Electronics Report in Elephant Moraine, Antarctica
By: Mark Sibenac
Robotic Antarctic Meteorite Search
Year 2000 Expedition
2000-01-10 -- 2000-02-02

Nomad is equipped with 1600 Watts of heaters inside of the electronics box along with ~150 Watts of heaters for external components. These heaters are switched on a few minutes after the generator warms up in the morning. After about 1.5 hours, the temperature is around 20 deg C inside the electronics box. The electronics are switched on at this time.

The wireless communications components used with Nomad are wireless ethernet for main communications with the robot and the wireless serial modem is used for DGPS corrections. Nomad has high gain omni antennae for the wireless ethernet and wireless serial. The base camp uses a medium gain omni antenna for the wireless serial and a medium gain pointed (~85 degree beamwidth) antenna for the wireless ethernet. So far at distances up to 500 meters with line-of-sight, no dropouts have occurred. 

The pan-tilt unit that moves the hi-resolution camera on Nomad's mast has failed once on startup so far. It needed human intervention at the robot's site to startup successfully. There were many problems with this unit during the tests at Williams Field.

Nomad's arm has a touch sensor mounted on the spectrometer probe to sense when the wrist is touching a rock. It does not seem to function correctly in the conditions here. It gets stuck in a tripped condition most of the time.

The panospheric camera had a fogging problem on the lens. The top was taken off yesterday to let the moisture out, and now most of the moisture is gone. It should all be gone tomorrow. The hi-res camera also had some frosting on its glass, but that went away after the first few days. The hi-res camera has a lot of dessicant near the lens and glass.

The science computer (meteorite.frc.ri.cmu.edu) had a hard drive failure. It needed to be replaced. No further information is available on the cause for the failure.

The INMARSAT terminal is functioning extremely well using the ComSat Land Earth Station. Internet connections occur one to two times a day at 64kb/s for about 5 minutes at a time. This should increase as we will start sending robot data. The voice works well also, and no problems are to be reported for over 30 minutes of use. The Stratos Land Earth Station does not work for the Internet connection. Their technical engineers are working on the problem.

The Iridium Satellite phone works okay. The signal drops out a lot when the satellites get low on the horizon. Many people have reported not being able to reach us on the phone.

No problems are to be reported for the electromechanical system.

The arm controller had a special "controller reset" relay connected to an I/O card on the science scomputer. When the I/O would get initialized, the arm controller would go into a reset condition. This caused the arm to drop to the ground because of the lack of servo control. The relay was removed to prevent the problem. We can still reset the arm controller by remotely power cycling it.

The arm controller returns an error message "SCA1 - stop code" every once in a while. It does not seem to affect perfromance, but it is carefully monitored.

The compass seems to be affected in a bad way according to heading of the robot. One would want a compass to read 0 deg for north, and 180 degress for south, but Nomad's compass will read 0 deg for north, but 175 for south. We verified the real heading according to the GPS which is accurate to about 3 cm. The compass was last calibrated at Elephant Moraine, and got a calibration score of <2 degree accuracy and a placement score of good. The compass results are worse than the specs according to the calibration score.

The laser developed a problem early on in the expedition. It seems like there is an obstuction about 5cm in front of the laser around the middle area. This corrupts the data and renders the sensor useless for obstacle avoidance.

Meteorite developed problems with the backup harddrive. The Linux OS does not come up into a state which can be worked on remotely. After powercycling two times, it came up okay. There were messages in /var/log/messages that the hard drive could not access some of the sectors.

Again, we experienced problems with VHF comms.  We can get VHF comms from 1 km about 70% of the time. The 20 Watt VHF base station that ASA provided did not work any better, even though it had a 1 m  antenna that was elevated about 2 m off of the ground. The VHF base station was the best at receiving transmissions from the field. I reccomend setting up a VHF repeater next time. The VHF comms just don't work on flat ground when the ground is ice and snow. There were no objects to bounce radio signals off of either.

The wireless ethernet radios (arlans) worked extremely well the entire expedition running at 2 Mbps. I saw a couple of dropouts when the robot was about 1.2 km away from the base station. A higher gain antenna at the base station would have helped out in that situation.

The wireless radio modem always seemed to run okay.

Nomad was left running overnight because of the scare with bad sectors on the meteorite harddrive. In the morning, many bad sectors had been encountered and the entire /usr/nomad directory was corrupted and inaccessable. Today is the last day of operations anyway, so it will not impact us much.

Burning CD's from a Pentium II 400 MHz laptop works at 4x write speed. This really helps out since it only takes about 20 minutes per cd.

