Fig 1. This diagram shows the test connections between the Microchip Technologies MRF24J40MB and Explorer 16 demonstration board ZigBee radio module/test board and the MDO.
Fig 2. In this capture, the orange bar represents the spectrum time of the frequency domain display relative to the time-domain measurements.
Fig 3. Shown are measurements of power and occupied bandwidth, with correlated RF amplitude versus time, and measurements of the power-supply current and drain voltage.
Fig 4. This capture shows a spectrum plot and measurements with resistance in series with the module power source to study low-power performance behavior.
Fig 5. Here, packet decode of SPI digital signals (SPI: master out/slave in and master in/slave out) are added to the display.
Fig 6. A subsequent acquisition triggered on an SPI command shows the delay between the command and the radio turn-on.
Fig 7. A wide scan of the 2.45-GHz spectrum enables a signal view of the entire industrial, scientific, and medical (ISM) band.
Fig 8. A wireless LAN interfering signal is shown to assess its impact during interoperability testing.
Driven by ongoing concerns over global warming and escalating energy costs, the demand for smart, radio-controlled devices to monitor, control, inform, and automate the delivery of energy and other resources continues to grow at a rapid pace. And that’s just one of the many applications for technologies based on the IEEE 802.15.4 physical-layer (PHY) radio standard.
While several competing protocols are based on this PHY, the world leader at the moment is ZigBee, with published standards spanning everything from home automation and smart energy to retail and telecommunications services to remote control and input devices. It provides a mesh network of devices enabling communication across large areas and hundreds or even thousands of devices. Assuming consistent implementation, ZigBee-compliant devices from different sources and vendors can seamlessly communicate with each other.
A vibrant industry has developed around these standards at the level of bare ICs and modules that typically come complete with the antenna and Federal Communications Commission (FCC) or other regional agency approvals. Embedded products are available with just the radio circuitry with the IEEE 802.15.4 lower-level protocol while requiring a separate microcontroller or microprocessor to handle the ZigBee software as well as the application.
Also, there are ICs as well as modules that have a microcontroller built in to run the ZigBee or other protocol software. Many of these ICs and modules have uncommitted I/O pins, so a complete product may need little more than the module and sensors and/or actuators and an enclosure. In addition, modules are available with power amplifiers (PAs) and receiver low-noise preamplifiers (LNAs). The PA and LNA can substantially increase the radios’ range, though at higher cost and power consumption.
For any of these choices, designers need a printed-circuit test board to support the IC or module. Designers also need a power supply with sufficient peak power and freedom from noise. If a chip-level radio is selected, the appropriate antenna interface circuits will also be needed.
As ZigBee becomes more commonplace in all types of embedded systems and applications, engineers need the ability to quickly and efficiently validate and verify ZigBee module performance. This system-level task becomes more complex with the presence of RF and the need to look at analog, digital, and RF signals in concert with each other. The mixed-domain oscilloscope (MDO), which incorporates a spectrum analyzer, reduces the efforts involved in performing ZigBee testing.
ZigBee Design Tradeoffs
Reflecting the variety of end-use applications and thousands of products that can be adapted to use ZigBee technologies, there is no such thing as “one size fits all” in the ZigBee world. ZigBee radio options from different vendors reveal different levels of integration, from bare-bones radio ICs to fully integrated modules with a microcontroller, PA, antenna, and LNA. Given this diversity, designers need to understand the tradeoffs involved. Key areas to consider include:
- Cost: There is a major tradeoff in material cost versus the cost of engineering and regulatory approval for modules compared to ICs. Modules cost substantially more than the radio ICs with their support components and assembly labor, even in large quantities. Part of the extra cost lies in the duplicated printed-circuit board (PCB) material, but most of it is in offsetting the engineering cost of the module and in providing a return to the module manufacturer. However, engineering the radio circuitry and gaining necessary approvals has substantial cost. For IC-based designs, ZigBee Alliance testing and approval adds to the cost. Experience suggests that the cost break-even between integrating ICs versus modules is typically around 10,000 to 25,000 units.
- Development time: Pre-certified modules can be marketed as soon as the product is ready. Regulatory agency approvals for IC-level designs can take as little as a month, but often much longer. Generally, this time is added to the development process because the product needs to be in close-to-final form. The software also needs to be functional before the approval testing can begin.
- Form factor: Designing a custom radio from the IC level provides flexibility in the configuration of the radio circuitry. With a custom design, the radio can use spaces that no module can fit into given the overall configuration of the product. Generally, the available modules have all of their parts on one side of the PCB so the module can be soldered to the main board. In a custom design, parts can be placed in any configuration and on both sides of the board.
- Protocol flexibility: Many manufacturers of modules and ICs with an embedded microcontroller do not provide access to the source code of the ZigBee or other communications software. This greatly limits your ability to add custom features.
- Special requirements: For some applications, there may be a need for hardware capabilities beyond what is available in modules or ICs that have an integrated radio and microcontroller. While adding a second microcontroller is always an option, the total cost may rise unacceptably. In other cases, one may wish to provide capabilities not commercially available. For example, U.S. regulations allow up to 1 W of radio output power, but there are few, if any, modules with this capability.
- Antenna type and placement: Modules are available with antennas on the PCB either as a printed pattern or as a “chip” antenna with an external antenna. An antenna on the module can have impaired performance if the antenna is inside a shielding enclosure or if it is located too close to other components in the end package design. Modules with connectors for external antennas are available. However, it is only legal to use antennas that have been certified with the module. If there is a reason to use an antenna not supported by the module vendor, such as a need for higher gain, agency approval with its accompanying cost and time are required.
Test Validation Of The Integrated Radio
Once the approach to the radio implementation is determined, the appropriate PCB laid out, and any necessary software written, several tests can be performed to ensure good communications.
For most applications, there will be serial communications between the radio system and other parts of the product. For example, many ICs and modules use a four-wire serial peripheral interface (SPI) connection to control the radio IC and any related components such as a PA. SPI commands set internal registers for the selection of the frequency channel, the output power level, and many other operating parameters. SPI controls general-purpose port pins to control the PA or other devices. In addition, SPI sends the data packet to the IC or module as well as the command to transmit the packet. Received data returns through the SPI bus as well.
Software in the microcontroller (whether integrated or separate) needs to provide the higher levels of the protocol (ZigBee or other) as well as control the power to the radio and run other aspects of the product. In many applications, the timing of the radio transmission is critical so the radio is not transmitting while some other power-consuming part of the product is running and draining the power-supply voltage below acceptable levels.
Some of the critical tests to verify radio operation include RF and power-supply measurement, digital commands, spurious signals, and interference. To illustrate these tests, we paired a Microchip Technologies IEEE 802.15.4 amplified radio module (MRF24J40MB) with an Explorer 16 demonstration board.
The screen shots are from a Tektronix MDO4000 series MDO, the world’s first oscilloscope to provide simultaneous time-correlated views of RF, analog, and digital signals. Setup and data commands are sent from a PC to allow manual control (Fig. 1). A direct connection to the radio facilitates power and other measurements. A calibrated antenna also could have been used to take the RF measurements.
RF And Power-Supply Measurements
The channel spacing for IEEE 802.15.4 (including ZigBee) is 5 MHz. The 20-dB channel bandwidth should be significantly less than the channel spacing. The measured occupied bandwidth of 2.3 MHz is well within the specification (Fig. 2). Output power should be in the range of 20 dBm. The screen shows the output spectrum in its lower part and direct measurements of bandwidth and power. The test cable drop is about 2 dB in this frequency range, so the power measurement is in the range of what is expected.
The orange bar at the bottom of the top half of the screen indicates the period in which the spectrum trace is displayed. The spectrum time is the window-shaping factor divided by the resolution bandwidth. In this example, using the default Kaiser fast Fourier transform (FFT) function (shaping factor 2.23) and the resolution bandwidth (RBW) of 11 kHz, the spectrum time is approximately 200 µs. Moving the spectrum bar across the time-domain window allows the spectrum and measurements to be taken at any time during the packet transmission. This acquisition correlates just after the radio packet transmission is turned on.
The MDO’s RF acquisition performs power and occupied bandwidth measurements on the RF signal. Because it also acquires a time record of the RF acquisition, a digital down-conversion process produces the I (real) and Q (imaginary) data. Each I and Q data sample represents the instantaneous deviation of the RF input from the current center frequency. With this analysis, the RF amplitude versus time is computed from the recorded data.
An illustration shows the added trace of the RF amplitude versus time (Fig. 3) added to the display of Figure 4. This demonstrates that the events of the current and voltage measurements shown in Figure 5 correlate with the turn-on of the RF transmission.
The green trace (Fig. 3, again) shows the current drawn by the module. During packet transmission, the current draw is almost 200 mA. (Note the direct measurement of 174 mA.) So, the power supply must be designed to support this load. The yellow trace (Trace 1) shows the effect of this current draw on the supply voltage. The drop is only about 70 mV, which should be fine. (Note the direct peak-to-peak measurement of 72 mV.)
The orange trace (Trace A) in the upper part of the screen shows the RF signal amplitude versus time. The input current rises in two steps. In the first step, the radio IC turns on. There is then a delay to allow the frequency synthesizer to stabilize before turning on the power amplifier. The rise of RF power coincides with the second part of the current step. The turn-on period appears to be approximately 100 µs.
It is often necessary to understand the performance of radio transmitters during low-battery conditions, or conditions when the power supply becomes current-limited, to understand the margins of radio-compliant performance. For example, designers can place a 1.5-? resistor in series with the module to simulate the effect of a depleted battery (Fig. 4).
The current drawn by the module is only a few milliamps lower, but the voltage drop is about 230 mV. The output power drops by 1 dB as measured by the RF power measurement, and there is a slight increase in the adjacent channel noise in the spectrum display. The lower output is evident in Trace A, which is amplitude versus time (Fig. 4, again).
Digital Commands
Radio ICs and modules need to be set up to meet the operating requirements of the specific application and any protocol-specific setups. The MDO allows decode of the SPI commands to the ZigBee module (Fig. 5). Figure 5 shows the digital capture of the SPI commands in the same time frame as in Figure 2. Decode is enabled, but it isn’t readable in this time scale.
In this case, the analog, digital, and RF acquisitions have been set to trigger on the drain current of Trace 4 occurring above the 130-mA level. All time-domain measurements in the upper display (left of center) show the events that occur prior to the current exceeding this level at RF turning on. This includes digital decode, analog (voltage and current), and RF versus time. From this information, you can see that a digital command occurs roughly 600 µs before the RF event turn-on.
The traces in purple show where the decoded data is in the time domain. Pan and zoom functions can be used to read digital waveforms and the decoded data. Designers also may read or trigger other commands on SPI (master in/slave out, or MISO) to confirm correctness and to verify radio operation.
The MDO architecture simplifies measurements between SPI command triggering and correlated RF events. Let’s suppose the trigger event is now changed to the SPI command \{37\}, the radio transmit trigger command (Fig. 6). Markers on the time-domain display show the SPI command to current draw (at the beginning of the RF Tx turn-on) is now 1.768 ms.
In the previous example from Figure 5, the command delay to turn-on was about 600 µs. The actual event in Figure 6 is almost three times longer. This demonstrates that the behavior of the ZigBee radio is complying with one of the PHY layer performance requirements of IEEE 802.15.4. The ZigBee radio uses a pseudorandom delay between the command and turn-on event to enable the radio to listen for other ZigBee radio transmitters or other radio interference channels.
Spurious Signals
In confirming a radio’s operation, it is critical to ensure that no spurious signals are causing interference. In the example shown, no significant spurious signals are present in the band in which ZigBee operates (Fig. 7). Note that the module is set to transmit in the center of the 2.45-GHz band. Here, the marker function is used to measure the peak signal. With the resolution bandwidth now set to 100 kHz, the spectrum time is now reduced to just over 20 µs.
It is also important to look for signals in other parts of the spectrum. For example, a next step would be to look at the frequency range of the second harmonic of the transmitted signal while it is still correlated to the triggered level of the current draw during the RF transmission turn-on. In this case, we found only a small signal at the second harmonic and nothing significant at other frequencies. The second harmonic signal at the marker is about 35 dB lower than the fundamental, which is well within the FCC rules applicable to this type of radio transmitter.
Interference
For certain applications, it is useful to take measurements with an antenna to identify other radio sources that might interfere with the radio being developed. The example shows a reference antenna used with the MDO to look for possible interfering radio sources (Fig. 8).
The wideband signal centered at approximately 2.46 MHz is a Wi-Fi basestation in the same building. This covers a number of channels that the ZigBee radio could use. In an application for this radio module, it would be wise to avoid using the channels around this frequency since the range of the ZigBee radio would be impaired or the radio blocked completely.
In this case, only the spectrum-analyzer portion of the MDO is used along with RF triggers to capture any signal in the band of interest. The main reference marker shows that this is a rather strong signal. The manual markers (a) and (b) show the range of frequencies of the interfering source.
The frequency range and power of this interference would make ZigBee channels 17 to 19 unusable. Of course, most protocols, including ZigBee, will scan for interference like this and move operation to a clear channel. Less sophisticated protocols may need manual adjustment of the operating channel.
Summary
There are many options to consider before implementing ZigBee or other IEEE 802.15.4 radios. The selection of the best approach depends on many factors including development time, the unit cost versus engineering, and approval cost, as well as other requirements such as space available, form factor, and special electrical requirements for the radio.
Regardless of the approach selected, several measurements are needed to ensure that the radio system is working correctly. RF measurements include checking the RF output frequency, output amplitude, occupied bandwidth, and spurious outputs. Confirmation of packet timing, current consumption, and any power-supply noise is important as well. In addition, it is valuable to confirm that the correct digital configuration information is being sent to the radio and correct data is being received. As demonstrated, the MDO, which can time-correlate analog, digital, and RF signals, is well suited to this task and can save the designer time and effort in validating and verifying ZigBee modules.