Pre-compliance testing for WLAN transmitters
22 July 2014
Using a wireless module lowers design complexity, but still involves a number of important steps.
It’s no longer enough to have smart gadgets. To grab consumer attention a gadget has to be both smart and connected. This means wireless must be part of the design as well. Wireless also opens up many new design possibilities and is starting to show up in some unusual places, says Xiao Li, Tektronix.
Some of the more interesting examples are Velo labs and from Oral-B. Start-up Velo Labs is developing a solar-powered bike lock that can hook up to WiFi networks to send alerts to the bicycle owner’s smartphone if the motion detector senses that the bicycle is being moved. The other is an intelligent electric toothbrush from Oral-B. It incorporates wireless to capture information about the user’s brushing habits – presumably it could become part of dental records.
There are literally thousands of products on the shelves, in development, or yet to be imagined that will incorporate low-power wireless capability to meet consumer demand or to become part of the Internet of Things.
One challenge is that product manufacturers – many of whom have little to no RF experience – need to learn how to add wireless capability to products. The most common and practical approach is to simply incorporate a pre-packaged WLAN module into the design. Not surprisingly, the market for such modules is growing at a strong double digit pace, with continued growth forecast.
While such modules eliminate many technical issues (see Figure 1), there are still decisions to be made. The most critical is to ensure that the end product meets FCC and international regulatory requirements. Compliance testing is exhaustive and time consuming, and a failure at this stage can cause expensive re-design and delay product introduction.
Regulatory pre-compliance checks (Step 6 in Figure 1) can avoid such worse-case scenarios. Fortunately, cost-effective and familiar test equipment can be used in-house to perform pre-compliance testing and ensure that wireless-enabled products have a high probability of passing first time.
From the very first wireless transmissions, spectrum emission has been a concern for design engineers. Regulatory agencies around the world have placed limits on the emission levels, and have defined measurement methods for compliance testing. Formal certification, which must be completed before a product can be sold, must be done at an independent lab and can cost between $5,000 to $10,000 per day (not including travel and other expenses).
The use of off-the-shelf modules, even ones that have been certified, does not necessarily make the job of certification much easier, as the complete assembly must be tested and qualified as well. Design issues like PC board layout, antenna design and placement or system interactions can lead to failure.
Pre-compliance testing vs full compliance
Pre-compliance testing is done after system integration, to determine any problem areas in the design. Pre-compliance testing does not necessarily need to map to every international standard, since the goal is to simply uncover potential problems and reduce risk of failure at the expensive compliance test stage. The equipment does not have to include every feature and specification required by the standard, and can have lower accuracy and dynamic range than compliant receivers if sufficient margin is applied to the test results.
General-purpose spectrum analysers with general purpose filters and detectors are a good starting point for pre-certification and EMI radiation testing. However, for more comprehensive analysis is it useful to have WLAN-specific test that supports the full range of test and analysis required for IEEE 802.11 standards. Such a solution, particularly when based on a mixed domain oscilloscope and coupled with vector signal analysis software allows designers to correlate events in the time domain with frequency domain analysis for fast identification of problems that could cause a product to fail certification. For instance, a glitch that only happens during wake-up in the time domain could be causing an out-of-band emission in the frequency domain.
External labs typically begin testing by performing a quick scan using peak detectors to find problem areas that exceed or are close to the specified limits. For signals that approach or exceed the limits, they then perform a QP (quasi-peak) measurement.
The QP detector is a special detection method defined by EMI measurement standards that serves to detect the weighted peak value of the envelope of a signal. It weights signals depending upon their duration and repetition rate. Signals that occur more frequently will result in a higher QP measurement than infrequent impulses.
A good rule to remember is QP will always be less than or equal to peak detect, never larger. Peak detectors give a margin so peak detection can be used for spectrum emission troubleshooting and diagnostics. There is no need to be accurate to a full compliance department or lab scan; if the lab report used the QP detector and shows the design was 3dB over and the peak detect is 6dB over, then you need to implement fixes that reduce the signal by -3dB or more.
Pre-compliance probing technology
In a full compliance lab, EMI receivers and well-calibrated antennae are used to test the electronic devices over a distance of three or 10m. The measurements might be done in the far field. Such a test can accurately tell whether the product passes or fails but cannot point the source of a problem. Using only the far-field test cannot isolate problems down to specific components or locations, like too much RF energy leaking from an opening in a metal enclosure or help identify a cable radiating too much RF energy. A near-field test is the only way to locate such emission sources and is typically performed using a spectrum analyser and near-field probe.
Near-field probes for EMI are electromagnetic pickups used to capture either the electric (E) or magnetic (H) field at the area of interest and are used with the spectrum analyser. Manufacturers provide kits of probes that offer the best compromise between size, sensitivity and frequency range. Selection between an H-field or E-field probe may be driven by location of a signal in your design, or by the nature of its source (voltage or current). For example, the presence of a metal shield may suppress the E-field, making it necessary to use an H-field probe for the application. Near-field probes must be used to either pick up the signal near the device under test.
Voltage probes are used with oscilloscopes and spectrum analysers to attach directly to the circuit of interest. Conventional oscilloscope probes can be used with spectrum analysers with loss in sensitivity depending upon the impedance of the probe. For example a 500? Z0 oscilloscope probe connected to a 50? spectrum analyser will result in a 10:1 divider and a reduction in signal to the spectrum analyser input of 20dB. However, when connecting directly to a circuit, the signals are generally large, and can be seen by the spectrum analyser even with the reduced signal level. Furthermore, the noise floor and sensitivity of a spectrum analyser is typically orders of magnitude better than an oscilloscope, so loss from a probe is rarely a limiting factor. Voltage probes must be attached directly to the circuit to pick up the signal.
Three basic steps for pre-compliance testing
For pre-compliance testing, the frequency domain is divided to three sub-domains or zones (Figure 4). There is a three-step spectrum pre-compliance test, consisting of in-band (channel) domain (check the transmit power output, the transmit bandwidth, and power spectrum density); out-of-band domain (check the spectrum emission or the adjacent channel power ratio (ACPR). The mask is usually defined by communication standards like IEEE) and the spurious domain, to check spurious emission levels.
When planning or updating a wireless device installation, it is often necessary to determine if the wireless equipment can achieve a certain transmission distance. This information is not printed in the specs for wireless devices and antennae. Standards and regulations have to be checked to ensure that the device is able to pass the compliance test. This is performed, using vector signal analysis software along with a spectrum analyser.
It is important to note that some WLAN signals exceed the bandwidth of an analyser to perform the transmitted power measurement. For example, an 802.11ac signal would require a bandwidth of at least 160MHz to perform the burst power test.
The power spectral density is the power within each unit of frequency. The FCC requires that the power spectral density conducted from the intentional radiator to the antenna shall not be greater than 8dBm in any 3kHz band during any time interval of continuous transmission. The spectrum analyser centre frequency needs to be set to the channel centre frequency, set the RBW to 3kHz, and use peak detector and marker to determine if the maximum amplitude level is greater than 8dBm.
Occupied bandwidth measurement
Occupied bandwidth is a measurement of the frequency band bandwidth that contains a specified percentage of the total power of the signal. Occupied bandwidth is a measurement of how much bandwidth a signal consumes within an allocated channel. Typically, the occupied bandwidth is specified as a percentage of the total power within the allocated channel bandwidth.
Once the transmit power output of a device meets the in-band compliance requirement, it can be moved on to test the out-of-band emissions. A spectral mask is a mathematically-defined set of lines applied to the levels of radio transmissions. This mask provides the limit under which the signal power is allowed to distribute over the channel. The transmit spectrum mask is defined for each variant of the standard. Generally, the spectrum emission mask (or out-of-band) domain starts at a frequency offset of 0.5 times the necessary bandwidth (allocated channel bandwidth) and extends up to 2.5 times the necessary bandwidth. For example, the IEEE emission mask domain of a 20MHz bandwidth 802.11g signal is from ±10 to ±50MHz frequency offset from its center frequency.
A spurious emission is any radio frequency not deliberately created or transmitted, especially in a device which normally does create other frequencies. A harmonic or other signal outside a transmitter's assigned channel would be considered a spurious emission. The local regulatory standards provide the limit (permissible value) of spurious emission power of a given unwanted emission domain.
Overcoming RF integration complexity
Many manufacturers assume that they can just buy a wireless module and have the product certified and ready for market with very little effort. Yet for even fairly simple integration efforts, there are many potential areas for problems and complex regulatory requirements that must be met. Given the cost of time of going to a compliance certification facility, pre-compliance testing is a must.
However, checking all the standards and regulations is difficult and time consuming. Even when all the regulatory information has been collected, the engineer must be familiar with test equipment and make sure all measurements are set up correctly. Pre-compliance procedures can add up to hundreds of clicks on spectrum analysers.
To simplify this process, test and measurement suppliers have started offering step-by-step guidance or wizards for using mixed domain oscilloscope or spectrum analysers to perform pre-compliance measurements for WLAN devices. Further, diagnostics are not limited to pre-compliance testing. These tools also support extensive diagnosis and troubleshooting to ensure that the RF subsystem performs up to specified levels without being degraded by other parts of the integrated system.
Contact Details and Archive...
