The total signal at the receiver is the vector sum of all these contributions, most of which are much larger than the wanted tag signal, with appropriate amplitudes and phases, most of which are unpredictable a priori. Thus, the actual effect of a given change in the load on the tag antenna on the receiver signal is completely unpredictable and uncontrollable. For example, modulating the size of the tag antenna current (amplitude modulation) may not result in the same kind of change in the reader signal. In Figure 17, we show a case where changing the tag reflection from a large amplitude (HI) to a small amplitude (LO) causes the received signal to increase in magnitude without changing phase (the "AM" case). Changing the phase of the tag signal without changing the size of the reflected signal in order to symbolize a LO state may change the amplitude of the reader signal at constant phase (Figure 17, "PSK" case). The only thing we can say with any confidence is that when we make a change in the state of the tag antenna, something about the phase or amplitude of the reader signal will change. In order to make a backscatter link work, we need to choose a way to code the data that can be interpreted based only on these changes and not on their direction or on whether they are changes in phase or amplitude.

Figure 17. The Received Signal is not Simply Correlated to the Tag Signal. The AM Case Assumes the Tag Reduces its Scattered Magnitude Without Changing Phase; the PSK Case Assumes Phase Inversion Without Amplitude Change.

As a consequence, all approaches to coding the tag signal are based on counting the number of changes in tag state in a given time interval, or equivalently on changing the frequency of the tag's state changes. Therefore, all tag codes are variations of frequency-shift keying (FSK). It is important to note that the frequency being referred to here is not the radio carrier frequency of (say) 900 MHz but the tag (baseband) frequency of perhaps 100 or 200 kHz. A binary '1' might be coded by having the tag flip its state 100 times per millisecond, and a binary '0' might have 50 flips per millisecond. Because the frequency being changed is the frequency at which a carrier is being amplitude modulated, techniques like this are sometimes known as subcarrier modulation.

Let's look at one specific example of tag coding, usually known as FM0 (Figure 18). In FM0, the tag state changes at the beginning and end of every symbol. In addition, a binary 0 has an additional state change in the middle of the symbol. Note that, unlike OOK, the actual tag state does not reliably correspond to the binary bit: for example, in the left-hand side of the figure, two of the binary '1' symbols have the tag in the LO state and another '1' symbol has the tag in the HI state. Remember, the reader can't reliably distinguish which state is which but can only count transitions between them. The right side of the figure shows the baseband signal corresponding to a series of identical binary bits to clarify the correspondence of binary '0's with a frequency twice as high as that of binary '1's.

Different tag coding schemes can be used to adjust the offset from the carrier frequency at which the signal from the tags is found. As we will find in Chapter 4, readers have an easier time seeing a tag signal when it is well separated from their own carrier frequency, so higher subcarrier frequencies help improve the ability to read a tag signal. However, if the separation is large compared to the channel size, the tag signal might lie on the signal of another reader in a different channel. Just as with readers, increasing the data rate of a tag signal tends to spread the spectrum out in frequency. To have a flexible choice of tag data rates while minimizing noise, the reader needs to be able to adapt the band of frequencies it tries to receive, adding cost and complexity.

Figure 18. FM0 Encoding of Tag Data.

In real receivers, noise and interference may be present as well as the desired signal. A certain minimum signal-to-noise ratio (S/N) is necessary for each type of modulation in order that it can be reliably decoded by the receiver. The exact (S/N) threshold depends on how accurate you're trying to be and to a lesser extent on the algorithms used for demodulation/decoding. For RFID using FM0, (S/N) of around 10 or better (10 dB or more) is usually sufficient. (Requirements for demodulation of reader symbols, like PIE, in the tag are generally similar.) As we will see in Chapter 8, modern protocols provide alternative modulations that can operate with smaller (S/N) ratios, at the cost of a reduction in the tag data rate.

Next: Link Budgets

About the Author
Daniel Dobkin is an RFID consultant, writer and teacher. He holds six patents as inventor or co-inventor. He is the author of such books as: Principles of Chemical Vapor Deposition and RF Engineering for Wireless Networks, and The RF in RFID. Additionally, he is a published author of 25 technical publications. He has taught RFID courses internationally in Singapore for the SMa/RFID Focus; and domestically at SDForum, Mitre Corporate University, and San Jose State University. Daniel is a Stanford University PhD in Applied Physics. He has MS and BS degrees from CalTech.

Printed with permission from Newnes, a Division of Elsevier. Copyright 2007. "The RF in RFID: Passive UHF RFID in Practice by Daniel M. Dobkin. ISBN-10: 0750682094For more information about this title and other similar books, please visit www.elsevierdirect.com.