1.1 What “Low Power RF” is All About
1.1.1 Defining “Low Power”
The obvious: a solution where the RF transceivers use a minimum of energy to communicate with each other, and where periods without communication are characterized by a minimal amount of energy spent idling. To quantify this statement for 2009, a low power RF technology worth its salt has no problem operating at an average current draw under 0.1 mA and a max current draw under 50 mA. Some technologies achieve far lower figures; for example, a well conceived ISO 18000-7 solution can easily average under 0.05 mA and max under 20 mA when using a low-leakage lithium battery.
1.1.2 Defining “RF”
RF stands for “Radio Frequency,” and it is used to mean just that. The nuance here has more to do with the application than the method of communication. Low Power RF products need:
• RF silicon parts, ideally with as much integration as possible (i.e. a single chip is better than two chips).
• Power supplies, which are usually batteries. Recently, no shortage of attention has been paid to so-called “energy harvesting,” where the idea is for the low power RF device to absorb energy from its environment.
• A microcontroller, which contains a small CPU and memory. Again, integration is important. For 2009
and beyond, designers should expect RF silicon and the microcontroller to be in one package.
• Some kind of antenna for conveying the RF energy.
• Optionally sensors, which are typically silicon parts themselves and hence also benefit from integration.
1.1.3 RFID
In the most basic sense, RFID (Radio Frequency Identification) encapsulates several low power RF technologies and product lines. These are referred to as “active RFID.” In the other corner is “passive RFID,” whose inherent asymmetry makes it a poor example of low-power; it requires a very high-power transmitter (often called an interrogator) while the transponder (tag) must exhibit very low power characteristics. These systems do not require batteries in the transponders, which, behaving in a similar way to RADAR, reflect and modulate the incidental signal from the interrogator. No one considers RADAR to be a low power technology – even though RADAR targets do not contain batteries – and neither should they consider passive RFID a low-power technology. Of course, logical arguments do not always win. Market forces have led to confusion when it comes to RFID. The general perception among laymen and even some self-styled industry experts is that passive RFID embodies the general term, RFID. For this reason, we also will attempt to divorce ourselves from the practice of using the term “RFID” with respect to any low-power RF system.
1.1.4 BLAST
DASH7 has been designed to operate using the “BLAST” concept: Bursty, Light-data, ASynchronous,
Transitive. Despite being another acronym of questionable genuineness, BLAST does actually correlate to the DASH7 operational philosophy on a one-to-one basis:
• Bursty: Data transfer is abrupt and does not include content such as video, audio, or other isochronous (i.e. streaming) forms of data.
• Light-data: In conventional applications, packet sizes are limited to 256 bytes. Transmission of multiple,
consecutive packets may occur but is generally avoided if possible.
• Asynchronous: DASH7’s main method of communication is by command-response, which by design requires no periodic network “hand-shaking” or synchronization between devices.
• Transitive: A DASH7 system of devices is inherently mobile. Unlike other wireless technologies DASH7
is upload-centric, not download-centric, so devices do not have to be to be managed extensively by fixed
infrastructure (i.e. base stations). Most wireless technologies throughout time have been designed to replace wired networks (it's called "wireless" after all). Wired networks cannot possibly be conceived to meet
the needs of DASH7 applications. DASH7 applications are inherently mobile; devices and infrastructure can be mobile, and it is even difficult to consider an alternate, wired network that could provide roughly similar function. BLAST as a concept fits into this application model, and it suits low power RF extremely well. DASH7 systems should be understood not as conventional networks.
2.2 A Survey of “Low Power” RF Standards
2.2.1 Standards and Their Proponents
There are many self-proclaimed “low power” RF products that have been available for years. At least by their names, most of them should be familiar to anyone who keeps in touch with the high-tech economy. The successful standards are backed by a couple of industry heavyweights, but when it comes to low power operation, not all of them are in the same league.
2.2.2 Strong Standards are Important
ISO 18000-7 is unique among all of the standards described herein because it is an ISO standard. While this is a seemingly redundant observation, it does emphasize both the ability to enforce the standard globally as well as the responsibility of its backers to cultivate effective structures for compliance and interoperability. Only with mixed success have the alliances that govern many of these standards been able to enforce compliance, interoperability, and industry cooperation:
• The WiFi Alliance has been generally successful but the internal battle between versions a and g cost
Members unnecessary time and money. The draft-n process is another example of mediocre cooperation in
Standardization.
• Since ZigBee’s inception, the alliance has had difficulty in getting solutions developers to adopt ZigBee for low power RF. The IEEE is now attempting yet another variation (f ) in hopes of success.
• Proprietary technologies deny their markets entirely of all of the re-sources available to technologies supported by alliances, including interoperability. ISO 18000-7 honors the ISO tradition by granting explicit terms of what complies and what level of functionality is mandatory, and also stipulating explicit compliance metrics via ISO 18047-7. Of course, it is possible to maintain a strong standard without ISO: Bluetooth is a good example of an older standard whose SIG has been successful in cultivating a sound level of interoperability as well as a marketplace full of low-cost silicon amid a constantly evolving strata of feature-sets.
2.2.3 Simple RF Interface, Low Power
Despite evidence that the “usual suspects” We have observed varying degrees of success as standards, there is room in the marketplace for all of them. They each perform certain tasks unquestionably better than do the others. However, when it comes to delivering an industrial, low-power RF system, it is hard to beat ISO 18000-7. It has been designed to perform a small but well-defined set of features with maximal efficiency, these being inventory collection and bursty, asynchronous communication between small transceivers (e.g. tags) and infrastructure products. In comparison the other standards seem less focused. Table 2.2a should open up the floor for some debate. The bandwidths, channels, and data rates are more or less objective. The operating power figures are more subjective, but they are founded on both empirical research (they are advertised attributes of real products) and assumptions.
Operating Electrical Current
The “best in breed” chips that were analyzed may not exemplify ceteris paribus comparison, but they are all
state of the art for their particular technology. Short of a breakthrough, future advances for each technology can be expected to improve along the same ratios that the existing implementations exhibit.
As we can see, the existing best in breed implementation for ISO 18000-7 has an impressively miniscule power budget. This is in part due to its use of the 433 MHz frequency band because, from a scientific standpoint, there are some inescapable rules when it comes to electric current requirements for semiconductors:
• In any given RF system there’s a point at which increasing maturity of the silicon delivers marginal returns on reducing system’s current draw. This is due mainly to electric current leakage.
• No matter how mature the digital silicon, amplifying the RF circuits will always expend more energy when the bandwidth is large and the band is high. This is due to the dynamic characteristics of the silicon-CMOS transistor topologies used in practically all modern, integrated RF chips.
The combination of these two rules yields this corollary: for a given application, if in place of wideband, higher data rate, higher frequency RF systems it is realistic to substitute narrowband, lower data rate, lower frequency RF systems, the latter will always yield the lower system power.
Communication Range:
The standards above do not all put the same limits on transmit power. For Bluetooth, ZigBee and ISO 18000-7, 0dBm @ 50 Ohms is the reference value which yields the nominal range as discussed in product or standards literature. It is certainly possible to improve range by increasing the transmission power or increasing the sensitivity of the receiver, although governments often have regulations regarding allowable transmit power. Incidentally, both of these techniques also increase the power requirement of the system.
The standards above do not all put the same limits on transmit power. For Bluetooth, ZigBee and ISO 18000-7, 0dBm @ 50 Ohms is the reference value which yields the nominal range as discussed in product or standards literature. It is certainly possible to improve range by increasing the transmission power or increasing the sensitivity of the receiver, although governments often have regulations regarding allowable transmit power. Incidentally, both of these techniques also increase the power requirement of the system.
2.2.4 Simple Protocol, Low Power
Referring back to Table 2.2a, we can see that in each solution the power differs for receive, transmit, and sleep modes. Techniques for optimizing low-power RF systems always seek to maximize the amount of time spent in sleep mode, or, from another perspective, minimizing the amount of time spent in active modes. More-so than data-rate, the protocol is the means by which time spent in active modes can be determined. Good low-power RF solutions have protocols that do not specify extraneous features. In other words, these solutions are defined by considering not what features you could use but instead what features you could do without. Of the depicted solutions, the simplest protocols belong to ISO 18000-7 and low energy Bluetooth (aka wibree), although they operate very differently. The diagram below intends to show time spent in active modes vs. sleep for these two protocols, while also showing the amount of power consumed during each operational state. As we can see, low energy Bluetooth does not adhere to BLAST principles, but because it is just a wire-replacement technology it can succeed nonetheless. The other technologies, ZigBee and WiFi, have protocols that are complicated enough that a diagram such as the ones below cannot come close to representing the many modes of operation. In section 2.4.4 we show how ZigBee cannot deliver low-latency (BLAST-like) behavior without expending a lot of power.
2.2.5 Symmetric Protocol, Flexible Use
A symmetric protocol is one where there is little or no difference between the way any sort of device communicates with any other sort of device. Symmetry does not necessarily make a standard low power optimized, but it does allow for more flexibility or innovation in the way that standard’s technology is implemented and ultimately used. ISO 18000-7 uses a symmetric protocol, and certain modes of ZigBee are symmetric, as well. Low energy Bluetooth, WiFi, and other modes of ZigBee, on the other hand, are asymmetric as they are predicated on the existence of base station or coordinator type devices.
2.3 Communications Theory 101
It has now been established that, from a purely scientific approach, lower frequency radio waves are more reliable than higher frequency waves are at delivering a signal over range, line-of-sight or otherwise. Communications theory is an engineering discipline focused on attaching rules (i.e. math) to phenomena involved in sending data via radio signals. When given a problem to solve, communications engineers go back to the rules to determine the best solution. There are always trade offs. Nonetheless, the primary
solutions criteria depend on the following:
• Allowable minimum data rate
• Allowable signal to noise ratio
• Allowable complexity of transmitter
• Allowable complexity of receiver
Data Rate
In today’s world, data rate is often confused with the term “bandwidth.” The two are related, but they are not the same. Data rate is a digital phenomenon, expressing the amount of bits that a communication system can deliver in a given amount of time. Bandwidth is the frequency range between which a signal’s energy can be realistically confined. Some modulation techniques are more efficient than others at cramming data into available bandwidth. Generally speaking, the more complex the modulation the more efficient it is at cramming raw data into a given amount of bandwidth, but sometimes further means are used to spread the band (i.e. spread spectrum technologies) in order to improve tolerance of noise. The latest exotic and complex methods manage to do both, although they are completely unsuitable for low power RF because the transmitters and receivers are too complex.
Signal to Noise Ratio
Maximizing signal to noise ratio (SNR) is a pursuit in which communications engineers put in a lot of time.
Noise refers to any energy received [by the receiver] that does not come from the appropriate transmitter. Noise can be hard to predict, but there are some guidelines. A typical model for noise is additive gaussian white noise, as this is how “static” is modeled. It exists wherever there are charge carriers moving around randomly, for example in an antenna, and is often called thermal noise. The larger the bandwidth of the communication, the greater the received noise. There are other types of noise, too, and they all have one thing in common: the larger the bandwidth, the greater the potential for noise ingress. We are interested, however, in signal to noise ratio, not just noise, and by increasing the bandwidth through modulation or encoding techniques it is possible to boost the signal energy in greater proportion than noise and interference. This is the basis for improving SNR and decreasing the affect of noise. By convention, DASH7 uses a marginally wideband FSK modulation (check appendix for more on modulations). It is set up to provide reasonably good resilience to noise and interference without expending too much energy doing so. Low energy Bluetooth’s modulation is very similar. ZigBee, on the other hand, uses a more complex modulation called QPSK that manages to be slightly more efficient at cramming data into bandwidth as well as better suited to delivering higher SNR. The added complexity, however, comes not without a price.
Simplicity vs. Complexity
Limits on targeted solution cost, development cost, and power requirements force communications engineers to be clever. Often we can evaluate two technologies, one simple and one sophisticated or complex, where the performance gap between the two can be closed by enhancing other areas of the total solution. One good area of study is the receiver. At the cost of higher power requirements, a more advanced modulation scheme may prove to have superior SNR than a simpler one, and a higher data rate may allow error correction coding to be part of the message. However, by changing the carrier frequency of the signal or by taking special attributes of the signal into account, the simpler solution may even out perform the more complicated offering. Without considering interference, ISO 18000-7 offers a greater SNR link budget than does the more complex, more sophisticated IEEE 802.15.4. When interference is in fact considered, the busy 2.45 GHz band has an increasingly negative impact on the performance of IEEE 802.15.4 vs. ISO 18000-7, and it actually becomes distrastrous if newer 802.11n networks are in place. Performance enhancements like these that trickle-down from total system design can be relied upon when a technology is well defined to attack a focused set of problems. For example, we can see how design simplicity is maximized to solve problems of communication with small satellites. Section analyzes how a simple, clever solution (ISO 18000-7) can excel in BLAST type applications, one of which is even validated in RF performance of embedded devices in shipping containers. The basic understanding is that ultimate performance for general purpose solutions will always require a complex system design, it will be expensive to develop, and it will be very difficult to test for and enforce interoperability. In such cases where a focused design philosophy can be applied – where simple, accessible technology can meet performance requirements – engineers can quickly develop products, testers can easily achieve interoperability, and marketers can immediately target users. It is the “prisoners dilemma” for wireless standards: pursue the holy grail or pursue the strategy of most probable success. Simplicity and focus lend themselves to success.
2.4 Conclusion
So, “why UHF?” If the solution doesn’t need high data rate and can be band limited (partly a function of data rate), then using UHF makes a lot of sense. Compared to 2.45 GHz systems, UHF systems operate better in non-line-of sight conditions, they use less power, and they are so much more permissive that they can still offer superior range even when using suboptimal antennas. For the bursty, asynchronous type of solution that DASH7 targets, the UHF band is the perfect choice to deliver the long range, highly reliable signal it needs, all the while preserving a tiny power budget....
