ATE facilitates WiMAX RF test and characterization
For proper WiMAX transceiver device testing, an RF semiconductor tester needs to test accurately and quickly and also assist in device characterization.
By Ron Waltman, Analog Devices, and Peter Higgins, Teradyne -- Test & Measurement World, 12/1/2009 2:00:00 AM
WiMAX transceiver devices have proven to be a boon to the consumer electronics market, where they have found multiple uses including connecting a WiFi hotspot to the Internet. To ensure the devices work as promised and to bring them to market quickly, device manufactures need sophisticated, multifunctional test equipment and equally sophisticated test software.
WiMAX capabilities
WiMAX is an RF technology developed to provide “last mile” broadband access as an alternative to wired DSL or cable. Based on the IEEE 802.16 standard, WiMAX technology has a range of a few kilometers, compared to the tens or hundreds of meters that WiFi (IEEE 802.11) offers.
Popularly adopted WiMAX carrier frequencies include 2.3 GHz, 2.5 GHz, and 3.5 GHz with channel bandwidths of 3.5 MHz, 5 MHz, 7 MHz, and 10 MHz. Like other digital modulation schemes, WiMAX provides longer transmission paths using a simpler modulation scheme with reduced data rates. If the path length is short, complex modulation schemes provide high data rates with low bit-error rates. To achieve high data transmission rates, WiMAX devices use numerous channels of MIMO (multiple input, multiple output) technology.
The “sister” version of WiMAX is WiBro (Wireless Broadband), developed by the South Korean telecom industry. Also known as mobile WiMAX and incorporated into IEEE 802.16e, the technology has a slightly different frequency band allocation around 2.3 GHz.
WiMAX uses OFDM (orthogonal frequency division multiplexing), a multiplexing technique that subdivides the bandwidth into multiple frequency subcarriers. In an OFDM system, the input data stream is divided into several parallel substreams of reduced data rate, and each substream is modulated and transmitted on a separate orthogonal subcarrier. In a 10-MHz channel bandwidth, data rates up to 63 Mbps are possible on the downlink between the base station and mobile unit, and rates up to 28 Mbps are possible on the uplink (Figure 1).
 Figure 1. WiMAX modulation schemes include quadrature phase-shift keying and 16-quadrature phase-shift amplitude modulation. |
In early mobile devices, in-phase (I) and quadrature (Q) information would pass from the baseband processor to the RF portion of the device in analog format. In today’s highly integrated devices, the ADCs and DACs reside in the same package as the RF circuitry, making the link between the RF device and the digital baseband processor a digital data bus. Moving the ADCs and DACs out of the baseband processor and into the RF device allows the processor to be fabricated on the smallest lithography possible, which reduces BOM (bill of materials) costs. Figure 2 shows the layout of a typical RF MIMO transceiver with digital interfaces and multiple RF ports.
 Figure 2. This WiMAX 2x2 MIMO transceiver block diagram illustrates a typical RF MIMO transceiver with digital interfaces and multiple RF ports. |
Requirements for WiMAX test systems
To test WiMAX transceivers on a high-throughput manufacturing line, an ATE (automated test equipment) system needs these key capabilities:
-
digital sourcing and capturing at the same speed as the DUT (device under test),
-
a low-phase-noise clock to provide a reference for the synthesizer,
-
clean power supplies and ancillary control circuitry for relay control,
-
RF sourcing and capturing,
-
multiple RF ports that can be easily calibrated for accurate signal level, and
-
methods of sourcing and capturing WiMAX modulated signals.
The ATE system also needs to have enough resources—both hardware and software—to perform multisite testing with a high degree of parallel capability. Using parallel testing, the system should be able to test several devices in an overall test time that is close to what a single-site system would need to test one device.
During test development, you should align the tester resources in a way that minimizes the complexity of the load board. This allows the calibration of the RF signal levels to the tester delivery plane to be automatic from the test engineer's perspective. Just as devices should be designed to minimize the number of components on the PCB of the final assembled product, and thus lower the BOM cost, ATE load boards should also have as few components as possible. A “clean” load board with minimal components requires less time to design, lay out, build, and debug, and also proves to be more reliable in a volume production situation.
To test MIMO devices, the tester needs to provide multiple receivers to capture the devices’ TX signals in parallel. It passes the captured waveforms to a modulation-analysis package that can interface with multiple input streams and analyze the combined information. The same process applies to the receive path, where multiple digital capture engines need to concurrently capture the digital data output of the devices’ receivers.
A 2x2 MIMO device has two input RX and two output TX RF ports. For such a device to be tested in a quad-site arrangement, the tester must provide eight RF source and eight RF capture channels. To avoid splitters or RF switches on the DIB (device interface board), the ATE needs to provide 16 RF ports.
A quad-site application requires four high-purity reference clock inputs, one for each DUT’s synthesizer. Low-phase noise of the clock inputs is paramount as clock phase noise can affect a device’s performance. A tester with good sources removes the need for dividers or crystals on the DIB. Crystals have good phase noise, but they are not frequency locked to the ATE, so they can cause digital synchronization issues. Thus, you will get better test results if your tester does not require them.
Test challenges for WiMAX devices
WiMAX devices must undergo a suite of tests to ensure they will operate properly when used in a radio. This range of tests typically includes:
-
continuity and leakage tests to ensure correct packaging and electrostatic discharge protection,
-
digital pattern tests (including some in scan format),
-
traditional INL (integral nonlinearity), DNL (differential nonlinearity), and THD (total harmonic distortion) performance measurements of the converters,
-
power-supply consumption measurements for the various operating modes of the DUT, and
-
RF receive and transmit operations to test specifications for both sine-wave and modulated signals.
A “loose functional” test may be performed first to determine if the device is alive and if further testing is justified. This step may not be time-efficient, though, depending upon yields and test methodologies.
You may not need to perform the tests in the order shown above, because some testers can perform DSP functions in the background while simultaneously performing other tests, such as ones requiring large digital patterns. Testers that offer such parallel testing can help you optimize the overall test time.
As devices become more complex, it becomes paramount for designers to include design-for-test features in them. For example, a test bus with test modes designed into the DUT assists in the routing of signals to observation points that are not used in normal operation. This visibility assists the test engineer in accurately testing a block or section of the DUT.
Classical CW (continuous-wave) tests on the transmit RF portion of a WiMAX transceiver include output power, carrier, and sideband suppression measurements. You can also perform a transmit test to measure the LO (local-oscillator) suppression. While this is not a conventional transmit test, you should know how much LO leakage is present at the transmit pin, and is thus being radiated by the antenna. This level is low and the phase noise of the LO cannot be tested using classical CW techniques.
On the receive side, gain, gain linearity, image rejection, and IP3 (third-order intercept) are all key CW tests. RSSI (receive signal strength indicator) is another test you should consider. For RSSI, a device’s own indication of the receive level provides a good basic test of receive functionality. The RSSI test usually involves the reading of a register value, a step that can be very handy, especially during wafer probing, when a fully loaded RF tester is often not available and a subset of tests are performed.
RF modulation testing
|
Modulated testing looks at the device as it will be used in its final application. This offers the advantage of measuring the performance of the radio as a whole system.
A typical modulation test on the transmit side is EVM (error vector magnitude), also known as RCE (receiver constellation error). EVM measures how far the constellation points are from the ideal; the lower the EVM the better (Figure 3).
In a perfect scenario, the constellation points of a modulated signal would be in their ideal locations. But device imperfections resulting from the phase noise of the LO, nonlinearities, image rejection, and other issues cause the constellation points to be in nonideal locations, thus limiting the data rate.
Channel mask testing is another common modulation TX test, where a bandwidth wider than the channel is captured, and the signal level outside the working channel is measured to ensure it is low and inside specification.
For the receive path, tests often measure EVM and BER (bit error rate). The BER is the ratio of the erroneous bits to correct bits—the smaller the better. BER tests measure a modulated RF signal received by the DUT and count the number of the correctly and incorrectly received bits. BER testing has typically been time-consuming because it takes a long time to test very low BER levels.
RF modulation testing can also be used for filter testing. A multitone signal containing a component in the pass band and containing three to six other components in the roll-off and stop bands can be used to quickly determine the 3-dB point and stop-band performance of a device’s filters. This multitone signal technique can be used for both RX and TX filters, with a key advantage of only one capture being needed in either the digital or RF domain.
Modulation tests offer useful information about how the DUT will perform in a complete system. DUTs that fail these tests probably won’t work satisfactorily. Unfortunately, it is difficult in a production environment to pinpoint which block of the device caused the problem. Both CW testing and modulation testing should be considered necessary to screen out marginal RF performance.
Bench characterization assistance
RF devices are characterized on the development bench with lab equipment designed to simulate the operation of the device in final usage and to test the device to the relevant standards. This process is extensive and time consuming and involves a large amount of bench equipment.
Incorporating the same tools in the ATE opens opportunities for the lab and production staffs to work closely together and use the same industry- and RF-standard compliant waveforms and analysis methods. The lab staff will see a reduction in the number of hours they need to spend in the lab, and the production staff will obtain quicker responses to questions on device-setup conditions, register values, and so on.
Today’s lab and production staffs can share data more easily than their predecessors, because most newer ATE systems are PC-based and run a Windows operating system. These systems can run tests quickly, over many devices, and can quickly rerun tests with different supply rails if desired, and the test results can be automatically exported to a spreadsheet such as Excel. Engineers can then plot the test results in a graphical form (Figure 4), which allows convenient visual analysis and sharing with other team members and management.
|
Using the same analysis tools in the lab and in production greatly increases the likelihood of correlation between bench and ATE measurements, but skill is still required to deal with factors such as different DUT sockets used in the two locations. In addition, an ATE board will probably be much thicker than a laboratory evaluation board, and supply decoupling locations and RF signal delivery routes will also differ, requiring test engineering skill.
Nevertheless, having a common tool for modulation work can produce an integrated team working less in the lab and more on the tester and delivering engineering samples quicker and in greater volume. Having identical ATE and bench modulation debug displays and setup files is also of great help. The bottom line is the ability to deliver comprehensively tested WiMAX devices in a timely manner that meet customer expectations.
No related content found.
- 0 rated items found.
Datasheets.com Electronic Parts & Inventory Search
185 million searchable parts
- Part Number
- Description
- Inventory
- Products
- Manufacturers
View All Webcasts
