176 R. Wetherington et al. Furthermore, the need for expensive anti-alias filters is eliminated since sigma-delta conversion technology is often used in these devices. There are some tradeoffs in using audio equipment for data acquisition. The most significant is the fact that most audio adapters are A/C coupled, which means they do not digitize DC values; however, many audio adapters will accurately resolve frequencies as low as 2–3 Hz, which in many cases is sufficient. There are a few audio adapters that claim to digitize all the way down to DC but these are less common and address a special focus of the audio market. Another tradeoff is latency (the delay between the sound input and the recording). Audio recording gear is designed for very low latency, resulting in very little digital buffering in the adapter. This characteristic has major impact on the computer and software that receives this digitized data in terms of the risk of missed samples. This can be an extremely difficult challenge to address. Some vendors address this challenge in specialized device drivers they provide with the equipment. These drivers are typically written for the Microsoft Windows operating system and may not be available for other operating systems. Some of the devices provide gain and filtering adjustments on a per channel basis. In some cases, these controls are available as knobs and switches on the front of the devices. While this can provide an incredible amount of flexibility when deploying the devices in the field, at some point the effects of the adjustments must be factored into the measurements before analytics can be performed. This can be problematic if note carefully planned for. Possibly the most imposing challenge to using audio adapters is the lack of vendor provided specifications and performance metrics. This means the burden is on the developer to determine how well this approach works and to what degree it can be certified. Environmental compliance is also important to consider. Testing and validation of these devices can be a sizeable amount of work because these devices are not marketed or intended for scientific research applications. Furthermore, the developer is essentially “on their own” for the applications software development and testing. This can be a significant commitment. However, the benefits of using audio adapters can be realized if the number of deployed units and channels is high. The investment to develop and qualify a design is then distributed across a large number of field deployments. ORNL has been using low-cost consumer grade audio devices for several years in high-speed data acquisition applications [1–4]. One of ORNL’s more recent adaptations of consumer audio gear for data acquisition is for a project that involves high-speed collection systems that are deployed around ORNL’s HFIR and REDC facilities. Data acquired by the collection systems is analyzed and correlated to operational ground truth information from the facilities. USB audio adapter are used as the foundation technology for three of the collection systems, which may be expanded in the near future. 15.2 Design The architecture of the MINOS high-speed collection systems is shown in Fig. 15.1. A Behringer UMC404HD audio adapter is coupled with a PC running Centos Linux. The Behringer unit digitizes four input channels by sampling each at 96 kHz with a resolution of 24-bits which is provided as a four-byte sample word. Advanced Linux Sound Architecture (ALSA) Fig. 15.1 Test Setup
RkJQdWJsaXNoZXIy MTMzNzEzMQ==