[an error occurred while processing this directive]

Recently, the automatic identification and data capture (AIDC) industry embraced both passive and active radio frequency identification (RFID) tags with improved designs for asset tracking and real-time location positioning applications.

Numerous market forecasts project significant volume deployment of these tags within the next few years as standards continue to become ratified, implementation costs decline further, and new features such as real-time location tracking become possible¹. In the near future, it is expected that most high-performance passive (battery-less) RFID tag technologies will transition from operation at 13.56 MHz to UHF frequencies between 862 MHz to 928 MHz across Europe and North America. It is also expected that active (battery-powered) tag technologies will seek to become protocol-compatible with existing low-cost communication system standards so as to “peacefully co-exist” in the unlicensed 2.45 GHz industrial, scientific, and medical (ISM) band. Doing so leverages design synergies with other ISM band products such as Bluetooth and IEEE 802.11, provides infrastructure added value, and reduces overall deployment costs.

Bluetooth- or 802.11 protocol-compatible RF tags will not necessarily share the same physical layer modulation scheme. However, the designs can leverage the same media access mechanisms and bit organization of the communication frames. This will facilitate simplified and cost-reduced multiprotocol baseband processor designs that can switch between different RF front ends.

A power transfer theory will be developed that presents the trade-off between maximum operating distances for passive tags and receiver sensitivity as this ultimately determines the maximum achievable data rates. Table 1 evaluates designs and key performance trade-off for active and passive RF-tag systems. The values associated with each parameter are typical for most designs.

Also, the article will address how receiver sensitivity and RF propagation models are combined to derive a single expression for range and data rate for active tag designs based on backscatter radiation rather than active RF transmission. Many RF tags use backscatter radiation to save energy during transmission. They do so by switching the antenna impedance so as to reflect or absorb electromagnetic energy in synchronization with the transmitted bit stream.

For reception, any traditional demodulation technique can be used. However, non-coherent ASK demodulation is generally used because it is simple and inexpensive. To extend battery life even further, some active tags also incorporate a carrier sense circuit that triggers a wake-up sequence to the controller logic. It will also be shown that increasing rather than decreasing the data rate achieves lower energy consumption per unit of information transmitted. Finally, in Section 4 results are provide from examples using the parameters described in Table 1.

In the far field^3, where r >> λ/(2π), and under line-of-sight conditions, the power received by the tag depends on its angular orientation (θ, ɸ) and distance r with respect to the transmitter, as well as the respective antenna gains and matching efficiencies as:

where γ_T, G_T and γ_r, G_r are the power-matching efficiencies and directional power gains of the transmitting and receiving antenna systems respectively, and s is the source power delivered into the matching network that drives the radiating element shown in Figure 1.

The source power is a function of the source voltage V_s and the source impedance Z_G in a typical implementation of the power amplifier that drives the antenna system. The matching efficiency γ is a function of the voltage standing wave ratio (VSWR) resulting at the interface between the power source and the antenna-matching network. This relation is derived in the appendix as:

For any antenna the directional power gain G is a function of the antenna radiating efficiency a_r and its directivity D, and is defined as:

The antenna radiating efficiency α_r is further defined in terms of the dissipative parameters of the radiating element:

where R_r represents the radiation resistance of the antenna and R_L the dissipative losses. X_A shown in the figure models the reactive elements, which restrict the signal bandwidth. However, we can eliminate this reactive element in the equations by considering only the practical case whereby the antenna's resonance frequency matches that of the carrier. The antenna directivity D is independent of transmission losses and is strictly a function of its radiating pattern.

The rate of power decay increases substantially in a multipath environment, depending on the amount of obstructions and reflections that the RF signal experiences in its path. Signal attenuation under various multipath conditions has been adequately modeled² as:

where the parameter N_B characterizes the rate of signal attenuation in a given environment, and R₀ the breakpoint distance where free-space path loss transitions into more severe path loss due to multipath interference and non-line-of-sight conditions.

N_B ≈ 4 and R₀ = 4 meters represent typical signal attenuation conditions that can be measured in a cluttered warehouse², particularly near docking bays where RFID tags are most frequently read in bulk. For example, the signal traveling towards a passive RFID tag will generally loose power at the normal rate of free-space path loss and will gradually transition to a higher rate loss near a distance R₀ from the transmitter as shown in Figure 2.

Active RFID tags customarily operate at much greater distances than R₀ and will, therefore, receive a much weaker signal than a typical passive tag.

This article will compute the expected maximum read range and data throughput for both passive and active RFID tags. It will be assumed that the passive tags will operate relatively close to the reader antennas, and as such will receive signal energy that has undergone only free-space path loss. On the other hand, it will be further assumed that the active RFID tags will receive and reflect (backscatter) signals that has undergone more severe attenuation due to multipath conditions.

Calculations will also derive the powering range for passive UHF tags by accounting for the antenna efficiencies, gain, and matching conditions.

The voltage realized across the load ZR is given as:

where AF is the antenna factor³, defined in the antenna factor section of the appendix, and E_in is a uniformly distributed electric field density across the aperture of the receiving antenna. The parameter α_cable represents any loss caused by cables running from the antenna to the receiver.

Figure 3 illustrates the receiver model showing the open-circuit voltage V_OC developed across the antenna due to the impinging electric field E_in. The power delivered into the matching network, which drives the load, is shown as P_M.

Substituting the antenna factor from (25) and the received electric field density from (28), both given in the appendix, into (6) yields:

It is assumed that there is negligible cable loss, α_cable, because the tag's integrated circuit is generally flip-chip mounted directly onto the antenna. This equation can also be rewritten in terms of the matching efficiency coefficients γ, and solved for the operating distance, r.

Therefore, (8) gives the maximum operating distance for a passive tag operating in the far field with equivalent operating load R_R and minimum turn-on voltage V_R. Equivalently, we can also write this in terms of the minimum power consumption required by the tag:

Under normal circumstances, internal rectification diodes limit the turn-on voltage for the tag. This is why the equivalent load resistance and minimum operating voltage is most often specified rather than the minimum turn-on power.

We observe from (10) that the wavelength of the transmitted RF energy provides a linear improvement in passive tag operating distance while the transmitted power provides only a square root improvement. This is also intuitively sound because the effective length of the RF collection aperture is linearly dependent on the wavelength whilst power in free-space diminishes with the square of the distance. This also explains why the operating distance⁴ for longer wavelength UHF (862 to 928 MHz) tags is significantly greater than that for shorter wavelength microwave (2400 to 2483 MHz) tags operating at identical output power levels and antenna gains.

Passive and active tags that use backscatter radiation to communicate do so by modulating their antenna impedance in synchronization with the transmitted bit stream, which can be patterned identically to that of a Bluetooth or 802.11 stream. Backscatter radiation results in a very low modulation index for amplitude modulated (AM) signals at the receiver. The rate of modulation or the change of phase of this AM signal typically encodes transmitted data on a frequency modulated (FM) or phase modulated (PM) sub-carrier, respectively.

When reflecting the base station continuous wave signal, the antenna characteristics are modified so that it becomes mostly a poor collector of RF energy. For example, shorting the two terminals of a dipole or switching in an extra capacitor or inductor tap is a simple way of configuring the antenna as a reflector or poor collector. In general, this does not significantly change the directivity D of the antenna because D depends mostly on the mechanical configuration and connection points resulting in an electrical length. However, the antenna gain G will be significantly modified by virtue of the change in antenna radiation efficiency α. Subsequently, the matching efficiency γ will also change because the antenna impedance has changed.

In the appendix, the following expression is derived for the reflected power at the antenna:

where Ψ_T is the transmitting antenna realized gain and ψ_r is the backscatter receiving antenna loss factor. This energy will in turn propagate toward the base station or the reader.

Therefore, as shown in the appendix, an expression:

for the signal power received by the base station can be derived. Because we are mostly concerned with the direction of maximum received energy, we have dropped the angle notations indicating beam directionality.

At the maximum operating distance rtag, the signal energy received by the base station under line-of-sight conditions is found by substituting (10) into (12) yields:

This expression is useful for determining when the ability to transmit sufficient power, rather than the receiver sensitivity, ceases to be the dominant failure mechanism for passive RFID systems. For example, we observe from this equation that the less power the tag requires and the greater the transmitted power, the lower the received signal strength is and the more difficult it becomes for the base station to recover the transmitted information from the tag. This is intuitive because the lower the required tag operating power and the greater the transmitted signal strength, the further away from the base station the passive tag will operate. At this increased distance, the reflected signal arriving at the base station would have attenuated significantly by an inverse exponential power of at least four under line-of-sight conditions.

It is also useful to note that (13) depends only on the tag antenna characteristics and not on the base station antenna characteristics. This is also intuitive because increasing the base station transmitter antenna gain will deliver more power to the tag and increase its maximum operating distance. However, the weaker reflected signal will also experience the reciprocal compensating gain of the receiving antenna. Therefore, the base station will see no change in its received signal power at the greater tag distance. Increasing the tag's receiver gain, however, will increase its maximum operating distance, but this results in a signal power reduction at the base station.

For active tags utilizing backscatter radiation, we expect that the distances would be much greater than four meters and so the signal received by the base station will experience typical multipath fading. By combining Equations 5, 11 and 12, we derive an expression for the signal received by the base station:

The noise power at the input of the receiver caused only by the source resistance is derived in the appendix and is given as:

For the receiver sensitivity, we simplify our analysis by ignoring any other independent noise sources that may be present at the input of the receiver, for example, noise from board layout and power supply coupling. If necessary, these could be lumped into the input-referred noise current as shown in Figure 4. We do, however, include internally generated noise voltages and currents from the active stages of amplification and filtering into a single factor called the noise factor. This is defined as:

The receiver manufacturer typically specifies this factor. The terminology generally changes to noise figure when this quantity is expressed in dB. The noise present at the input of the demodulator is then conveniently defined as:

The minimum signal power the receiver requires to demodulate the received signal into a valid bit stream with some specified probability of bit-error is generally referred to as its sensitivity. This implies some level of signal-to-noise (SNR) at the input of the demodulator. Each modulation scheme requires a different level of SNR and bandwidth for a given probability of bit error, Pe. Backscatter radiation results in an amplitude-shift-keyed (ASK) data stream at the receiver. This modulation rate is also not necessarily phase-coherent with the carrier.

Therefore, the base station must employ non-coherent demodulation of the ASK signal. For non-coherent ASK demodulation, the required SNR [5] is given as:

where E_b is the energy per bit in Joules, and N_o is the single-sided noise power spectral density in W/Hertz.

For example, we can see from a plot of this equation in Figure 5 that to achieve a bit error rate (BER) no greater than one part in one million (10-6), we will require an E_b/N_o at the demodulator that is slightly greater than 14 dB. The ratio E_b/N_o is related to the average SNR of the modulated carrier. It is derived in the appendix and is given as follows⁵:

where P_av is the average signal power, N_av is the average noise power, R_bit is the data rate in bits per second, and B_w is the signal bandwidth in Hertz. For ASK, the ratio B_w/R_bit is 2.

We can now relate the SNR at the base station demodulator to the required BER as follows:

Substituting the expressions from (14), (17), and (19) yields (21):

which is the maximum data rate possible with non-coherent ASK demodulation at the base station. This expression can be intuitively interpreted as showing that the bit rate is linearly proportional to realized or net power gain including losses due to the transmitting and receiving antenna systems, respectively. Bit rate is also inversely proportional to the signal propagation losses at a given distance based on the channel propagation characteristics.

The expression also shows that for a given bit error rate, the bit rate is inversely proportional to the base station receiver noise figure. Hence, we can conclude that given no other changes, we can improve the bit rate or throughput of the communication system by the same factor that we can reduce the receiver noise figure. That is, we can achieve higher system throughput at a given desirable maximum operating distance, maximum regulated output power, minimum achievable antenna system losses and the maximum desirable bit error rate “simply” by reducing the receiver noise figure. This of course implies that we can increase the active tag modulation speed, perhaps at the expense of higher burst power consumption.

Likewise, we must also increase the base station receiver bandwidth to accommodate the higher modulation speed while seeking to simultaneously improve its noise figure. Doing so implies increasing the receiver power consumption.

Given the same amount of information to transmit, a bit rate improvement results in shorter packet transmission times. This means that the radio will access the airwaves less often and for shorter time slots. Additionally, shorter packets are statistically less susceptible to interference from other equipment using the same frequency band. Therefore, fewer collisions and, correspondingly, fewer requests for re-transmission result. Higher bit rates will result in a decrease in the average power consumption of the radio per information unit communicated because the radio will tend to use the airwaves less often.

Hence, given a semiconductor process, the penalty for increasing the instantaneous power consumption to achieve lower noise figures with higher bandwidth or more bandwidt- efficient modulation schemes can be offset by the gains realized from reduced use of the airwaves per information unit communicated.

This expression also provides the following additional insight. The maximum possible bit rate for a given communications distance r improves with the fourth power of the carrier wavelength l. The implication here is that given the greater signal collection capability of antennas at lower frequencies, and baring interference, we can use more bandwidth at lower frequencies to increase data rate without sacrificing bit error rate. Unfortunately, greater bandwidth is not always available within the lower frequency UHF bands — compared with higher frequency microwave bands, for example.

Given passive tags operating in the 915 MHz industrial, scientific and medical (ISM) band under FCC power levels of 1 W delivered into an antenna with the parameters listed in Table 1, we calculate the range degradation and data rate increases with increasing power consumption of the tag.

The results are normalized to the maximum operating range (at the minimum power consumption level), and the maximum data rate (at the maximum power consumption level.) These are 12.1 meters and 62.8 megabits per second, respectively. These results show that under free-space conditions the most significant range improvements occur when tag power consumption is reduced below 35 µW. This still leaves a potential for about a six Mb/s data rate, which is generally not necessary for passive tags.

Figure 6 exemplifies the fact that the ability to deliver power to the tag is typically the dominant factor that limits the maximum operating distance and not the receiver sensitivity or desired data rate. However, by quantifying this information, we gain some insight as to the degree to which we can relax the receiver sensitivity (for example by increasing the noise figure specifications and lowering cost) for full-duplex passive RFID tag readers given the expected read range. This is a key consideration for reader design optimization because of the challenges in recovering a signal of tens of nano-Watts that is buried within a carrier signal in the hundreds of mW level.

Given tag readers operating under reduced power transmission levels of 100 mW, which is identical to that of 802.11b wireless LAN systems, we calculate the range and data rates based on backscatter radiation from the tags using the design parameters of Table 1.

The results are plotted in Figure 7 and indicate that within the 915 MHz frequency band, and at a distance of about 10 meters from the base station, data rates between 10 kb/s and 1 Mb/s are possible depending on the multi-path conditions and receiver sensitivity. In contrast, data rates between 200 b/s and 30 kb/s are possible when operating in the 2.45 GHz frequency band.

We see that to achieve data rates similar to 802.11 and Bluetooth systems, the tags must operate within four meters of the transmitter. This has important implications for Bluetooth or 802.11-compliant RFID tags operating in the backscatter mode within the 2.45 GHz frequency band. For maximum throughput efficiency, these types of tags should be designed with adaptive rate scaling so that neither range nor data rates will be sacrificed when propagation conditions are favorable.

Because better receiver sensitivity and higher SNR is possible with active (non-backscatter) transmission, we should expect orders of magnitude improvement in the range and throughput over backscatter based systems. However, we give up the low power consumption, and hence long battery life (or smaller, lighter, cheaper battery) characteristics that backscatter radiation would provide. Battery life is generally the most important feature amongst similarly capable active RFID tags.

We have derived expressions for the maximum expected operating distance and data rates for passive tags given basic antenna characteristics and line-of-sight conditions. For active tags, we assumed multipath signal propagating conditions and derived a single expression for data rate as a function of operating distance, bit error rate, operating frequency and net realized power gain due to system losses. We then compared, via examples, range and data rates for backscatter-based RF-tags operating in both UHF and 2.45 GHz frequency bands at identical (also typical) power levels, and at the extremes of expected signal propagating conditions. Results show that within a range of multipath fading conditions, both active and passive tags, which operate in the UHF frequency bands, will provide significantly improved range and throughput over those operating with similar parameters in the microwave (e.g. 2.45 GHz) frequency bands.

From this analysis, we further conclude that to benefit from the power efficiency of backscatter-based communications at 2.45 GHz, we should expect significantly reduced range and data rates from that possible with active transmit systems (e.g. Bluetooth and IEEE 802.11). Many active RF tags typically utilize backscatter-based communications because battery life is often more important than either range or data rates in their applications compared with wireless networks.

The analysis indicates that both improved channel utilization and greater power efficiency results from improving the bit rate at the expense of reducing the receiver noise figure, even though doing so may require somewhat higher instantaneous power consumption during communications. Higher bit rate per information unit results in shorter packets that are more likely to “squeeze” through channels characterized by frequent but random interference bursts. We saw that without sacrificing range or bit error rate, we can increase the bit rate by moving to a lower frequency (longer wavelength) band that provides similar channel bandwidth (e.g. from 2.45 GHz to 915 MHz.)

The reflection coefficient ρ_v is a complex valued variable that characterizes the power transfer ability of the matching circuit between the source and load. For example, a reflection coefficient with unity magnitude and zero phase indicates that all of the power delivered into the matching circuit was reflected back into the source. For mathematical convenience, we can define a “matching efficiency” coefficient to describe the power transmission characteristics of the matching circuit:

Because the magnitude of the reflection coefficient is related to the VSWR, Г³ by:

we can rewrite the matching efficiency as:

From antenna theory, we define the ratio of the electric field strength E in the direction of the antenna's polarization to the received voltage V_R as the antenna factor AF. This quantity has units of inverse meters and is derived in numerous texts₃ as:

where R_R is the dissipative or real part of the complex load Z_R.

The electric field intensity is:

were S is the transmitted power density:

Substituting (27) into (26) yields:

The power density at the aperture of the tag antenna is given by (27) in w/m². The maximum collection aperture of a lossless antenna is given as:

and that of a practical, lossy antenna is given as:

where both quantities are in square meters. Therefore, the reflected power will be:

Substituting for the power density S yields the reflected power at the antenna as:

For notational convenience, we can further define a realized backscatter antenna loss-factor as:

and a transmitting antenna realized power gain as:

We can now rewrite and simplify the expression for reflected power at the receiving antenna as:

The base station will receive a signal that propagates from a source represented by the reflecting tag antenna as:

From reciprocity in antenna theory:

and

Substituting for P_re we get:

or equivalently,

Referencing Figure 4, the input voltage to the LNA and down-converter block is:

For maximum power transfer:

The noise source is, therefore:

where B_w is the noise equivalent bandwidth of the system. The substitution of (42) and (43) into (41) yields:

Therefore, the input power is:

At room temperature (298° Kelvin), this quantity is approximately -174 dBm/Hz.

The ratio between the energy per bit, E_b and the single-sided noise power spectral density N₀ can be expanded as follows:

1. http://www.aimglobal.org/technologies/rfid/

2. McCune, E., and Feher K., “Near-Far Interference in Digital Wireless Communications,” Applied Microwave and Wireless, January 1997.

3. Smith, A. A., Jr., “Radio Frequency Principles and Applications,” I.E.E.E. Press, N.Y., 1998.

4. Bridgelall, R., “UHF Tags — the answer to the retail supply chain prayers?,” RF Innovations Magazine, August 1999.

5. Skalar, B., “Digital Communications Fundamentals and Applications,” P.T.R. Prentice Hall, N.J., 1988.

Raj Bridgelall
Symbol Technologies, One Symbol Plaza, Holtsville, NY 11742
rbridge@symbol.com