Oscilloscopes are not commonly associated with Bluetooth measurements. But a digital scope's long memory records and multiple screen views make it an ideal tool for analyzing the packetized bursts that comprise Bluetooth baseband data. Key oscilloscope characteristics for performing Bluetooth baseband analysis include adequate memory size, sufficient number of trace displays with zoom capability, and sufficient display resolution to reveal critical information about these relative slow, real-time baseband signals.

By configuring an oscilloscope for single-time capture and acquiring a long memory record, an initial zoom trace can be used to navigate through packets in the serial data stream. At the same time, a second zoom trace can be used to isolate individual packets in order to identify the access code, header, and payload components of the Bluetooth packet structure. A third zoom trace can be used to identify specific bit values comprising the sync word, preamble, and trailer within the access code. A fourth zoom allows close inspection of physical level characteristics of isolated bits. Since the preamble is a predefined sequence, the nominal bit size can be determined by inspection of the sync word pulse width. Once this is known, the entire bit stream encompassing the access code, header, and payload can be identified on a bit-by-bit basis. Simultaneous views using a single time capture allow system-level characterization of Bluetooth baseband signals while isolating any edge of any bit within any packet in the serial data stream (Fig. 1).

Captured at a slow sweep rate, the high-level interactions between transmit and receive signal lines can reveal initial verification of communication between Bluetooth devices. Transmitted data packets, poll transmissions, and null transmissions can be monitored on the transmit line, while received data packets, responses to polling, and nonresponses to null transmissions can be monitored on the receive line. An oscilloscope's ability to create a surface map of incoming waveforms allows a continuous display of the time-varying data which is useful for this verification. The amplitude of the signal is denoted by its color, where the largest amplitudes are represented in red. As the scope continuously acquires waveforms, an outline of the packet shape, and boundaries between the header and data are formed. Areas of constant waveform activity in the access code are differentiated from the dynamic variations of the payload, and the variance of logic values in the payload from one packet to the next is depicted. Note that the receive window is a different size from the transmit window as the receiver (Rx) side becomes active in preparation for possible early arrival of a packet. For a packet that arrives with nominal arrival time, we would expect to see some random bits prior to the arrival time of a packet. Surface mapping the serial bit stream allows a useful way to view signal variations in the Bluetooth baseband, identify payload and access code boundaries, and measure the transmit and receive window size. In addition, oscilloscope surface maps (Fig. 2) can be used to monitor the Universal Asynchronous Receiver-Transmitter (UART) interface and to verify that the Synchronous Connection Oriented (SCO) link between Bluetooth applications is functioning.

RF test equipment such as spectrum analyzers and vector signal analyzers is useful for making Bluetooth spectrogram measurements and for testing phenomena such as carrier frequency drift, burst profile, and frequency-modulation (FM) characteristics. An oscilloscope is able to provide unique time-domain analysis of packet data not available on other instruments such as time correlation showing cause and effect between events, measurements based on time differences between edges, and analysis of rare intermittent events.

Using a complex trigger and zoomed view of a single data packet, an oscilloscope can automatically detect the data's frequency and set an internal reference used to compute instantaneous phase changes throughout the acquisition for each cycle in the waveform. The resultant plot of time interval error reveals underlying modulation characteristics of the measured symbol timing of a packet and the jitter of the derived clock from the 1-MHz data. The resultant time interval error waveform is used to view the characteristics of the symbol timing variation. Maximum, minimum, and standard deviation values scaled to unit intervals can be computed and read directly from the oscilloscope display (Fig. 3).

The master transmitter (Tx) timing test, which is a measure of clock drift, is a difficult-to-perform Bluetooth test. The basis for the measurement is to determine that the master device will keep an exact timing interval while the piconet is active. The master is transmitting to a second Bluetooth device in the connected state, and the timing from the start of a master packet to the start of another master packet 5000 slots (6.25 s) later is measured to calculate clock drift.

The pass verdict for measured timing drift of the device under test over 5000 slots is less than or equal to 125 µs. Channel 1 is a time capture of Tx data, and Channel 2 is the Tx-on debug signal, used to signify the start-of-packet timing. The baseband transmits every other slot, with 1.25 ms per frame. Allowable tolerable clock drift is 20 PPM, which for 6.25 s is 125 µs. Capturing 10 s of continuous data using long memory and double zoom isolates two preamble pulses 5000 slots apart with a delta measurement of 6.25006079 s. This 60.79 µs measurement of drift confirms passing the master Tx timing test within the Bluetooth standard specification (Fig. 4).

Another difficult-to-make Bluetooth measurement is the computation of packet-error rate, which is a ratio of missed to transmitted packets. TXON is the sync for the master transmit and this signal envelopes the transmit packet. PKDET is the packet-detect signal of the slave and produces an edge whenever a packet is received. If there is a case of a rising edge of the master Tx sync not followed by a rising edge on the packet detect, the slave has missed a packet. The packet-error rate is the ratio of (the number of TXON edges minus the number of PKDET edges) divided by (the number of TXON edges).

Within the same time period, collecting as many edges of Tx sync as possible and using the oscilloscope to count rising edges, the ratio of missed versus sent will determine packet-error rate of the Rx. An oscilloscope's capability of counting waveform edges greatly simplifies this difficult measurement (Fig. 5).

The Bluetooth baseband device can generate an access correlation signal to determine if the access code of the received packet matches the expected bit values. We want to verify that if an access code does not correlate in the baseband it is because there were errors in the received data arriving from the radio. Using Interval trigger to synchronize the scope capture to a missed packet, simultaneous zooms onto a correct packet access code and invalid packet access code quickly shows which bits in the access code resulted in the access code correlation error and missed packet (Fig. 6).

As has been seen, an oscilloscope can be a useful tool for analyzing Bluetooth baseband signals. As was demonstrated, the instrument can perform system-level analysis, transmission and receipt verification, master Tx timing, packet-error rate, identification of a missed packets, and the computation of time interval error. Today's digitizing oscilloscopes can provide powerful system-level analysis capabilities for debugging Bluetooth baseband systems.