Brunner

TracVision started life as a new product development focused on bringing IoT monitoring and control capabilities to medical facility power conditioners.

The primary constraint on the project was to build an extremely low-cost solution, including all manufacturing processes and labor. The product had to fit within a large range of existing equipment with varying form factors while still being available for new original equipment manufacturers. Finally, TracVision was envisioned to develop into a platform solution for a family of related products, and thus it had to support future expandability of sensors, transducers, actuators, and communication channels and protocols.

I chose to base the TracVision’s control system on the extremely popular and inexpensive Raspberry Pi line of single board computers (SBC’s). The Raspberry Pi Compute Module is the industrial version of the computer and was extremely robust in specification while staying extremely low cost.

The Raspberry Pi supports a large number of open source Linux based operating systems, and I chose a compact distribution based on the popular Debian system. Under the hood, I created a system of Python programs that functioned as the system monitor, application programming interface (API), and sensor module driver.

I built TracVision with a number of unique and modern features to fulfill its purpose as a low-cost IIoT device:

• a containerized application layer based on the popular Docker architecture that supported in-place software updates distributed via the internet and a recovery mode base image.
• a modern web interface developed with Javascript and React featuring encrypted and authenticated access restrictions using tabbed interfaces, real-time graphical plots, and
• a Git-based CI/CD build system that implemented robust unit and integration testing and automated deployment of the software to the production facility.

I was part of the team that developed the original bench-top passive intermodulation distortion analyzer (i.e., “PIM Analyzer”) and later became the Director of Engineering for Summitek Instruments, the world leader in PIM instrumentation.

At the dawn of the wireless cellular revolution (circa 1990), companies were starting to deploy huge networks of cellular basestations which were used to provide the link between their customer’s mobile phones and the telephone network.

The key components in wireless basestations are the antennas used to transmit and receive the signals from the subscribers’ phones, and it turned out that the quality of these antennas was critical to the optimum performance and revenue generating capability of the basestations. Poorly designed or incorrectly assembled transmit antennas would generate a type of harmonic distortion that would directly block the relatively small signals from handsets in the basestation’s area. Dropped calls due to this effect meant a significant hit to revenue.

Intermodulation distortion caused by the passive device (i.e., the antenna) was the culprit, but antenna manufacturers had no easy and standardized means to test their antennas for the distortion. In 1998 I was part of the team at Summitek Instruments that created and sold the first complete Passive Intermodulation Distortion Analyzers. These so-called PIM Analyzers were turnkey boxes that could measure and analyze antennas (as well as other passive devices) in a rapid enough manner that the test could become part of the manufacturing production line. It soon became the world standard that antennas to be used in wireless basestations be tested on Summitek equipment to show their PIM quality.

Although I was intimately familiar with the entire PIM analyzer, its operation and capabilities, I was the engineer in charge of developing the instrument’s user interface. I developed a robust control API for the instrument and used my software design experience to create a very intuitive interface for a relatively complex device.

Over the years, the original Summitek Instruments PIM analyzer evolved and combined with another company’s mobile version of the test set. Summitek Instruments and Triasx became Kaelus, and I developed the user interface for their mobile PIM analyzer, the iPA. The iPA relied on the same underlying principles of PIM measurement, but used a more modern interface. Users could operate the device from its built in screen, or from any web-connected device such as their computer, tablet, or smart phone. I built the interface to work on any of these devices while preserving the simplicity and capabilities of the original interface.

The Spartan Quality Management System (QMS) was a traditional client-server enterprise application used to streamline automated testing and data management for midsize RF and microwave component factories.

With Spartan QMS, companies could get automated testing and data analytics running quickly and inexpensively. The target market for Spartan QMS was midsize manufacturing operations producing ten to hundreds of assemblies per day that required significant RF testing operations as part of the manufacturing process.

Spartan QMS was installed onto one of the company’s servers or was supplied on a new server. Test stations throughout the production line were connected to the company IT infrastructure via TCP/IP (Ethernet/LAN) and set up within the Spartan QMS configuration database to carry out the automated testing. Spartan QMS automatically distributed the appropriate test runners to the test stations and collected and indexed the resulting test data.

The key technologies leveraged with the Spartan QMS system were the enterprise server program model and database. Additionally, the local test equipment interface drivers distributed from the central server proved effective in reducing the complexity of testing allowing new or minimally skilled operators to perform the testing. Finally, the automatic data repository index created by a crawler program made the production line test data immediately available for reporting and production quality metrics.