By Sh r e h a rS h a r ao T e x aS I n S T r u m e nT S

Estimating the ZigBee transmission-range ISM band

DESIgnErS oF ShorT-rAngE wIrElESS DEvIcES In ThE 900-Mhz AnD 2.4-ghz BAnD nEED To unDErSTAnD whAT AnD how PArAMETErS AFFEcT ThE TrAnSMISSIon rAngE BASED on ForMulAS AnD BE ABlE To APPly ThEM In ForMulAS For STATISTIcAlly cAlculATIng ThE PATh loSS AnD rAngE For BoTh InDoor AnD ouTDoor EnvIronMEnTS.

s home, building, and industrial-automation applications go wireless, short-range wireless devices are receiving a lot of attention. Typically, these applications use either proprietary or standards-based approaches, such as ZigBee in the 900-MHz and 2.4-GHz ISM (industrial/ scientific/medical) bands. With the increased popularity of short-range wireless devices, its more important than ever for end-system designers to fully understand the range of wireless communications. This article discusses wireless propagation and develops models to estimate the path loss and range for short-range wireless devices in indoor environments. These models give system designers an initial estimate on a wirelesscommunication systems performance. Before exploring range-estimation formulas, designers need to understand the wireless channel and propagation environment. The wireless-radio channel is the transmission path between the transmitter and its intended receiver. Unlike wired channels, which are stationary and predictable, wireless channels are random, time-variant, and difficult to model. So, designers need to use statistical modeling for these random channels. Radio-wave-propagation models have traditionally focused
Table 1 average signal loss for radio-paTh obsTrucTion by differenT maTerials
material Type Metal concrete-block wall loss from one floor loss from one floor and one wall 6-in.-diameter metal pole concrete wall concrete floor Transmitter turning at right angle along corridor loss (dB) 26 13 20 to 30 40 to 50 3 8 to 15 10 10 to 15 frequency (Mhz) 815 815 1300 1300 1300 1300 1300 1300

on predicting the average received-signal strength at a given distance from the transmitter, as well as the signals strength variability in close proximity to a location. Propagation models that predict the mean signal strength for an arbitrary transmitter-receiver separation are large-scale propagation models and are useful in estimating the transmitters range. Conversely, propagation models characterizing the rapid fluctuations of the received-signal strength over distances of a few wavelengths are small-scale, or fading, models. This article focuses on the large-scale propagation model, which estimates the range of wireless transmission. The free-space-propagation model predicts the received signals strength when the transmitter and the receiver have Equations for MS4238 a clear, unobstructed line-of-sight path between them. The free-space model predicts that the received-signal strength decays as a function of the transmitter-receiver separation distance raised to the nth powerthe power-law function. The free-space power that 1 receivers antenna receives is Equation the separated from a transmitting antenna by a distance, which the Friis free-space equation defines:
PR (d) = PT G T G R 2 (4 )2 d2 ,

(1)

where PT is the transmitted power; PR(d) is the received power and is a function of the transmit-receive separation, d; GT is the transmitter-antenna gain; GR is the receiver-antenna gain; Equation 2 d is the distance between the transmitter and the receiver in meters; and l is the wavelength in meters. The Friis free-space equation shows that the received power falls off as the square of the transmitter-to-receiver separaPT G GR 2 P result suggests that the T 2 2 , power de= log received 10 tion distance. This L = 10 log P (4 ) d cays with distance at a rate of R dB/decade. 20 An important term in estimating the wireless-transmission range is path loss, which represents signal attenuation in deciEquation 3 bels. Path loss is the difference in decibels between the transmitted power and the received power at the antenna. From Equation 1, you can deduce the path loss as the transmitted power divided by the received power. Equation 2 defines the P 2 PL = 10 log T = 0 log 1 . path loss as: P (4 )2 d2
R

may 24, 2007 | EDN 67

Equation 6

Equation 2 Equation 2 PT G GR 2 = log T 2 2 , 10 PR (4 ) d P G GR 2 PL = 10 log T = log T 2 2 , 10 PR (4 ) d PL = 10 log

d PR (d) = PR (d 0 ) 0 . Equation 6 d

(2)

d Table Equation 7 2 faf (floor-aTTenuaTion facTor) for PR (d) = PR (d 0 ) 0 . signal peneTraTion across mulTiple floors d
no. of floors one Two Three Four Five floor-attenuation frequency standard deviation d factor (d (Mhz) 0 . (dB) P (d) = P(dB) ) + 20 log
R R 0 10 915 d

Equation where PL is path loss. To3simplify Equation 2, assume that both the transmitting and the receiving antennas have unity Equation gain, and this assumption results in: Equations 3for MS4238 Equations PT MS4238 for PL = 10 log = 0 log 1 . (3) 2 PR (4 )2 d2 PT PL = 10 log 1 . You can also express this P = 0 log the following usable Equation 1 equation in 4 )2 d2 ( R form: Equation 1 PEquation 10(fMHz)120log10(d)228, 520log 6 (4) L PT G T G R 2 PR (d) = 6 2 2 , or Equation (4 ) d 2 P G G PR (d) = T T 2 R2 ,2 PR5P 2P (5) (4 ) Td L, d PR (d) = PR (d 0 ) 0 . d where d is the distance in meters. 0 2 d PR (d) = PR (d 0 )can estimate the received-powThe Friis free-space formula d . Equation 2
2

Equation 7 13.2
18.1 24

9.2 8 5.6 6.8 6.3

915 915 915

d0 PR (27 8 d) 915 Equation = PR (d 0 ) + 20 log10 d .

27.1

Using the values in Equation 1, you obtain

2 0. 8 Equation0063(1)(1)(1 / 3) = 0.44 109 W. PR (100) = 2 2 (4 ) (100)

(8)

To calculate the power in decibels referred to milliwatts, you must express the power in milliwatts as: Therefore,
Equation 9 0.0063(1)(1)(1 / 3)2 PR (100(100)50.4431026 mW.0.44 109 W. (9) = PR ) = (4 )2 (100)2 100 1200

er level only for values of d that are in the transmitting antenEquation 2 Equation Fraunhofer region, of a transmitting nas far field. The far field, 7 antenna is the region beyond the far-field distance, dF. For an 2 Equation 7 2 antenna, dF is 2DL/l,10 log PT = log G T G R largest physical P = where D is the antennas2 , 10 2 2 ( d PR G4 ) than D and must linear dimension. Also, dFPmust be greater0Rd TG . PR = ) = PR T =+ 20 PL (d10 log (d 0 ) log10 2 2 , 10 be in the far-field region. This path-loss4formula applies only PR ( d ) d to ideal systems with)clear (d ) +of sight, d 0 . should use it and you PR (d = PR lines 20 log10 0 d Equation only for initial estimates. 3 Propagation models use3the close-in distance, d0, as the reEquation ceived-power reference point. You must calculate the received Equation 8 power, PR(d), at any distance greater than the received-power P 2 PL = reference 0 log . reference point with 10 log8 T =to1PR(d0), whose value you can Equation P 2 2 T predict from equations logand =Alternatively,. you can meas1 PR 4. 10 log (4 ) d PL = 10 PR ure it in the radio environment by taking daverage received (4 )2 2 . close distance2 power at many points from0a0063(1)(1)(1 / 3) from44 109 W. PR (100) = = 0. the transmitter. You must select the close-inreference distance so that the (4 )2 (100)2 2 0 Equation than the )(1)(1 / 3 = 0.44 far-field region is R (100) = 6 .0063(1close-in)distance. 109 W. Pgreater 2 Using this information, 6 (4 )2calculate the received powEquation you can (100) er at any distance using the following formula: Equation 9
d PR (d) = PR9d 0 ) 0 2 . (6) Equation ( d d PR (d) = PR (d 0 ) 0 . 100 d The referencePdistance = PRpractical systems operating at 1 , R (1200) for (100) + 20 log to 2 GHz is 1m for indoor environments and 1200 for outdoor 100m 100 Equation) = PR (100) + 20 log PR (1200 7 , environments. 1200 Equation 7 are either in decibels referred to 13 Most RF power-level units Equation
2

PR (1200) = PR (100) + 20 l mW)5263.6 dBm. (10) PR(100)510log(0.4431026og ,

Using Equation 7 to obtain the received power at 1200m yields: Equation 13

100 PR (1200) = PR (100) + 20 log , 1200

Equation 9

(11)

and

PL (d) = PL (d 0 ) dBm221.58 + X , PR(1200)5263.6 + 10 n log10 ddB5285 dBm.

0

Equation 13

(12)

Using Equation 5, you can verify the same value of received power. Thus, for an ideal, unobstructed-outdoor-line-of-sight envi d ronment, the PL (d) = PL power 10 nalog10 d distance when the received (d 0 ) + at 1200m + X , transmit power is 8 dBm is approximately0285 dBm. The actual received power will be lower because the real-world environment will likely have obstructions in the line-of-sight path or, worse, no line-of-sight path at all. For the previous example, you calculate the path loss as PT2PR. Therefore, path loss is 8 dBm2(285 dBm)593 dB. Practical Path-loss formulas For any practical wireless-sensor system, its important to know the maximum reliable data-transmission range. This wireless-system range directly depends on the link-budget parameter: LB5PT1GT1GR2RS, (13) where LB is the link budget in decibels, PT is the transmitted power in decibels referred to milliwatts or watts, GT is the transmitter-antenna gain in decibels, GR is the receiver-antenna gain in decibels, and RS is the receiver sensitivity. Sensitivity is the minimum RF signal that the system can detect with an acceptable SNR (signal-to-noise ratio). Equation 14 shows the receiver sensitivity: S52174 dBm/Hz1NF110logB1SNRMIN, (14)

milliwatts or decibels referred to watts rather than absolute Equation 13 power levels. You can rearrange Equation d 0 as: 6

PR (d) = PR (d 0 ) + 20 log10 . d d PR (d) = PR (d 0 ) + 20 log10 0d . (7) PL (d) = PL (d 0 ) + 10 n log10 d + X , d0 d The following example explains these concepts. Assuming PL (d) = P (d ) + 10 n log + X , a transmitting frequencyL8 0 MHz, 10 d 0 of 900 the transmit power of Equation

6.3 mW (8 dBm), and the unity-gain transmitting and reEquation 8 ceiving antennas, determine the received power at 1200m distance in an outdoor-line-of-sight environment. For an outdoor environment, the reference distance is 100m, and 0.0063(1power )2 you must determine the=received )(1)(1 / 3at 100m. 10waveThe 9 W. PR (100) = 0.44 2 length at 900 MHz is 0.33m.(4 )1)(1)(12 3)2 0.0063( (100) / 9
PR (100) = (4 ) (100)
2 2

= 0.44 10

W.

68 EDN | may 24, 2007 Equation 9

Equation 9

Equation 3

PL = 10 log

PT 2 = 0 log 1 . PR (4 )2 d2 Table 3 paTh-loss exponenT and sTandard deviaTion in differenT buildings

frequency (Mhz) 914 914 1500 900 1900 1300 900 1300 path-loss exponent, n 2.2 1.8 3 2.4 2.6 2 3 3.3 standard deviation (dB) 8.7 5.2 7 9.6 14.1 3 7 6.8

where 2174 dBm/Hz is the thermal noise floor, NF is the overall-receiver-noise figure in decibels, B is the overall receiver bandwidth, and SNRMIN is the minimum SNR. If the Equation total path loss between6the transmitter and the intended receiver is greater than the link budget, loss of data ensues, and communications cannot take place. Therefore, its important 2 for designers developing end systems to accurately characterd PR and=compare it0with the link budget to obtain (d) PR (d 0 ) . ize the path loss d initial estimations of the range. Path loss in indoor channels Equation MORE indoor-radio channel differs from the outdoor chanThe AT EDN.COM 7 nel because the indoor channel has shorter distances to cover, higher path-loss variability, and, MORE AT EDN.COM thus, greater d variance in the rePR (d) = PR (d 0 ) + 20 log10 0 . ceived-signalpower. However, varid + Go to www.edn. ability in the received-signal power com/ms4238 and is negligible for stationary wireless click on Feedback devices. Building layout, type, and Loop to post a comEquation 8 construction materials strongly affect ment on this article. indoor propagation. Research classifies indoor channels as either line-ofsight or obstructed channels with varying degrees of clutter (Reference 1). A buildings internal and external structures 0.0063(1)(and 3)2 have a wide variety of partitions 1)(1 /obstacles. Partitions dePR (100) = = 0.44 109 W. pend on whether the structure)2 (100)2 or an office environ(4 is a home ment. Partitions in a buildings structure are hard partitions, and partitions that can move and do not span to the ceiling are soft partitions. Houses typically use wood-frame partitions, Equation 9 whereas office buildings use soft partitions with metal-reinforced concrete between floors. Partitions vary widely in their physical 100 electrical charand PR (1200difficult to apply generic models for in, acteristics, making it ) = PR (100) + 20 log 1200 door channels. However, extensive investigations tabulate signal losses for common material types (Table 1). Floor-atEquation 13 tenuation factors represent the partition loss between floors (Table 2). Equation 15 shows the practical path-loss model for indoor channels using the log-distance path-loss model:
d PL (d) = PL (d 0 ) + 10 n log10 + X , d0

building retail store grocery store office, hard partition office, soft partition office, soft partition Factory, line of sight Suburban, indoor street Factory, obstructed path

PL(d)520log10(fMHz)110nSFlog10(d)2281FAF,

(17)

where nSF represents the path-loss exponent value for the same floor measurement and FAF is the floor-attenuation factor (Table 3). You can determine the FAF value from (Table 2). The following examples demonstrate how to use the foregoing tables and equations: For example, calculate the path loss for an outdoor free-space environment at a distance of 1200m at 915 MHz and 2.4 GHz. Using 20log10(fMHz)120log10(d)228, you can deduce PL at: 915 MHz520log10(915)120log10(1200) 228592.8 dB, (19) and PL at: 2400 MHz520log10(2400)120log10(1200) 2285101.2 dB. (20) (18)

(15)

where X is a zero-mean gaussian random variable in decibels and s is standard deviation. If the devices are stationary, you can ignore the effects of Xs. Calculating the value of path loss at a distance of 1m using Equation 4 and plugging it into Equation 15 results in: PL(d)520log10(fMHz)110nlog10(d)2281Xs. (16)

The value of n does not vary much with frequency and depends on the surroundings and the building type (Table 3). An in-building propagation model includes the effect of building type as well as obstructions. This model provides flexibility and can reduce the standard deviation between measured and predicted path loss to approximately 4 dB compared with 13 dB when you use only a log-distance model. Equation 17 represents the attenuation-factor model:
70 EDN | may 24, 2007

Propagation at a higher frequency results in higher path losses, which results in the reduction of wireless-transmission ranges at higher frequencies: For example, wireless devices operating in the 2.4-GHz range suffer from an approximately 8.4-dB reduction in path loss compared with a device operating at 915 MHz in an outdoor, free-space environment. In another example, using the information in Table 2, calculate the path loss for an indoor-office environment with hard partitions at a distance of 100m at 915 MHz and 2.4 GHz across the same floor and three floors. For the same floor, from Table 3, the average path loss is 3 dBm. Using this value of n53 in: 20log10(fMHz)110log10(d)2281Xs, yielding PL at: 915 MHz520log10(915)110(3)log(100) 2281Xs591.2 dB, (22) (21)

where s57 dB. And PL at: 2400 MHz520log10(2400)110(3)log (100)2281Xs599.6 dB, (23)

20log10(fMHz)110nlog10(d)2281Xs, you can deduce PL at: 915 MHz520log10(915)110(3)log10(100) 2281245115.2 dB, where s55.6 dB, and PL at:

(24)

where s514 dB. From Table 2, you can calculate the FAF for propagation for three floors as approximately 24 dB with a standard deviation of 5.6 dB. Using the information in

(25)

2400 MHz520log10(2400)1 10(3)log10(100)2281 245123.6 dB, (26) where s55.9 dB. In a third example, estimate the transmission range at 915 MHz for the above two examples assuming a system with unity-gain transmitting and receiving antennas, a transmitting power of 8 dBm, and a receiver sensitivity of 2100 dBm. The systems link budget is 82(2100)5108 dB. Its a good idea to have a link-budget margin of approximately 10 dB to account for the standard deviations in the path-loss formulas. Thus, the available link budget is 98 dB, which exceeds the path loss of 92.8 dB from the first example; therefore, you can consider 1200m to be outdoor range. In the indoor environment, path loss is 91.2 dB, and the available link budget is approximately 98 dB, assuming a 10-dB margin, which exceeds the path loss. Therefore, you can consider 100m to be the indoor range of that system.EDN au t h o r s b i o g r a P h y Shreharsha Rao is a systems engineer with Texas Instruments RFID group. His responsibilities include research into emerging architectures and technologies in the near-field-communication market and application development for low-power-wireless and RFID systems. He has a masters degree in electrical engineering from the University of TexasArlington. His personal interests include hiking, sports (hes a big fan of the Dallas Mavericks and Dallas Cowboys), and a poker league. r e fe r e nce
1 rappaport, Theodore S, Wireless Communications Principles and Practice, Prentice hall, 2001.

72 EDN | may 24, 2007