Colorado State University
Design constraints required that RoverSat be a completely autonomous vehicle with both
electrical and mechanical systems capable of withstanding a wide range of temperatures
pressures and humid conditions. In addition to these basic conditions the rover was expected to
detect landing and survive the impact of landing so that it could complete it’s mission of leaving
the landing site and storing at least one image of the surrounding area. For the case of the rover
we constructed weight was a very crucial factor in the design freedom as our school was to be
sending off two projects instead of just one on the launch date of August 2, 2003. For this reason
all design was done with an overall weight limit of 750 grams for the rover and all of its systems.
This was one half of the allotted 1500 grams that each school had been given for their balloon
payloads. Near the launch date it was discovered that there was more available weight for
schools in need so some additional reinforcement and revisions were done to take advantage of
the much-needed weight. After completion of the revisions and reinforcing of potential problem
areas in the rover structure the total weight was 830 grams. When the rover was retrieved on
launch day there was not a single sign of mechanical failure or any indication of fatigue anywhere
in the mechanics of the vehicle. For this reason the rover was considered to be a complete
With weight being the most influential force in our design of RoverSat it was decided from
the beginning of the project that our goal for every part of the vehicle should be to minimize the
number of systems and components used. To allow more weight in the vehicle itself we chose
not to place the vehicle inside of a carrier. Allowing more mass in the vehicle itself would allow
for more weight on the wheels thus helping RoverSat conquer rough terrain like thick grass and
Sending a vehicle through such conditions as those seen by the DemoSats presents the
designers with many variables that are often difficult to account for completely. The largest one
we, as a team, had to deal with was the temperature gradient between the ground and the edge
of space. Many materials and devices will not function properly if at all at such low temperatures
as those to be experienced by our payload. With the decision to eliminate the container it was
necessary to take the proper precautions to ensure that the mission was not a failure simply
because a single component got too cold. With a little research, components and materials were
found that could withstand the low temperatures at hand.
Another problem with a vehicle that did not have a container to protect it was the impact
during landing. To protect our vehicles structure and components during the landing sequence a
simple bumper was constructed that surrounded RoverSat’s body. The intent of this bumper was
to prevent the wheels from directing the impact of landing directly into the driving servomotors
connected to them. With several tests conducted at various velocities and vehicle orientations it
was found that such a bumper system was sufficient to protect the vehicle.
With a final design concept that involved no container prototype after prototype was
constructed and tested with modifications, small and large, made to each one until our final
vehicle was fabricated and assembled.
The rover itself was separated into several separate mechanical systems during the
design process to simplify the manufacturing and assembly of all the necessary parts. By
building separate subsystems modifications and replacement of broken parts after testing would
be much simplified. The entire vehicle was compiled from five simple subsystems. Each of these
systems are explained in further detail in the following order: Main Body, Drive System, Release
Mechanism, Camera System and Electronics.
Due to the weight restrictions of the vehicle it was necessary that the material
used for the main structure be very light yet strong enough to survive the landing without
damaging any of the other mechanics within the rover thus preventing its mission from
being completed. Much searching was done in hopes to find metals that would be
suitable for our needs, however, in the end our mass budget simply did not afford us the
use of any type of metal for the body. At that point it was necessary to find some sort of
plastic or composite that would be possible to tool and form the way we needed.
Common plastics posed serious concerns because of the temperatures the vehicle was
going to experience during its ascent and descent stages. Finally it was decided that a
fiberglass composite structure would be our best option.
Once our body material was selected we purchased several weights of fiberglass
cloth and the epoxy necessary to form it. Sizing of the body was done in a crude fashion.
We simply took a small cardboard box and went into a field where we could see how it
compared to the grass and rock in the area. From there a body dimension of
approximately five inches wide eleven inches long and two inches deep was selected.
Since we could easily control the weight of the fiberglass the sizing of the body was not
critical as long as it was kept within reason. The only crucial part of the body sizing was
wheel spacing. For obvious reasons the largest possible body was desirable to aid the
vehicle in covering the landing terrain.
Construction of the body shell was done using a simple mold and hand lay-up of
the fiberglass cloth and epoxy resin. Figure 1 shows the mold used for the body. The
mold was constructed from common particleboard and then coated in epoxy in efforts to
yield the best surface finish of the body as well as ease the removal of the body from the
mold. This process for making the body proved to be very effective as time passed
because it allowed fast manufacture of prototypes with little effort in comparison to
assembling the body from several pieces of fiberglass sheeting. When tests were done
on the body it was found the fiberglass was rather easy to tear and that the rigidity of the
piece was much to low to use such a piece for our application.
To prevent a huge increase in the mass of our body we had to be careful in how
we reinforced the fiberglass body piece. Eventually, we found expandable polyurethane
foam that was incredibly stiff and had a very low density. This foam was poured into the
body shell and allowed to cure at which time the excess was cut away leaving the
necessary space to mount all of the necessary mechanics. Figure 2 depicts the main
body fully constructed with all of the mechanics mounted inside.
Construction of the vehicle body out of fiberglass and polyurethane foam was far
greater in strength than initially anticipated while keeping the weight well within the
acceptable range. Several tests were conducted using this body design testing for a
landing impact at approximately twenty-five miles per hour. Such velocity was predicted
to be the upper limit the vehicle could experience while attached to the parachute as it
made its descent. Through all testing there was no sign of damage so, as a
demonstration of bravado, a Jeep was driven over the final prototype only to find a crack
in one corner of the body.
Of all the systems in the rover the drive system was the most time consuming
and difficult to fabricate. Included in this system were the axles, wheels, servomotors,
and connecting linkages. Because of the size of the rover it was difficult to find any
readily available parts that could be purchased at a reasonable price to use for
assembling the drive system. The wheels also posed many problems with their strength
and stiffness. The most time was spent in building the drive system for the RoverSat for
these various reasons.
Our first step was to choose a motor that would be suitable for our needs. After
looking at standard DC motors and plastic transmissions it was decided unanimously that
such a combination would be mediocre at best. A metal gear servo was chosen for it’s
superb torque output and small size. The brass geared servos were designed for use on
radio controlled airplanes and thus came from the factory made only to turn slightly more
than one quarter of a turn in either direction from a known center position. Because we
were going to use them to drive the wheels of our vehicle rather than controlling the flaps
of a model airplane we had to modify the gear assembly and circuitry to allow the motors
to continuously turn our driving wheels. This modification was completed by removing a
potentiometer connection and a steel pin from the gear assembly that prohibited the
output gear from turning past a given point. With the servos ready drive it was necessary
to develop a way to link them to the wheel axle.
Connecting the output gear of the servo to the wheel axle involved the fabrication
of a joint similar to that of a universal joint in a cars driveshaft assembly, however much
more simple. Below, Figure 3 depicts one of the servos with the joint piece attached and
ready for mounting within the body. The joints were made from aluminum spacers
purchased from a local hardware store. The spacers were cut down in length drilled and
filed so that the spacer could be connected via a music wire pin to the output gear of the
servo. The size of the servomotor and the spacers were so small that the holes for the
connecting pin of less than one thirty-second of an inch in diameter had to be drilled by
The axles that connected to the servomotors were constructed from brass, as it
was the stiffest material available that could be tooled by hand. Using brass of one-
eighth inch in diameter was found to be just sufficient to withstand an impact of twenty-
five miles per hour. When aluminum axles were tested in the same fashion, they
consistently bent when the vehicle landed sideways on the wheels. In efforts to reduce
the potential impact from landing on the servomotors a larger tolerance was used in the
universal joint system of Figure 3 in conjunction with a simple bearing in the wall of the
body system. Figure 4 is an image of one of the drive wheels on the final vehicle
assembly. The bearing in the body was also constructed from a common aluminum
spacer and securely mounted to the body with epoxy resin.
Connection from the axle to the wheel was done using a simple pin system. This
pin connection can be seen as well in Figure 4. Assembly in this way allowed for the
servomotors to be mounted within the body of the vehicle without the wheels being
connected at all times. During testing of the electronics this was very beneficial.
The wheel design was quite time consuming as our weight budget greatly
restricted the size of wheel that could be used. Several prototypes were constructed and
tested only to find that they were insufficient in strength for our requirements. Eventually
a plastic similar to what is found in common circuit boards was discovered at the local
plastic supplier. This G-10 plastic was made from fiberglass and epoxy resin as was our
body piece however it was much thinner and stronger than what we were able to produce
by hand. Using a single layer of G-10 a four-spoke wheel of approximately 5.75 inches in
diameter was cut. The strength was phenomenal to say the least and it weighed in at
less than half of our next best prototype. Because the G-10 was so thin and smooth the
area of the drive wheels that were to be in contact with the ground were made three
layers thick and a rubber fly fishing line was laced around the outer area to help RoverSat
maintain traction on gravel and sand.
The final drive system was designed and built with simplicity as the driving force.
No part in RoverSat’s drive system involved any complex technique to fabricate nor was
it difficult in any way to assemble. The rated torque output was approximately five pound
inches, which was decided to be more than the one-pound vehicle would ever need. The
wheels were nearly indestructible and the overall weight of the wheels reduced our
expected mass for the drive system substantially allowing for more weight to be added in
other areas of the vehicle.
Of all occurrences that could be reasonably predicted regarding RoverSat
mission, the landing sequence was the most difficult to predict. Parameters such as
landing area, wind conditions as well as location and weight of other payloads were
absolutely impossible to predict which made guaranteeing that our vehicle would
disconnect from the tether line impossible. After contemplating every possible
hypothetical situation we could imagine, we developed a simple mechanism that, in
theory, appeared to have the highest likelihood of success of anything thought up so far.
Employing another servomotor, a spool, a small cable, and an aluminum pin one
half inch in diameter, a very reliable mechanism was prototyped. The fully assembled
vehicle in Figure 5 shows the release mechanism assembled and ready for connection to
the balloon tether. The tube running along the length of RoverSat’s body (see red arrow
in figure) was used to hold the paracord tether and help orient the vehicle in a way that
would prevent it from becoming entangled in the cord during first portion of the high-
altitude descent. Reinforcements at the ends of the tubing held the tube in proper
alignment and the pin in the middle (see yellow arrow). At the base of the half-inch pin in
the center of the vehicle is a spring that was designed to eject the tether clear of the
vehicle allowing it to drive freely from the landing site. On the inside of the rover body is
a music wire pin that runs through the spring loaded aluminum pin keeping pressure on
the spring and holding tether close to the rover’s body. A servomotor with a small spool
attached to its output gear is connected to the music wire pin that runs through the large
aluminum pin. This music wire pin is pulled from the aluminum pin when the servomotor
is engaged. Once the spool has wound enough cable and the music wire pin is removed
from the aluminum pin the tether along with the steel tube is ejected from the body and
RoverSat is free to leave its landing site.
Although the release mechanism was not tested during the launch itself, due to
an electronic failure, it was believed, with strong confidence that it would have worked.
Given that the primary objective of RoverSat’s mission was to survey it’s landing
site it was crucial that the camera system be protected as much as possible from any
potentially damaging encounters during ascent or descent. To ensure that the camera
was operational, it was decided that the camera remain in the vehicle until the descent
stage was completed. The camera was kept within the confines of the body and braced
by two foam-covered plates that would were to be ejected from the vehicle immediately
after the vehicle detached from the balloon tether. The same servomotor that was used
to disconnect from the tether was used to eject the camera covers in the same fashion as
the release mechanism. Once ejected the camera was in a position to be deployed from
either side of the vehicle depending on which side of the vehicle was determined to be
upright. Orientation was determined using a simple sensor assembly constructed from
some metal scrap and photo-interrupters. With vehicle orientation known, the camera
was raised using another servomotor and pictures could be taken looking in the direction
In order to cut down the vehicles mass a digital camera was used instead of a
standard film camera. The camera was made by a company called Aiptek and had a
resolution of 1.2 Megapixels. Using a digital camera such as this allowed for more
complex picture taking abilities as we could control the direction and frequency of the
images recorded. To help keep image quality as high as possible a sensor array was
constructed using photo-resistors, which allow the light intensity to be determined in
different directions. Mounting sensor assembly directly above the camera enable
position of the sun to be known so that RoverSat could be positioned so that no images
were recorded with the camera pointed directly at the sun.
More weight was saved by disassembling the camera and removing the factory-
installed housing. In addition, this allowed mounting of the camera to the deployment
mechanism to be done exactly as desired. Using a lightweight piece of G-10 as the
mounting base, the camera circuitry and lens was easily mounted and then connected to
the servomotor that controlled the camera deployment. Figure 6 shows the camera
assembly in the body. The green arrow locates the sensor array used for orientation of
the vehicle while the blue arrow shows the lens portion. The camera circuitry is on the
right side of the lens in the figure and the servomotor is highlighted with the red arrow
and is connected to the mounting piece we manufactured.
From a future mechanical engineers perspective the electronics posed little
concern in this project. The most important thing was that nothing happened to the
components during any part of the mission, especially the landing portion. In regards to
the electronics, our efforts were directed strongly towards surviving the impact of landing
and the extreme temperatures during ascent and descent.
To protect the electronics during landing impact much care was put into the
mounting of the circuit board. Because there was a strong possibility that the
components within the vehicle would be very cold at landing the circuit board was
mounted to the body with bolts but in a loose fashion to prevent any chances of brittle
fracture of cold plastic components on the board. Furthermore silicone sealant was used
to cover the circuitry of both the camera and the control system. This was done to
prevent the possibility of a short circuit occurring if a component were to break off inside
the vehicles body and fall onto the circuitboard. It was likely that such measures were
excessive but they were taken anyway.
Because the mechanical system appeared to work just as expected but the electronic
and programming portion was a complete failure, we currently intend on rebuilding the circuitry
and reconstructing the programming necessary to complete the mission in time for a launch in
November 2003. We have much faith in the mechanics of RoverSat and would like to see the
vehicle succeed at its intended purpose.
The DemoSat program allowed for a huge degree of design freedom and allowed our
group members to learn not only about the design process but group dynamics and about how to
succeed not only as an individual but as a team. Leaving the project at the end of the ten-week
program left us more prepared for our senior design projects starting in the upcoming semester.
Although the mission as a whole was not a success, it was a good experience and something that
would be done again with no hesitation.