Bluetooth® wireless technology is a short-range communication system intended to provide connectivity of voice and data between information appliances. Bluetooth enables point-to-point and point-to-multipoint networking without the need for a formal wireless infrastructure. Two or more devices sharing the same wireless channel form an ad-hoc network or piconet. With one device acting as a master, up to seven other devices or slaves can be actively operating in the piconet.1 As Bluetooth wireless technology found its way into a variety of consumer products, applications requiring higher data rates have emerged such as streaming CD-quality audio, digital image transfer and printing. In addition, consumer demand for short-range wireless connectivity is moving beyond running a single application into a desire to run multiple applications simultaneously within the same piconet.

To meet consumer demand, Bluetooth wireless technology continues to evolve with recent improvements for higher data rates and longer battery life with the introduction of the enhanced data rate (EDR) mode. The Bluetooth EDR mode is a physical layer addition to the core specification2 that provides a two to three times increase in effective data rate over earlier versions while maintaining backwards compatibility. As a result of the higher data rates provided by EDR, the radio operates at a reduced duty cycle resulting in lower power consumption and longer battery life for the wireless appliance. The EDR mode also allows multiple applications to more effectively utilize the available bandwidth and achieve higher overall performance.

Bluetooth Evolution

Bluetooth systems, as originally defined in version 1.0 (v1.0) of the core specification, operate in the unlicensed industrial-scientific-medical (ISM) band at 2.4 GHz. Low power RF transmission provides communication between devices over a range of 10 to 100 m. The system uses a frequency hopped spread spectrum (FHSS) method for multiple access and interference mitigation. The hopping rate is nominally 1600 hops per second occurring over 79 channel frequencies in the ISM band.

The core specification was modified in version 1.2 (v1.2) with the introduction of adaptive frequency hopping (AFH). One reason for the change was the result of co-existence problems occurring between Bluetooth and 802.11b/g WLAN. Using AFH, the Bluetooth system is capable of measuring interference and avoiding those frequency channels that could result in lower system performance. The system can adjust its number of usable channels from 79 down to 20 as needed.3 Another major improvement implemented in v1.2 resulted in faster connection times. Improvements to the inquiry and paging operations provided connection times below 0.5 seconds compared to 4 to 5 seconds in devices compliant with v1.0. Additional improvements were obtained in the quality of the link by using data retransmission when errors occur and an improved flow control, with the introduction of new packet types, enhanced the functionality of v1.2 devices while maintaining backwards compatibility with v1.0 specifications.

The core system specification was recently updated to include higher data rate modes with the introduction of EDR in version 2.0 (v2.0+EDR).2 This latest specification has all the functional characteristics of v1.2 with the addition of two new modulation schemes implemented in the payload section of the packet. These EDR packet types provide peak data rates of 2 and 3 Mbps. The increase in the peak data rate beyond the basic rate of 1 Mbps is achieved by modulating the RF carrier using differential phase shift keying (DPSK) resulting in a two to three times increase in the number of bits per symbol transmitted. To maintain backward compatibility to v1.2 and provide simultaneous operation of v1.2 and v2.0+EDR radios in the same piconet, all devices use the same access code, header and frequency-hopping scheme.

Packet Structure and Modulation Format

The Bluetooth system uses a time division duplexing (TDD) scheme where the physical channel is sub-divided into time slots. The time slot length is a function of the frequency hop rate resulting in a nominal length of 625 ?sec. The data is transmitted between the master and slaves in packets that are contained within the time slots. All packets contain an access code, header and payload. The access code is used for synchronization, DC-offset compensation and identification of the packets in the physical channel. Access codes are also used in paging, inquiry and park operations in a Bluetooth system. The header contains link control information that includes the packet type. The payload contains voice and data, and may also contain error correction and control information depending on the type of packet transmitted. To maintain backwards compatibility to the earlier versions of the core specification, the access code and header information is modulated onto the RF carrier using Gaussian frequency shift keying (GFSK). The GFSK modulation scheme provides a 1 Mbps peak data rate by modulating one bit per symbol resulting in a symbol rate of 1 Ms/s. The data is modulated onto the RF carrier using a shift or deviation in the carrier frequency of a minimum of 115 kHz. The binary one is represented by a positive frequency deviation and the binary zero is represented by a negative frequency deviation. The modulation format in the payload portion of the general rate packet is also GFSK. The general rate packet is now referenced as the basic rate packet in v2.0+EDR to distinguish the 1 Mbps GFSK scheme from the two higher data rate EDR packet formats. The GFSK modulated signal uses Gaussian pulse shaping to provide spectral efficiency by maintaining a –20 dB bandwidth of 1 MHz.

For the higher data rate EDR modes, as specified in v2.0+EDR, the payload data is modulated onto the RF carrier using one of two DPSK modulation schemes. The EDR packet for 2 Mbps transmission, mandatory for an EDR device, uses a payload modulated with ?/4 rotated differential encoded quaternary phase shift keying (?/4-DQPSK). The optional 3 Mbps EDR packets use 8-ary differential encoded phase shift keying (8DPSK) modulation. As an example demonstrating the two different modulation formats in an EDR packet, Figure 1 shows a measurement of the amplitude versus time for an EDR waveform using GFSK modulation during the access code and header and an 8DPSK modulation during the payload. For this waveform, the packet length is approximately 450 msec, which is contained within the specified 625 ?sec time slot. Spectral efficiency in an EDR packet is achieved by using a root-raised cosine pulse shaping over the DPSK modulated portion of the waveform. This pulse shaping technique results in a –20 dB bandwidth of 1.5 MHz, which is larger than the bandwidth for the GFSK modulation format. The FCC has accepted the use of Bluetooth EDR radios in the 2.4 GHz ISM band by relaxing the –20 dB occupied bandwidth requirement from 1.0 to 1.5 MHz.

As a result of the modulation change in the EDR packet, additional timing and control information is required for synchronizing to the new modulation format. Following the header in an EDR packet is a short time period that allows the Bluetooth device time to prepare for the change in modulation to DPSK. This short time or guard time is specified to be between 4.75 and 5.25 msec. The guard time is followed by a synchronization sequence that contains one reference symbol and ten DPSK symbols. This sequence is required for synchronizing the symbol timing and phase for one of the two modulation types used in an EDR packet. Figure 2 shows a measurement of amplitude versus time for an EDR packet during the time when the modulation changes from GFSK to 8DPSK. It shows the 5 ?sec guard time and the eleven synchronization bits at the beginning of the EDR payload.

A differentially encoded phase modulated signal used in the EDR mode has the advantage that the signal can be demodulated without estimating the carrier phase. In this case, the received signal in any given symbol time is compared to the phase of the preceding symbol.4 As stated, the differentially encoded modulation format defined for 2 Mbps transmission is ?/4-DQPSK. The ?/4-DQPSK constellation can be viewed as the superposition of two QPSK constellations offset by 45° relative to each other. Symbol phases are alternately selected from one QPSK constellation to the other for each symbol time. As a result, successive symbols have a relative phase difference that is one of four angles ±?/4 and ±3?/4. The symbol transitions from one constellation to the other always guarantees that there is a phase change between symbols, making clock recovery easier.4 Figure 3 shows the ?/4-DQPSK constellation for the EDR portion of a packet. It shows a measurement over many symbols resulting in the eight desired constellation points. Note that, during any one-symbol time, only four constellation points or transitions are available resulting in the transmission of two bits per symbol. The figure shows the combination of two separate QPSK constellations separated by 45°, labeled A, B, C and D for one constellation, and 1, 2, 3 and 4 for the other.

The second EDR modulation format defined for 3 Mbps transmission is 8DPSK. The further increase in data rate is achieved through the addition of four more constellation points for each symbol. The total of eight constellation points allows a transmission of three bits per symbol resulting in a three times improvement in data rate over the GFSK modulation scheme. This type of modulation has many of the same benefits as ?/4-DQPSK, including the use of non-coherent demodulation schemes. Demodulation of an 8DPSK occurs by examining the relative phase difference between successive symbols resulting in phase angles of 0, ±?/4, ±?/2, ± 3?/4 and ?. As all eight constellation points or transitions are available between symbols, three data bits per symbol can be transmitted. The increase in data rate does not come without a penalty, as an 8DPSK modulated signal is more sensitive to noise due to smaller separation between constellation points when compared to ?/4-DQPSK signals.

EDR Test Procedures and Test Cases

With the introduction of EDR to the Bluetooth core specification, additional EDR-specific measurements have been added to the RF layer test procedure and specification (TSS/TP).5 The new measurements allow provisional testing of Bluetooth devices under non-loop back operation, which may be very useful during the early stages of radio development. The EDR tests specific to transmitters include relative transmit power, carrier frequency stability, modulation accuracy and differential phase encoding. The EDR tests specific to the Bluetooth receiver include sensitivity, bit error rate (BER) floor performance and maximum input level.

EDR Transmitter Test Cases

The EDR relative transmit power test verifies that the difference between the average transmit power during the GFSK modulation and the average transmit power during the DPSK modulation is within the specified range of +4 to –1 dB. The relative power is calculated from the difference of the average power measurement taken over at least 80 percent of the GFSK portion of the packet to the average power measurement taken over at least 80 percent of the DPSK portion. Figure 4 shows the relative transmit power measurement of an EDR signal using ?/4-DQPSK modulation with the RF carrier at the mid-band frequency of 2441 MHz. As shown, the average power measurements for the GFSK and ?/4-DQPSK waveforms are –14.4 and –16.22 dBm, respectively. The relative transmit power is calculated as +1.82 dB and is within the specified difference of –1 and +4 dB.

The EDR carrier frequency stability test begins with a determination of the initial center frequency error in the GFSK header. The frequency deviations in logic 1 bits and logic 0 bits are measured and reported as ??1 and ??2, respectively. The initial frequency error is calculated as the average frequency error between logic 1 and logic 0 bits and reported as the initial frequency error, ?I (?I = [??1+??2]/2). The initial frequency error is specified between ± 75 kHz. The frequency error in the EDR portion of the packet is corrected using this initial frequency error, wI. The corrected waveform is then partitioned into 200 blocks of 50 symbols in length. The remaining frequency error in each block is reported as ?0. The worst-case block frequency error, ?0, is specified to be within ±10 kHz. Lastly, the Bluetooth specification limits the maximum value of the combined frequency errors, ?I + ?0, to ±75 kHz. This value represents the maximum excursion of the frequency error, which includes the initial error in the access code and the frequency drift that may occur over the measured blocks. The given measurement example shows the frequency stability of an EDR waveform. In this case, the initial frequency stability is measured as –5.997 kHz, the block frequency error as –0.857 kHz and the combined frequency error as –6.854 kHz. All these measured values are shown to be within the required specifications.

The EDR modulation accuracy test verifies the quality of the differential modulation and is intended to highlight errors that would cause problems to a real differential receiver. The modulation accuracy is tested using a differential error vector magnitude (DEVM) measurement that is similar to the traditional error vector magnitude (EVM) measurement specified in other digital communication systems.6 The DEVM is defined as the magnitude of the error between two received signals spaced one symbol apart in time. The DEVM measurement is made over the synchronization sequence and payload of 200 blocks of 50 symbols in each block. The modulation accuracy is reported as three separate values, the 99 percent DEVM, RMS DEVM and Peak DEVM.1 The modulation accuracies, measured and reported in percentages for the measured EDR waveform using ?/4-DQPSK modulation, are 10.24, 11.57 and 5.5 percent, respectively. As shown, all measured DEVM values for this waveform are within the required specifications.

The differential phase encoding test verifies the operation of the differential PSK modulator used in the transmitter. For the EDR payload, the modulator is required to map correctly the binary data stream into a set of specified phase angles in the complex plane. The EDR payload is modulated with a PRBS9 sequence and a packet error rate measurement is performed over 100 packets. It is specified that 99 percent of the packets be received with no bit errors or, in other words, that the packet error rate is less than one percent. The packet error rate for the EDR waveform measured is zero percent or, in other words, no errors found.

EDR Receiver Test Cases

Bluetooth EDR receiver testing requires measuring the BER performance using test signals containing a variety of frequency and timing impairments. The BER performance measurements of all receivers are calculated over 16,000,000 bits by comparing the received data to the original PRBS9 sequence transmitted by the test source or test equipment.

The EDR sensitivity is measured using three groups of 20 packets corrupted with different timing errors and frequency offsets.5 The first group of packets contains no impairments. The second group of packets contains a carrier frequency offset of +65 kHz and a symbol timing error of +20 ppm. The third group of packets contains a carrier frequency offset of –65 kHz and a symbol timing error of –20 ppm. The receiver BER performance is required to be 10–4 under these conditions.

The EDR BER floor performance is a BER measurement at a received power level of –60 dBm. The BER is calculated by comparing the received data to a transmitted PRBS9 sequence. Under these conditions the BER performance is specified as 10–5.

The EDR maximum input level test shows the receiver BER performance when the input signal level is –20 dBm. This test shows the receiver performance under possible front-end compression when driven with a high input power level. The BER performance is specified as 10–3 using this input power level.

Conclusion

The need for higher data rates and improved power consumption required in portable, multi-media consumer appliances is expected to drive the transition to Bluetooth EDR technology. The EDR evolution will provide multi-use scenarios where numerous devices operate concurrently in the same piconet. In addition, new portable devices are anticipated which combine several wireless interfaces, such as GPRS and WiFi with Bluetooth EDR, in order to provide simultaneous and seamless connectivity across multiple network types.

Bluetooth and the Bluetooth logos are registered trademarks owned by Bluetooth SIG Inc., US, and licensed to Agilent Technologies Inc.

References

1. Bluetooth Enhanced Data Rate (EDR): The Wireless Evolution, Agilent Technologies Application Note 5989-4204EN, November 2005.

2. Specification of the Bluetooth System, Version 2.0+EDR, Specification Volume 1: Architecture & Terminology Overview and Volume 2: Core System Package, November 4, 2004.

3. “How 802.11b/g WLAN and Bluetooth Can Play Together,” Philips Semiconductors White Paper, Document 9397-750-13426, June 2004.

4. J.G. Proakis, Digital Communications, 3rd Edition, McGraw-Hill, New York, NY, 1995.

5. “Radio Frequency Test Suite Structure (TSS) and Test Purposes (TP) System Specification 1.2/2.0/2.0+EDR, Rev. 2.0.E.3, Document Number RFTS/ 2.0.E.3,” March 21, 2005.

6. Digital Modulation in Communications Systems — An Introduction, Agilent Technologies Application Note 1298, 5965-7160E, March 14, 2001.

Helen Mills received her MS degree in electronic and electrical engineering from the University of Glasgow, Scotland. She joined Hewlett Packard/Agilent Technologies Inc. in 1996. During her nine years at HP and Agilent, she has worked in a variety of product marketing and planning roles for signal generators, signal analyzers, power meters and wireless connectivity test equipment. She is currently a product manager with Agilent Technologies, Wireless Division for the N4010A wireless connectivity test set.