Robin Irwin, Wireless Product Manager, Cobham Wireless
Continual pressure to maximise the production throughput of mobile terminal devices has made manufacturers look for ways to minimise both test time and the capital costs associated with it, since testing represents a significant proportion of the overall production time. The introduction of non-signalling tests, which removes the need for protocol-based testing by utilising the proprietary test modes within the chipsets, was a significant step in achieving this. The simultaneous testing of multiple devices then became a key goal. There can be a variety of different scenarios for testing multiple devices under test (DUTs), including an analysis of various device configurations, radio frequency (RF) routing and test control options, along with the respective design considerations and strategies. A solution offering parallel multi-DUT testing significantly increases the efficiency of utilisation of production test equipment.
Multi-DUT design considerations
The main elements of a device test system that can be reconfigured for multi-DUT testing are:
- RF test resources: RF signal generator and RF signal analyser are key items of the test equipment required for non-signalling tests. The RF signal generator provides a stimulus to the devices, and the RF signal analyser is used to analyse the output from the devices. An RF channel is defined as a combination of one signal analyser and one signal generator.
- RF conditioning: RF routing and interconnection between the RF test resources and the antenna ports of the DUTs.
- RF fixturing: Additional RF routing, device handling or external equipment that is considered external to the test equipment but is still necessary for production flow.
Configuring both the RF conditioning and the software that controls it for improving test system efficiency is the key
to performing simultaneous multi-DUT testing. In practice, the devices that need to be tested will have multiple antenna ports, each either serving different radio technologies and/or supporting the use of diversity or multiple-input-multiple-output techniques, thereby introducing additional challenges. Several multi-DUT testing options have been identified and evaluated and these are described in the following sections.
Option 1: Multiple RF channels
In this scenario, each RF channel is mapped with a one-to-one relationship to a specific device and each device is tested asynchronously. This approach is best implemented with modular test equipment, where channels can be added as required and the RF signal conditioning can also be specific to the device requirement. A traditional one-box-test approach is possible, but is likely to be less flexible as, for example, it may not provide a range of RF signal conditioning options. This means that function would need to be provided externally to the test equipment, typically through additional fixturing.
When moving from a single RF channel to multiple channels, careful designing is required to avoid interference. It is also necessary to consider how to identify a given device in the test flow within the test executive when the channel count per test station has been increased.
Option 2: Multiplexing RF resources
This is where RF resources are multiplexed sequentially between different DUTs. Two examples are shown in Fig. 1.
The arrangement in Fig. 1(a) allows for full duplex multiplexing where each device is tested sequentially, which suits a frequency division multiple access set-up where the transmitter and receiver can be tested in parallel. Fig. 1(b) shows a receiver on one device being tested, while on the other device, the transmitter is being tested, after which the connections are reversed to complete the tests. This is sometimes referred to as Tx/Rx ping-pong testing. In this scenario, each device can use an independent RF resource – an example of half duplex testing that is suited to a time division duplex /time division multiple access set-up.
Option 3: Sharing RF resources
The example in Fig. 2 shows RF resources being shared simultaneously between multiple DUTs.
Option 4: Combinations – Multiplexing and sharing RF resources
The final scenario combines elements of the previous approaches, enabling the resources to be both multiplexed and shared. It illustrates a high utilisation of the RF resources with a broadcast downlink (that allows for parallel receiver testing) and switched uplink.
Modular RF resources in a test platform
In order to enable multi-DUT tests in practice, it is essential to choose a test platform architecture that is flexible enough to meet the demands of different multi-DUT scenarios. Firstly, it is necessary to be able to scale the test system to add an RF channel, providing flexibility and expansion without increasing the footprint. Secondly, it requires the ability to dynamically assign and un-assign an RF resource so that the test executive (which integrates the modular software controls with the test instrumentation) is in charge of the test flow and its flexibility, and is not inhibited either by the architecture of the instrument or by its application programming interface (API). A further prerequisite is the ability to control an RF resource generically and independently of the parameter being tested, which gives better flexibility in the software and hardware architecture.
Electronically configurable RF conditioning
It is necessary for the test engineer to understand how to use RF conditioning to address the DUT’s multiple ports. If a range of modular RF conditioning options is available, it allows the engineer to think more flexibly and control the RF routing.
Simplifying the adoption of parallel multi-DUT testing is dependent not only on the specifications of the hardware being designed into a system, but also on the ability of test vendors to supply these parallel multi-DUT set-ups with optional modular device driver plug-ins that support leading cellular chipsets. If these plug-ins are available, then the time taken to deploy multi-DUT techniques in volume for specific chipsets can be dramatically reduced for both research and development, and production.
Modular software architecture
There are also some more general advantages of software modularity that can help the test engineer beyond the
ability to support or plug in different chipsets. Software architecture is important for the system and test engineer to control and specify tests for individual requirements. Although it may appear that complex test solutions incorporating many system components are closed architectures that cannot be customised, this is not the case with modular software architecture.
Equally, there is a balance to be found in simplifying the process that test engineers need to follow in order to get to the market quickly, avoiding what could otherwise be significant test integration efforts and complexity. This is achieved by abstracting away from specific RF conditioning/routing, power supply units and devices, allowing the test engineer to develop customisable modules. An example of such a software architecture is shown in Fig. 3. Conventionally, the vast majority of logical functions is contained within the customer application or test executive, where the software uses instrument and device-driver API calls to execute testing.
Conclusion
Using modular hardware and software, a variety of options and approaches exist when considering the design of a multi-DUT system. Moreover, a commercial production-ready ATE system solution with multi-DUT testing and integrated chipset control is equally beneficial to both large and small manufacturers. This addresses the pressure and growing challenges concerning system engineering complexity and test efficiency.
This wireless test system offers unprecedented speed and flexibility, based on a mature and proven production test platform. Up to eight mobile devices can be tested in parallel, giving an almost four-fold improvement in test throughput.
