Machine Design

Mars Rover to Earth: "Where to now?"

Harmonic drives boost pointing accuracy in high-gain antenna drives aboard Mars Explorer Rovers.

A high-gain antenna will beam data back to Earth.
Team members gather around Mars Exploration Rover 2 and its smaller predecessor, a flight spare of the Pathfinder mission Sojourner rover.

If all goes as planned, two Mars Explorer Rovers launched last summer will land on the Red planet this January. Once on the Martian surface, the go-cart-sized, 185-kg rovers will travel about 100 m daily, searching for signs of ancient water.

The Rovers will stay in touch with Earth during their 90-day missions using three different systems: a low-gain antenna (LGA), a UHF antenna, and a high-gain antenna. The omnidirectional LGA transmits at a low data rate to the Deep Space Network when rover orientation is unknown. DSN antennas communicate with far-flung spacecraft at frequencies of 2.2, 8.4 and 32 GHz. The omnidirectional UHF antenna communicates through orbiters Mars Odyssey and Mars Global Surveyor passing overhead.

About half of all communications will go through the HGA, a 0.28-m-diameter antenna that beams data directly to earth receivers over the X-band (8 to 12 GHz) at 1,850 bits/sec. The HGA is not omnidirectional so a two-axis high-gain-antenna gimbal (HGAG) points it to a receiving antenna.

Identical azimuth and elevation drives steer the antenna through a hemispherical field of regard. Each axis uses a 34-Vdc REO 20 brush motor from Maxon Motor, to spin an integral three-stage, 81.37:1 planetary gearbox which, in turn, drives a 1.333:1 spur-gear stage. The spur-gear stage then powers an HD Systems Size-14 SHF 50:1 harmonic drive, for a final reduction ratio of 5,425:1. Harmonic drives are noted for their ability to retain out-of-box backlash specs without adjustment for the life of the device. They also pack high output torque and stiffness into a small, lightweight package -- all are important metrics for the application.

Slashing backlash

A typical planetary gearbox may have about 2° of backlash. MER's smaller-sized predecessor, Pathfinder, moved its instruments with precision spur-gear reducers. But the reducers required high levels of preload to cut backlash and boost stiffness. Backlash in harmonic drives, in contrast, is typically about two to three orders of magnitude less than in planetary drives. The use of harmonic drives for HGAG final output compensates for backlash in the planetary and spur stages. Boosting HGAG mechanical accuracy in this way allows a more generous error budget for the fairly simple control system that runs it. A high level of stiffness helps the antenna hold position without motor power when Rovers traverse rough or sloped terrain.

"The antenna drives move slowly and running torques are low, the very conditions at which harmonic drives are least efficient," explains Randy Lindeman, an engineer with the Jet Propulsion Laboratory. "But mechanical efficiency is not a big issue in this case because there is more than enough power to run the drives." Solar arrays deliver between 100 to 140 W to storage batteries. Motors consume only a few watts. However, torque margin is critically important. Specs say drives must supply double the amount of output torque needed to overcome a 40° slope when temperatures plummet to -70°C and battery power is at low ebb, considered a worst case in normal operations. And they must not be so powerful that they break delicate hardware.

A 5,245:1 final drive ratio provides the required torque margin (about 5.4 N-m) and limits motor revolution count to 2.5 million over mission life, another design requirement. The low-pressure, CO2-rich and nearly desiccated Martian atmosphere happens to be tough on motor brushes.

There was also a question of how lubricated harmonic drives would behave at deep subzero temperatures. For example, the kinematic viscosity of grease base oils can reach 12,000 cST at -70°C, a factor of 50 higher than at room temperature. (The viscosity of water is about 1.00 cST at room temperature for comparison.) No-load-torque and efficiency tests conducted by JPL engineers showed that is was better to grease-coat drive bearings and gears rather than fill them with grease to a higher percentage.

Torsional stiffness and ratcheting were other considerations. The drives must meet the HGAG pointing spec but also be compliant enough to limit loads when commanded into hardstops at the motor's 34-Vdc maximum operating voltage. Then, harmonic-drive peak torque may reach 52 N-m. Radially flexible members -- circular spline, wave generator, and housing -- must be adequately stiff to prevent ratcheting. Ratcheting happens when the flex spline doesn't mesh properly with the fixed outer ring gear in the housing. A finite-element model from engineers at HD Systems estimated minimum ratcheting torque at 77.4 N-m, well within the peak torque spec, and significantly above the worst-case operating torque.

Make contact

HD Systems Inc.
Hauppauge, N.Y.
(800) 231-4374

Maxon Motor
Burlingame, Calif.
(800) 865-7540

Get the point?

The HGAG uses a PID digital controller with feedback from motor encoders and from a potentiometer for absolute position reference. Stored in memory is what is termed an Ephemeris file containing the relative motions and positions of Mars, Earth, and the Sun. The program and a high-accuracy clock directs the HGAG to track an antenna on Earth. The Earth from the perspective of Mars moves across the sky at a rate of roughly 15°/hr.

Link margins (a measure of signal strength) require that the antenna stay aligned within 2° of its Earth-based target during a communication session. To slew the antenna, the controller ramps up motor current, runs the motors for 4 sec at maximum speed which equates to a slew rate of 3°/sec, ramps down, waits 20 sec, then does it again. This on-off duty cycle repeats for an entire communication session. A session lasts about one hour and happens three times each day of the mission.

Ultimately, pointing accuracy is the litmus test of a successful antenna drive design. HGAG pointing accuracy is about 0.2°, less than half the 0.5° "ceiling" error spec, and a factor of 10 less than that needed for link margins.

Results of no-load torque tests show a clear relationship between input torque, speed, and operating temperature from room temperature down to -55°C. Data at -70°C indicate a similar relationship, though frictional heating at higher speeds probably warmed the lubricant and lowered input torque. In all cases, starting torque was consistently higher than low-speed running torque.
Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.