Special Section: Temperature & Pressure
A temperature change
New reliability lifecycle test system flexes muscle of Naval Warfare Center
By William Crespo and Anthony Bryan
When it comes to reducing costs and increasing lifecycle, the U.S. Navy is making great strides. Transferring complex test, deployment, and field systems to full or partial commercial-off-the-shelf (COTS) software and hardware makes the government flexible enough to adapt to changes in user needs. So when a complex test system began to fail at the Naval Surface Warfare Center (NSWC) Crane Division in Crane, Ind., Technology Service Corporation (TSC) stepped in for a complete redesign to provide increased test data quality and testing options.
The accelerated reliability lifecycle test system (ARLTS) comprises reliability test stations in which the devices under test (DUTs) are exposed to variables such as arbitrary radio frequency (RF) and microwave (MW) signals in the 2 GHz to 18 GHz range, temperatures (thermal cycling from 15ºC to 250ºC), and voltage and current conditions for drain and gate sources and overall device gain.
While this may be a military application, the technology can also see use in the manufacturing realm.
The main challenge was to integrate a variety of test equipment into a single cohesive system that exercises multiple test devices simultaneously, each with less than a 0.1 in2 footprint. Also, the system is required to execute individual thermal profiles reliably for each test device for periods lasting several weeks. In response to these requirements, TSC developed a low-cost solution using a combination of COTS components, devices, and custom software.
Setting up these tests is time-consuming, often requiring reconfiguration of instruments on the test station and software changes to meet the requirements of each DUT. Once a setup is complete, a typical test lifecycle can last 1,000 to 2,000 hours. Testing requires occasional monitoring by lab personnel and several hours to convert data into a presentable format. During original delivery, the system was not completely integrated, so the only support for this system comes from internal NSWC Crane resources.
The improved ARLTS consists of a lifecycle test system with the capacity to test 16 RF amplifiers/transmit receive modules. Testing involves varying operational conditions while changing temperatures. These temperatures range from 15ºC to 250ºC and are maintained for periods up to 2,000 hours. Temperature can be extended to 300ºC with additional attachments. All these units are tested in a single tower array of “Qubes” (individual test chambers). N2 valves, H2O valves, and individual heaters control the temperature during the test cycle. While the DUT is exposed to these elements, input and output RF power are recorded to determine the gain for these units. In addition, voltages and current are measured at the gates and drains of each DUT.
The new system will incorporate RF/MW sources, power meters, temperature chambers, thermal reading devices, voltage sources, and meters. The system comprises three main systems; two of them will have two subsystems. The three main systems are the host PC and uninterruptible power supply, the tower with its16 Qubes, and the support system racks (which include amps, source meters, loads, switches and control subsystem). The host PC sees use for development, remote system monitoring, data storage, reduction, and analysis. The tower of Qubes houses up to 16 independent temperature units, where each Qube contains a DUT, which stresses it to different temperatures, RF levels, and varying voltage and current levels. The third main system is the subsystem racks that contain one dual head source meter for each Qube, one set of loads for each Qube, one set of eight S-band Amps, one set of eight X band amps, power supplies for each amp, one RF source for each band, and one power meter for each band. It also has a switch system to control all the loads/signals and couplers and a separate host PC to control, deploy, and monitor the Qubes. The system is able to individually set the temperature, voltage, current, and RF signals to each DUT. Then it will measure the temperature, voltage, current, and RF signals received from each DUT for periods in excess of 2,000 hours.
The host PC can download the saved data from each Qube and process and deliver the data in a presentable form. Timing and budgetary constraints required development in phases. For Phase I, the system was a 1-up bench top system to demonstrate a proof of concept (POC). Phase II was an incorporation of the POC into a 4-up system, and the final Phase III is a full 16-up system.
The equipment for Phases I, II, and III include a Qube, with a single COTS programmable automated controller (PAC) to replace each Qube’s custom set of cards and mini microcontrollers. A small custom inexpensive interface board can pass connections from the front of the Qube and the DUT to the rear with all the measurement cards on the PAC. The PACs comprise compact field points used as the primary controller for the test during executions.
Inside the Qubes are thermocouples, a custom LED interface board, a heater, a micro valve, a pinch valve, custom LED relay board, and two thermocouples located at the Qube junctions used to monitor plate temperatures. LEDs offer front and rear interface connections for thermocouple, front panel drain, and source-feed lines for DUT voltage and sense feeds. The heater is the plate underneath the DUT or junction plate that heats or cools to stress the DUT. One valve controls the chilled water flow and completes the closed-loop temperature control system. Another valve controls the nitrogen flow into the DUT cavity to prevent humidity sweating.
External PAC controlled devices are the drain and gate source meters. Due to budgetary constraints the use of nanoamp resolution on some source meters had to be moved from a single unit per Qube controlled by the PAC to be one dual source unit per pair of Qubes. For this reason the control of this parameter passes on to the controller PC and not to each individual Qube controller. Hence this removes a parameter of independent control and decreases a level of fault tolerance.
The software package has a real-time module and developer suite with add-on toolkits and a test stand option, control box toolkit, diadem toolkit, and connectivity toolkit. The software is the heart and soul of the project; it combines all the toolkits and lets us compile and run the code. The real-time option lets us deploy our code to the target (Qube). The test-stand option is for future use of scheduling tasks of sections of code in a queuing method.
The code created in this project for Phase I, II, and III is in a continuous communication architecture in which the host computer monitors the target hardware as the test proceeds. The host program will start and prompt the user in an XML wizard to input all the test parameters. Then the host program sends the XML file over to the target and starts the test setup and subsequently monitors the test parameters. All the test parameters are measured and recorded inside the PAC in a compact flash card and copied back to the host PC, or they can download to another location if necessary. The current maximum data rate is collectable at a rate of one measurement per second for 2,000 hours of the RF in, RF out, voltage at the gate, current at the gate, voltage at the drain, current at the drain, junction, plate temperature, surface, and ambient temperature. The data throughput can be greater if the system changed to have individual input and output RF power meters, currently the system limitation is the use of 8-to-1 coaxial switching for the input power monitoring for the X and S band load and drive chassis.
During the test sequence, the target program monitors, stores, and processes the data as it runs through. The program analyzes and compares the current temperature and setpoint set by the XML file to set the heater temperature/disturbances (water flow) to control the temperature ramp-up or ramp-down. The proportional-integral-derivative (PID) algorithm is the most common control algorithm in industry. In PID control, you must specify a process variable and a set point. The process variable is the system parameter you want to control, such as temperature, pressure, or flow rate, and the set point is the desired value for the parameter you are controlling. A software PID controller determines a controller output value, such as the heater power or valve position. The controller applies the controller output value to the system, which in turn drives the process variable toward the set point value.
This is done using a PID algorithm with gains set at 5 to 1,000 and delta time set at 1 second. The acquisition analysis and set-point update is all done within one-second intervals; moreover the data acquisition to the local storage occurs at one second.
ABOUT THE AUTHORS
William F. Crespo (email@example.com) is an engineer with Technology Service Corporation in Bloomington, Ind. Anthony Bryan is with the Crane Division, Naval Surface Warfare Center (NSWC), in Crane, Ind.
Testing is a necessity
The development of new semiconductor technologies, such as gallium nitride, gallium arsenide, and other exotic wide-bandgap devices presents issues related to the verification and sustainability of its applications in high power radio frequency and microwave systems. The Naval Surface Warfare Center Crane is the Navy’s steward of sustainment and support of radar systems and components currently deployed in the fleet. This means reliable, accurate, and consistent methods of testing and validating designs as semiconductor devices is a must when replacing high-power-amplifiers vacuum-tube technology.
The accelerated reliability lifecycle test system is a key tool in verifying and validating design, reliability, and process control of these emerging technologies. Reliability testing, verification, and quality engineering are well established in the military, research, and commercial industry. While the ability to trust a company’s capability to create a quality durable product is ideal, the reality in today’s world is acceptance, validation, and control process testing is a necessity at the user end.
Return to Previous Page