Measuring space

MTC team members (Ref. 1) have backgrounds in electrical engineering, computer science, physics, mechanical engineering, and measurement sciences. But if you ask what they do, the engineers will say "we're measurement systems integrators." Because every project requires knowledge in several engineering disciplines, each engineer takes on tasks outside his area of expertise. "Even the computer science guys know how to wire a chassis," said Yates. "They resist at first, but we wear them down."

MTC engineers provide measurement expertise to projects at JPL and at other NASA sites. They also assist engineers and scientists at Pasadena's California Institute of Technology (Caltech), located a few miles away. Thus, not all MTC projects revolve around space-based systems. MTC engineers work on 20 to 25 projects at any time. Project engineers at JPL, other NASA sites, and Caltech may request MTC's help at any point in a project. "About 50% of the time, we're involved at the beginning of a project, which we call 'prephase A,' and that's what we prefer," said Yates. "Other times, we're called in when an engineer isn't getting predicted measurements."

Phil Yates manages engineers with backgrounds in electronics, software, physics, and mechanical engineering.

A JPL project consists of milestones, schedules, budgets, demonstrations, and design reviews. A project must pass through a requirements review, a preliminary design review, a critical design review, and a pre-ship review. MTC engineers participate in project design and delivery reviews.

MTC engineers often hold their own design reviews where they discuss the details of the project from a measurement perspective. At these reviews, MTC engineers discuss the details of a measurement system's hardware, software, and workmanship.

Longtime JPL engineer George Wells (left) mentors Erik Peterson and others in software development.

For example, they will open any custom chassis they build and inspect the chassis for workmanship and safety compliance. They review front and rear panels for clear labeling and for a logical and effective layout. They also inspect software for its architecture, implementation, and documentation. Finally, they review a spacecraft system's requirements for test and validation.

Because the customer isn't present at these reviews, MTC engineers freely discuss ways they could have done better. From each project, they learn how to improve future projects.

In a recently completed project, MTC engineers built a test system for a spacecraft's radar controller. The test system emulates a spacecraft's data bus—a BCP 2000 from Ball Aerospace—by generating and receiving commands from the spacecraft's instruments. MTC engineers have also emulated the Japanese ADIOS spacecraft bus and are currently working to simulate the bus that will go into the Argentine SAC-D spacecraft, scheduled for launch in 2008.

To emulate a spacecraft bus, MTC engineers often build custom hardware that includes a bus interface that lets a PC communicate with the spacecraft's payload, which are usually scientific instruments. The interface often consists of an RS-422 (or open collector) three-wire interface with data, clock, and enable lines. A bus may also contain analog and digital control lines.

To use a standard bus such as MIL-STD-1553, the MTC engineer simply buys an interface card for a PC. Even for standard buses, the spacecraft bus emulator will need custom software to generate the command sequences and display telemetry.

Brian Franklin designs embedded systems and builds electronics chassis for testing spacecraft and earth-based systems.

Dawn's three ion engines will use xenon gas to propel the spacecraft once it is out of earth orbit. For the engines to work, they need pure xenon. Chemical contaminants from Dawn's xenon tank must outgas before the spacecraft can launch.

To perform the outgassing, JPL engineers place the spacecraft in a vacuum chamber, pump out the air, and heat the spacecraft to 50°C before pumping pure xenon into the spacecraft's fuel tank. At the end of the test, JPL engineers will remove the xenon and analyze it for purity. Only when the xenon is sufficiently pure can Dawn get the gas it will use in flight.

Figure 1.  To prepare the Dawn spacecraft for flight, MTC engineers placed it in a vacuum chamber and heated the spacecraft with lamps.

The data-acquisition system monitors 48 thermocouples, eight in each of the six temperature zones. A PC collects the data through an IEEE 488 port and controls the power to each lamp based on the highest temperature measurement in each zone. A National Instruments analog-output card in the PC drives silicon-controlled rectifiers that regulate power from the supplies to the lamps.

Figure 2.  Custom power supplies power the lamps that heat Dawn while it is in the test chamber. A data-acquisition system measures Dawn’s temperature during outgassing.

The tank test system is one of three projects that MTC engineers participated in for the Dawn mission. Others were a tester for the spacecraft's Ion Propulsion Gimbal Actuator motors and a tester for its digital control interface unit (DCIU), which controls Dawn's ion engines. MTC developed a DCIU simulator that mimics the input and output signals. The simulator went to Orbital Sciences for integration testing.

Because of changing priorities, some of the products that MTC engineers test never go into space—yet they remain viable, useful products. One example is a cesium-fountain atomic clock. Intended as a science experiment in accurate timing, the clock could have provided a 10x increase in timing accuracy because of zero gravity.

Although earthbound, the clock is still one of the most accurate in the world. From its location at JPL, it can provide a timing reference for the Deep Space Network (DSN) where precise timing and low noise signals contribute to improved navigation and tracking and a greater dynamic range for the DSN receivers. JPL can compare this clock to other clocks on the campus or to Universal Coordinated Time. Other cesium-fountain clocks are located at NIST and at USNO (Ref. 2, 3).

Figure 3.  A cesium fountain clock uses a vacuum chamber to excite the cesium atoms, which generate a known, repeatable frequency.

To develop the clock's control system, Yates assembled a team of MTC engineers with expertise in electronics, microwaves, real-time control systems, and physics. Physicists collaborated with hardware engineer Brian Franklin and software engineer Erik Peterson so they could develop the clock's hardware and software. Senior engineer James Granger designed a microhertz resolution X-band synthesizer with ultra-low phase noise for this task.

Three computers run the clock. A single-board computer (SBC) in a VME chassis runs the clock's equipment, a PXI chassis handles timing calculations and data acquisition, and a Windows-based PC handles the user interface. If the clock had gone into space, another embedded controller would have replaced the PXI chassis. A PC now emulates the Space Station interface and sends commands and receives data packets from the SBC.

All three computers run LabView, with the SBC running a prototype version of embedded LabView on the VxWorks real-time operating system. "When we took on the project," said Yates, "we knew that we had to write the software in LabView because that's where we have our greatest depth in programming, although we use C, C++, and Visual Basic, too." JPL's experience with LabView dates to the 1980s with version 1. "MTC and LabView," traces MTC's experience with the programming language.

MTC engineers support projects outside of JPL, too. Recently, Franklin and Peterson automated a 10-nm atomic-force microscope for Caltech that researchers use to study molecular interactions.

The original microscope used a commercial off-the-shelf controller from Digital Instruments. The Caltech researchers who built the microscope manually operated its control loop. "When we arrived," said Franklin, "we saw them controlling optics' offsets and gains with knobs mounted in 14 little boxes." So, they reverse-engineered the manual control system and designed an automated control system that drastically cuts focus time. "The reverse engineering cost was 50% of the entire cost of designing and building the control chassis," noted Yates. "Brian had to figure out what the Caltech people were doing and why before he could design the hardware."

Figure 4.  A chassis contains control electronics and power supplies for an atomic-force microscope.

Franklin explained that as the stages move, the probe tip oscillates just above the sample. "At the top of its movement, the tip's output signal is down in the noise," he said. "As it approaches the sample, the signal amplitude grows because the probe tip detects more photons." A National Instruments' high-speed data-acquisition card in a PXI chassis digitizes the height signal, so the embedded controller knows the tip's height from the sample. Software in the PXI chassis subtracts the noise from the acquired signal.

As the stages move, a photon detector in the microscope probe counts photons at each stage location and produces a pulse train that represents light intensity. In effect, it counts photons. A counter card in the PXI chassis counts the pulses, from which software in the PXI chassis' controller creates images and sends them to the host PC over an MXI link. Researchers can view the images and study how a sample's molecules interact.

Whether they need a test system for spacecraft hardware or they need automation and programming expertise, project engineers from JPL, other NASA sites, or Caltech rely on MTC for help. With measurement systems developed at MTC, missions such as the Mars Rover and Voyager will keep expanding our knowledge of the universe for years to come. T&MW