler and lower cost microcontroller to reside in the many lower cost and distributed nodes. As we noted earlier, cost is key. You may expect that people will pay around $20.00 for an interface box added to their Wi-Fi network, but they will balk at paying that for a light switch. Unless that light switch enables a higher level of functional control (like turning on the lights from your driveway with your smartphone or your automobile’s control console), customers will not see the benefits of spending $10 or $20 as opposed to the 99 cent cost of a regular light switch.
Figure 4: A Wi-Fi Gateway Box lets Wi-Fi be the main transport for computers, tablets, and media devices. Lower cost RF standard and simpler protocols can reside in distributed, price sensitive devices like switches and lights. Access to lighting and building control is still possible with the added security of the gateway box.
An interface gateway also is a security benefit. It can still provide access to/from the cloud, but in a procedural way. The interface can have its own security and embedded intelligence to restrict certain operations. For example, someone with malicious intent may want to turn all your lights on full and jack up the thermostat if he gains complete access to you internal network. A rule- based interface gateway may have the intelligence to not permit this, or even override it if it occurs.
Wireless MCUs and radios
As a developer who must ensure that products work and play well with other devices, it boils down to three options: a single chip radio/micro, a two-chip micro and radio, or an OEM module. The three are not mutually exclusive. You may want to start out with a module for feasibility studies and then roll your own when ready to reduce cost.
Many well-engineered solutions are ready for implementation from suppliers who provide chips, development kits, protocol stacks, and IP. Take, for instance, Atmel, a company that is a very strong ZigBee supporter. A part like the 8 bit, 16 MIP ATMEGA256RZBC-8CU has 256 K of Flash, 8 K of RAM, and 4 K of EEPROM to help support the IP stack and the application at hand. It is supported by a variety of development tools like the ATAVRISP2 in system programmer and the ATAVRRZUSBSTICK ZigBee evaluation board, which supports the AT86RF230-ZU transceiver. Introductory training modules on MCU Wireless Solutions and An Introduction to ZigBee Networks can be found online on the Hotenda website.
Even chipmakers are unsure which wireless protocol will dominate a given application, that is why parts like the Atmel AT86RF230-ZU support ZigBee but also 6LoWPAN, RF4CE, and other ISM applications.
An interesting solution comes from Cypress with its CYWUSB6953-48LFXC wireless USB device. This single device Micro and Radio embeds an 8-bit controller along with a 2.4 GHz direct sequence spread spectrum transceiver to provide ranges up to 50 meters with 62.5 Kbit/sec throughput rates. While wireless USB is not as popular as Wi-Fi it is a protocol, consumers are comfortable with and this may be advantageous.
Not surprisingly, Wi-Fi is increasingly coming within reach, as it is no longer the sole property of chipmakers who demand usage in the millions before they will even speak to you. Take a look at the Maxim MAX2828ETN+ 5 GHz Wi-Fi solution. This part targets 802.11 b and g so it can act as an 11 Mbit/sec transceiver. Maxim’s MAX2839 is a corresponding 2.4 GHz part.
Not wanting to be left out of a significant market opportunity, Texas Instruments offers ISM and ZigBee processors, transceivers and SoCs within its CC2x family of devices. The CC2511F32RSPR, for instance, is a 2.4 GHz 500 Kbit/second RF transceiver for ISM band applications. The part is supported by the CC2511EMK eval module, the CC2510-CC2511DK development kit, and multiple TI training modules such as Introduction to Wi-Fi Technology.
For dual-mode gateways, modular solutions are also readily available. For example, combo modules like the Taiyo Yuden WYSBMVGX8, Murata LBEE5ZSTNC-523 or RFM’s WLS1271L combine Wi-Fi and Bluetooth on a single, reasonably priced module. These provide fast and low risk ways to test and determine the feasibility of a proposed design.
Do not get caught in that trap where you assume modules will always be a higher cost option. Check out the relatively low cost CC3000MOD modules from TI. These 2.4 GHz SimpleLink family parts tout a 54 Mbit/sec data rate and take aim at network processors and protocol gateway box applications and have an embedded IPv4 and IPv6 stack.
Manufacturers of LEDs and lighting fixtures will be working together with chipmakers to support wireless protocols and communications standards. The result will be a wide variety of chips, chipsets, and modules that allow quick development and certification as well as reduced cost – thanks to (often) free IP and stacks.
For more information on the parts and technologies discussed in this article, use the links provided to access information pages on the Hotenda website.