Click for All vehicles design.

Picture of the ground station
JUNE 04, 2013 10:58 PM 3D printed parts of ground station/rover, Main PCB, etc.
ong>
In software development exist rule - if source code working from the beginning - that mean only one -- something wrong with a source code.
Such excitement discovery was made last week by Alex -- he was wondering why main computer board starts to function right away, but 3.3 voltage regulator during last month was overheated like boiling water -- well it was simple
== 6 SMD components was soldered upside-down. Current was 180 milliamps, instead of 18. Fix was quick -- it takes 10 minutes to rework PCB, but positive part is
- components was chosen properly - all 6 SMDs was working perfectly after 1 month stress.
Integration continue on GPS and Camera. HD Camera waiting it turn together with Orbitcomm backup communication modem. Integration at the rest of the week was not a full scale SNAFU, which was bothered Alex little bit.
And finally -- all 3D printed parts for a rover and ground station showed up last Friday. Some imperfections was fixed - like this on a HD camera sealed box, or on camera box -- probably the thickness of a wall (target == made from carbon fiber) will be OK, but for rover testing (that part additionally to a camera box functionality will be used as a leg for a movement) plastic is not strong enough.
For reduced size ground station tubes was made from aluminum, and assembly was in a progress
MAY 26, 2013 10:42 PM 3 configurations, Mission Ctrl + Ground station integration
Last 2 weeks - Rover - was done 3 configurations.
a) Rover as moon rover - it is a "virtual rover". Exists in 3D model formats. Majority parts has to be made from carbon fiber. To make each (carbon fiber) part need to design molds, and to print molds from PVA (like this) on a 3D printer (like that). Today we know that on that printer, to make all molds, and to make all carbon fiber frames (like this), with "space capable to fly" epoxy (like this) it is require total 45 days. That is why we named that moon rover as a "virtual one". Before arrangements of the flight to the moon, it is not practical to spend time and efforts to build such device. Only demonstration can be a purpose for rush, to build a "virtual rover".
b) Rover as a ground station for a test flight. That is exactly same rover but designed to operate on the earth surface. In this case it is not require use of carbon fiber. Same parts can be made from less strong plastic.
c) Compact version of a rover. It is not a rover truly speaking -- it looks like regular ground station, but parts for such configuration is exactly same as in configuration (b). That compact version also is for a debugging software purposes. Configuration (b) and (c) has totally different mechanical properties, but software which will control 4 stepper motors in first configuration and 2 steppers motors in second configuration, must be the same, software should be self adaptable for such different mechanical configuration, to perform task for orientation the antenna to a moving target -- flying cubesat.
Plus convenience to use 3D printing on factory allows, to separate process of design and manufacturing, make ground station on 2.4Ghz for a cubesat repeatable, and, who knows!, can allow to support main project by taking orders to build 2.5 GHz ground stations. Parts can be ordered on next week via shapeways factory.
Antenna will be regular helix (like this one), or reduced size helix (like that one). Reduced size helix has major advantages -- 3D print is twice less expensive than regular helix; different winding of a conductor allows to reduce interference of two transmitters working on the same frequency at the same time, which is important in a case of constellation of cubesats flying together.
Ground station electronics and software. The same main processor board for ground station (like this) is reused from a cubesat main processor board. Last week was done integration of a board with mission control, today mission control can operate flash memory and main computer commands. Some improvements was done on simulation of a flight, and session data visualization.
ON MAY 13, 2013 10:19 PM Ground Station, rover ver 6, clearance, solar power harvesting surface.
Last week was spent in redesign of a rover and ground station. It was done by three main reasons –
a) ground station needs to be build and 3D models for 3D printer has to be finalized.
b) clearance of a rover needs to be increased.
c) needs to increase solar power harvesting surface.
In old design on frame was mounts for flexible solar panels (cells).
each cell has size 25.4 x 63.5 mm = 16 square centimeters.
length (long side )of the frame was 600mm, that gives a place for 23 panels with total surface of 0.037m^2
two frames allows to use 4 times more - total 92 cells, or total .14 m^2
with efficiency of 0.09(9%) and power of a sun per square meter = 1000Wt surface can give 13.35 watt total.
Top side of a rover can harvest some energy, but at the same time bottom did not, it is in shadow – outcome power probably is half == 6 watts.
Solar panels mounted on frame with 45 degree angle.
Problem (b) frame is not flexible, in can stuck on terrain - as a result needs to increase clearance.
Second problem (c) - needs to increase harvesting power surface. (6 watt capable source with 40% efficiency of a 10 watt transmitter, means that transfer of data can be done only in 1 minutes with another 3 minutes 20 seconds waits to harvest the energy.
Flexible solar panel side is 63.5 mm. Assuming it is hypotenuse - that geometrically gives in prime triangle with 45 degree angle, side sizes 45+45mm – which is bigger that original 63mm.
Sizes of each cell will be smaller (like 25.4 mm x 44.9 mm) but on same frame now it is 46 cells with 0.052 m^2.
On 2 frames can be placed 4 groups of 46 cells, calculating same efficiency and same Sun’s power, that surface allows to harvest total 18.8 watt.
Same top-bottom and max outcome power is 9 watt. That is better, than previous design, and it mean == 1 minutes to transmit + 2 minutes to harvest.
Placing additional 2 stand (length 0.5m) which can hold 19x4 = 76 small size panels outcome additional 0.17m^2 with total power 15wt (and half of tit is 7wt). Gives the grand total is 34wt from all surfaces, and halfing gives max s 17 watt power. Which is not bad – on top of each minute to transmit it will require 35 seconds to harvest.
Now the question where to place that two stands. Plus needs to balance weight- i.e - 350grams for each 4 stepper motor, but gear motors can be reduced to have less torque. That can save around 50-100 grams on a motors in favor for stands. Reducing clearance makes smaller dimensions and also saves weight.
Total amount solar cell (sizes 45x25mm) to accommodate on rover today is 168. With weight limit of 400 grams for all system it is 2.3 grams per flexible cell including weights of holders and mounts. Plus needs to account that after 3 week flight cells will lost 1/2 of it power capability. Realistically it will be 8 watt of power at best.
That is a theory – up today Ground station's 3D models probably ready to be ordered, but rover's energy harvesting mounts not finalized yet.
work in progres on rover/ground station version May 2013:

===use IE on Win to see 3D model====
Design not finalized yet (needs couple of weeks to verify all 3D printing imperfections, and to fix mistakes in 3D design). Following are three design
(a) rover (model)
(b) rover - ground station
(c) simplified ground station
"Rover-ground station" is a main mechanical design intended to communicate with test flight of a cubesat "10-A". It's task is to orient the antenna to a flying CuibeSat.
"Simplified ground station" designed for a software debugging process. Software for both "rover-ground station" and "simplified design" must be the same but result - pointing antenna to the flying CubeSat should be the same in both cases. Making adaptable software for totally different mechanical configuration will allow to reuse software on the rover on the moon.
"Rover (model)" is 3D virtual design only - all 3D model parts for the rover needs to be build instead of regular plastics, but from carbon fibre. For each 3D model needs to make a mold. Then, needs to 3D print the mold from a dissolvable material (the best for today is PVA). Then, needs to knot a carbon fibre "sweater" for the 3D part. And finally make the 3D part, qualified for a space flight. For each part it takes one day for a mold design, one day for 3D printing, and one day for epoxy filling. That allows to build rover, capable to flight to the moon, in 15 x 3 = 45 man-days from today. Software design and debugging will take at least 2-3 times more.
In each cell of the table specified quantity of each 3D part, requre in each design. Some part can be build from aluminum (instead of 3d printing). Some parts requre drilling.
(a) 2; (b) 2; (c) 1.
|
(a) 2; (b) 2; (c) 0.
|
(a) 1; (b) 1; (c) 0.
|
(a) 1; (b) 1; (c) 1.
|
(a) 1; (b) 1; (c) 0.
|
(a) 1; (b) 0; (c) 0. drill pin-hole.
|
(a) 2; (b) 2; (c) 1.
|
(a) 2; (b) 2; (c) 1.
|
(a) 2; (b) 2; (c) 1.
|
(a) 1; (b) 0; (c) 0. drill pin-hole
|
(a) 8; (b) 8; (c) 4.
|
(a) 4; (b) 4; (c) 2.
|
ON MAY 13, 2013 10:19 PM Ground Station, rover ver 6, clearance, solar power harvesting surface.
Last week was spent in redesign of a rover and ground station. It was done by three main reasons –
a) ground station needs to be build and 3D models for 3D printer has to be finalized.
b) clearance of a rover needs to be increased.
c) needs to increase solar power harvesting surface.
In old design on frame was mounts for flexible solar panels (cells).
each cell has size 25.4 x 63.5 mm = 16 square centimeters.
length (long side )of the frame was 600mm, that gives a place for 23 panels with total surface of 0.037m^2
two frames allows to use 4 times more - total 92 cells, or total .14 m^2
with efficiency of 0.09(9%) and power of a sun per square meter = 1000Wt surface can give 13.35 watt total.
Top side of a rover can harvest some energy, but at the same time bottom did not, it is in shadow – outcome power probably is half == 6 watts.
Solar panels mounted on frame with 45 degree angle.
Problem (b) frame is not flexible, in can stuck on terrain - as a result needs to increase clearance.
Second problem (c) - needs to increase harvesting power surface. (6 watt capable source with 40% efficiency of a 10 watt transmitter, means that transfer of data can be done only in 1 minutes with another 3 minutes 20 seconds waits to harvest the energy.
Flexible solar panel side is 63.5 mm. Assuming it is hypotenuse - that geometrically gives in prime triangle with 45 degree angle, side sizes 45+45mm – which is bigger that original 63mm.
Sizes of each cell will be smaller (like 25.4 mm x 44.9 mm) but on same frame now it is 46 cells with 0.052 m^2.
On 2 frames can be placed 4 groups of 46 cells, calculating same efficiency and same Sun’s power, that surface allows to harvest total 18.8 watt.
Same top-bottom and max outcome power is 9 watt. That is better, than previous design, and it mean == 1 minutes to transmit + 2 minutes to harvest.
Placing additional 2 stand (length 0.5m) which can hold 19x4 = 76 small size panels outcome additional 0.17m^2 with total power 15wt (and half of tit is 7wt). Gives the grand total is 34wt from all surfaces, and halfing gives max s 17 watt power. Which is not bad – on top of each minute to transmit it will require 35 seconds to harvest.
Now the question where to place that two stands. Plus needs to balance weight- i.e - 350grams for each 4 stepper motor, but gear motors can be reduced to have less torque. That can save around 50-100 grams on a motors in favor for stands. Reducing clearance makes smaller dimensions and also saves weight.
Total amount solar cell (sizes 45x25mm) to accommodate on rover today is 168. With weight limit of 400 grams for all system it is 2.3 grams per flexible cell including weights of holders and mounts. Plus needs to account that after 3 week flight cells will lost 1/2 of it power capability. Realistically it will be 8 watt of power at best.
That is a theory – up today Ground station's 3D models probably ready to be ordered, but rover's energy harvesting mounts not finalized yet.
Technical recording. Carbon Fiber. Mold == PVA
Not much to say – was video recoded to analyze what went wrong –
(a) dimensions of a mold – PVA’s 3D printed parts are 0.3 mm tolerance and depend on a tracing process (g-code generator) it has “quantum” effects, small changes and part become 0.3mm longer. That effected round parts. Configuration for 3D printer depend on 3 stretching coefficients. Better they will be equal 1 and then in 3D model will be adjustment of dimensions.
(b) Tolerance of any dimensions has to be remover (because of the reason (a). And adjustment should be done manually.
(c) Carbon “sweater” is little bit thick especially soaked in epoxy – it takes extra pressure to insert. Distance btw walls (in main stepper motor's holder) has to be increased by 1mm.
(d) Final size for stepper motor's holders must be 0.3 mm plus.
(e) Wall for second stepper motor (camera stand/ antenna mechanism) needs to be 15mm low – mistake.
(f) No “impact adsorption” epoxy layer. Stepper motor’s mounts (on rover – not on ground station) needs to be treated by additional layer of epoxy with microspheres to accommodate impact. 203C, 10h baking will be for fully assembled frame.
Technical recording. Frame. Mold assembly. 3D printed parts. PVA.
All parts of mold was 3D printed from PVA. PVA are tricky to print, slow speed are the only choice.As a result mold consists from small parts.Need to glue parts of the mold together. With PVA glue.
But before assembly needs to dry parts - Under Vancouver's weather it is essential step. Bake parts in stove / oven under 50-60C for 1 hour, and cool down inside zip-lock bag. Last step after assembly - PVA from the glue makes mold soft. Reason - water from a glue.
To solve this (it was not recorded on video) mold needs to be placed together with 1 pound of the sugar inside zip-lock bag. Baking in oven also can help.
Now mold ready for composite manufacture -- knotted carbon fiber + epoxy will do job.
Sep 27, 2012.
Molds- molds.
Mold for left tube holder/longeron:

stepper motor sleeve / ceramic bearing holder.

Extrusion of PVA filament is tricky. Filament needs to keep dry – In Vancouver
it is hard – as a result needs to cut “spaghetti” 10-25 m long. Estimation of
required length in 3D printing software kind of sketchy. Mistake in manual
measurement can ether ruin part or extruder can stop thread's extrusion. Loosing
part better then stops – PVA in extruder coagulates if left for a long time
under 190C.
Video capture and remote desktop essential in this case – from cellular phone it
is possible to do check process remotely. Safety is mandatory also – nobody
knows what happen if experimental technology will generate sparks. Pictures in
last posts was not captured intentionally but rather as a "side effect" of 3D
printing process.
Even with all of printer’s problems it is not comparable with time savings.
Sep 25, 2012.

Another part done
Sep 24, 2012.
Funny part of the molds/rover frame creation – first mold was made from alumini
– epoxy and carbon sweeter (knotted) of a frame was placed into mold – first
(93C) and second (163C) cure was done and after process comes the truth – mold
can not be disassembled – all attempts was failed. After some consideration was
chosen method for disassembling combined with a drop test (14 floor in downtown
Vancouver) strata for the building do not allow to throw from the windows
cigarettes butts, but nothing in regulation was about lunar rover. Attempt was
made with precaution under cover of the night, mold with carbon part of a frame
fly freely – and – mold + frame part survived impact solidly.
Fortunately manager of the building was on vacation.
Sep 13, 2012.
Mold creation. Rover and ground station. Material PVA.
... formula” creates cycles of temperature by itself. Period is 1min 20 sec and
period is stable, even after 20 minutes it stay amplitude and period the same.
To repro situation needs to compile firmware (Sprinter) and monitor temperature
in ReplicatorG.
That is definitely creates a problem with extruder in case: for PVA stating flow
point is 180C, starting coagulation temperature for PVA is 200C.
Setting target temperature to 190 makes extruder jam when temperature drop to
180, and melt flow index = 0. That make pressure inside extruder jump, and small
jam created.
On another hand when temperature reach 200 it coagulates PVA on a surface of
extruder (that will contribute to a jam in a future).
Firmware with PID enable (I do not know correct name = “#define PIDTEMP 1”
uncommented) is not capable to print PVA without jam.
To avoid problem need to find solution ether by releasing extra pressure, or by
different algorithm of regulation.
Tried setting HEATER_CHECK_INTERVAL to 50 ms – amplitude of cycle dropped to
+-5C, period was same, and I believe this is not because of algorithm as it is –
just coincidence – when I debugged formula in spreadsheet – in each interval of
time it jumped for one bigger value for heater_duty to a smaller value. With
50ms looks like it just “skipped” some of that jumps. When I tried different
interval it just returned to same +-10C.
For a plastic with more stable parameters of melt flow index (PLA/ABS) that
problem with temperature regulation in extruder is not big impact – plastic
flows in a range of +-10C and index is different but varies not much.
Also checked “regular” temperature control (with PIDTEMP commented out) in this
case when temperature reaches 191 it switches off, when it is low then 190 –then
it is on. Cycles become shorter, and temperature variation is +-5C.
Then I did experiments with heat absorber (I hold heater by big pliers – good
contact and good heat adsorption). PID formula works better then on-off
algorithm when there is a flow of a heat to cold pliers, when pliers become hoot
(around 60 C – temperature when plastic insulation on pliers did allow to
hold it in hands more) then “on-off” started to work better.
Definitely to improve extruder for 3D printer needs to control temperature
better then +-5C.
Task by itself is interesting not only for dissolvable mold creation, but also
for a temperature control in vacuum when only radiation of a heat can be done
for “heat flowing out”.
Aug 15, 2011
- All molds were received – now it is possible to make different thickness’s
springs: 1.5, 2.0, 2.5, 3.0 mm. Also was proved – amount of carbon fiber /
threads inside the mold 1.5mm – 4 layers of carbon fiber , 3 layers and just
threads (no fabric) gives a different linear elasticity (easy to measure – on a
electronic scale – amount of a weight require to connect sides of a spiral).
Also was discovered a mistake (may be not) – if to place the assembled mold into
a vacuum chamber, then excessive air pockets inside mold squeeze out epoxy from
the mold. From one point it is not looking good – a cross section of a spring is
not regular – but spring become lighter 25%, epoxy stays just on carbon fiber,
and coefficient elasticity is bigger. For future use – placing in a vacuum
chamber has to be done before closing the mold (for nice looking springs), or in
assembly (for saving weight and increase elasticity). Also was found a way to
speed up process (up to 5 times) and make spring more controllable. The carbon
fiber has to be in a braid (or a plait) – different technique can be used to
control amount of fiber and physical dimensions (5 x 1.5 x 500 mm). Also was
proved an easy way to control temperature for epoxy’s curing – the induction
stove from Wal-Mart + connecting sequentially to a thermo element variable
(10-100kOm) resistor and the temperature inside the chamber wrapped by aluminum
foil will stays with 1 centigrade range.
June 31, 2011 –
Tweaking of a mold was done. Download to www.shapeways.com and ordered.
view
eDrawings : Top mold (Will install plug-in):
and view bottom mold (Will install plug-in):
download top mold,
bottom mold
or pictures:
June 30, 2011 – Parts
for a rover springs arrived. Plastic is not strong enough to assemble into
wheel. This is okay as I wanted to see how it molds – final goal is to make same
part(s) from carbon fibre). Nice idea to use 3D printing but for a spring
fabrication it is not good. Needs to make 3D mold directly – without step.
Nothing available to withstand thermal curing of epoxy. Another solution is to
use aluminum – the same 3D printing but with aluminum powder – should work at
175C – more than enough for composites. Spring model needs to rebuild to make 2
mold parts.
This page will be updated step by step to reflect current changes on recent
drawing in attempts to resolve problems appeared in design.
Blender files (http://www.blender.org/download/get-blender/)
design (.blend+.STL) available here:
teamPlamB_vehicle.zip
View in SolidWorks eDrawings here. (Will install plug-in).
SolidWorks assembly and parts here.
Wheels.
Each wheel has 16 springs made from ether steel or carbon fiber. For steel’s it
is easy to cut parts from specific thickness steel’s sheet and a coefficient for
a Hooke’s formula can be easily calculated, confirmed and controlled. More
complicated process is for a carbon fiber – all controllable parameters are:
amount of layers (one, two, or three layers of a carbon fabric), different types
of fabrics, and amount of epoxy (solo depends on a mold). The mold manufacturing
can be complicated. It was chosen simple approach - first made 3D model and
manufacture it at
http://www.ponoko.com/showroom/TeamPlanB.
Ponoko Personal Factory can produce parts
from a plastic which
is not suitable for real springs, but plastic model can be used in mold’s
manufacturing. Then depend on a results (carbon fiber’s molding or steel springs
cutting) can be made the decision: what is better. Arbitration are usual - the
weight of a wheel, and performance at different temperature.
Absence of a gear will definitely reduce weight, and springs flexibility can
accumulate energy to get momentum enough to travel over obstacles and craters.
Coefficient in Hooke’s formula and precision delivery of a momentum by stepper
motors can pump energy (using resonance frequency) into a spring system and a
precision release (of that energy) will give jump’s type of movement for a
vehicle. This is similar to a case (if somebody remember from past) when a car
was stuck with one wheel at road’s hole, small oscillation (as back and forward
push) can release car from road trap. From another hand (to save a time) plastic
parts can be used in vehicle software development, some flexibility plastic
already has and braking plastic part can be indication of a proper resonance
frequency achieved. Wona stated that this type of wheel is totally useless
without controllable software.
Problems to solve:
1. How to make a mold for carbon fiber? Probable answer on that question – to
order mold from Ponoko Personal Factory from
a different type of ceramic/plastic.
2. How to control stiffness/elasticity for carbon fiber’s springs? Probable
answer – to make 3 different (1, 2, or 3 layers of carbon springs) and use
combination of different springs.
3. How to calculate resonance frequency of a wheel in assembly? Probable answer
– software has to be adaptable to different parameters/performance of springs.
4. Does it require carbon fiber clothe wrap to make bigger surface contact?
Answer for this question is actually: Yes.
Frame design.
Two aluminum tubes with two aluminum plates, holding tubes in parallel each
other. Four stepper motors – two for wheels and two for antenna and camera
stands. Motors for a camera stand and antenna was to be with gears with 1.x 3.5
ratio. Gears made from ceramic grinding tools. On a frame mounted an aluminum
mirror allowing HD camera’s observation in holding position.
On tubes mounted flexible solar panels with angles 45 degree to the center line,
solar panels both sides orientation. Aluminum frame and plates will be warped
for enforcement with one or two layers for a carbon fiber. This will create
“composite’ structure – soft aluminum inside and carbon fiber outside. For
mockup all parts except aluminum tubes ordered over
Ponoko Personal Factory.
Problems to solve:
1. Needs to find flexible solar panels with good performance in temperature
range and under of UV and high charged particle’s conditions. Current power
capabilities are around 6-8 W total, and in real life it will be around 2.5-3 W.
Which is not acceptable at all.
2. Orientation for aluminum mirror.
3. What a minimum distance from solar panel to antenna’s stand rotation axe? (To
avoid interference with communication).
4. Protection for a two wheel’s stepper motors. Originally wheel’s springs will
give protection at impact, but looks like it will require some carbon’s
bees’-comb-like structure for additional protection.
5. Manufacture molds for top layer of carbon fibber (for tubes and plate) with
bees’-comb-like pattern.
6. Protection for a gear needs to be decided.
7. Holder for solar panels does not look good.
Antenna stand.
Helical antenna for a 2.4 GHz communication made from fiberglass. Conductor for
antenna glued inside the cylinder body of antenna.
Reflector made from cuprum layer on top of carbon fiber layer. On the same stand
connected hermetically sealed box for main HD camera (requirements for
competition), capacitors, and main onboard computer. HD camera has a window
allow observation from a low point at vehicle movement.
In holding position box with camera located near aluminum mirror, this will give
additional angle of observation. Box has solar panes glued to its sides. Solar
panels are different type than mounted on a frame.
Problems to solve:
1. Center of mass of antenna stand needs to be “adjustable”.
2. Which direction to shift center of mass (related of a rotation axe) to the
antenna side or to the main sealed box?
3. How to clean lunar dust from exterior surface of the antenna?
4. Power wire’s elasticity should be accounted at antenna stand movements.
Camera stand.
Made from carbon fiber. On top of stand 4 cameras mounted with pinholes. Two
sealed boxes for capacitors. All parts for mockup ordered from
Ponoko Personal Factory.
Problems to solve:
1. Center of mass needs to be “adjustable”.
2. Does it require two or (for bigger space) one sealed box?
3. Power wire elasticity has to be accounted as for antenna stand.
Air bag system.
For now was consider “composite” design with 3 layers – polyurethane internal
shell to cover vehicle, inflatable foam, and thin film bag. Shell cut in
sections and each section glued to deflated airbag with deflated foam inside the
bag. Two similar “big” air bags covers top and bottom of vehicle, and two
“smaller” on a sides. In deflated position bags formed two “rolls”, with
polyurethane side out. To protect polyurethane from UV exposure cut sides can be
covered with thin flexible solar panels, which can give additional power at
orbit flight. After landing the solar panels (if survived and recovered) can
give source for additional power. Inflation of bags has to be in sequence – top
and bottom bags will be inflated first then side bags after. Inflating of bags
will unroll internal polyurethane shell and covers top and bottom of the
vehicle, then side shell covers and hold all polyurethane’s shell together. Time
for inflating foam inside bags needs to be considered. Impact will eventually
tear sections from bags and released vehicle on the Moon. For now it is just an
idea – no experiments was done yet, except attempt to make internal shell (video
available here soon).
Problem to solve:
1. inflating foam time has to be calculated.
2. Performance of polyurethane at vacuum has to be checked.
3. Deterioration of polyurethane under UV and temperature needs to consider.
4. Foam – where to get such?
5. Bags has to be deflated with good vacuum pump, otherwise needs to do drainage
at orbit (at vacuum).
6. System for inflating totally not designed yet.
Assembly.
Problem to solve for all assembly:
1. Center of mass in plane perpendicular to a frame has to be projected to
wheel’s axes.
2. Air bag mount is not finalized et.
3. 4 solid state engines mounts is not finalized yet. It has to be connected
somehow to frame of the vehicle.
4. Small thrust engine mount is not finalized yet. The same problem – mount must
be on vehicle frame.
5. Release mechanism of each solid state / small thrust engines - no idea how to
do this. Pyrotechnic fastener definitely is out of questions.
Apr 24, 2011. In a
vehicle development was discussed an idea to use carbon fiber instead of
springs (steel springs can eliminate vehicle’s gears). Made from carbon fiber
spring needs to be longer and spiral shape. Question how to make a mold for that
shape. Was tried bent aluminum sheet to make desirable shape. 6 hours was spend
in experiments with no results. Mold for a wheel’s mount does not bent properly.
Total 64 springs needs to be manufactured for two wheels, one or two can be, but
it can be a challenge to make 64 exact molds.
Mar 20,2011. Spring
tension wheel was build for a testing.
Mar 19, 2011. Concern
about molybdenum lubricant on stepper motors. Momentum 0.35Nm will be in this
case smaller. Needs (at least for wheels) to use something to increase momentum.
Previous decision not to use gears make it difficult. Was considered to use some
alternative way. If wheels-motor system will be controlled by microprocessor it
is possible to use springs and controllable oscillation movements to accumulate
energy in springs with releasing it (energy). Oscillation frequency in this case
can be controlled by springs parameters.
Mar 18, 2011. thermal
test -78C. All test passed as expected. It is good. Counting that previously all
materials was passed thermal test +125C just give confidence in all approach.
Soldering with tin also passed test, but needs to repeat that test for 10
thermal circles +125 -78 to make sure that for onboard electronics possible just
simply to place manufacturing order. Test was filmed. In low temperature better
performance was stepper motor without lubricant at all. Molybdenum lubrication
was working but it was feasible that it has more traction.
Mar 17, 2011, was spent
in preparation for thermal tests -78C. All materials was sorted, stepper motors
was disassembled. Two stepper motor’s bearing was cleaned from lubricant, one
was left as it is (this will confirm failure of a test). From two cleaned
stepper motors in one was applied molybdenum lubricant. Main onboard computer
was checked - no electrolytic capacitors was on board, microprocessors,
insulation material, aluminum tapes, carbon fibre components was selected,
antenna itself, wires, soldering components with tin soldering and etc.
Mar 14, 2011. Sketches
for 2 wheels design and two type of movements was made. Next step to create
drawing. Considered making composite structures. Parts for carbon fibre
production will be done in two steps – first manufacturing at Ponoco plastic
component, then testing, then using same component to make forming using
modeling ceramic. Aluminum parts will be manufactured as it was done previously.
Mar 1, 2011. 16 month
left to a dead line. Was made an attempt to re-design a vehicle. It will be
third re-design from beginning of a project, first was flip-flop movement with 5
segments, 8 stepper motors and 2 segments/antennas used as helping legs. First
design was abandoned because of a high weight (8kg) and sophisticated gears
which does not reduce mass also.
Was considered different methods of movement
http://www.romela.org/main/Robotics_and_Mechanisms_Laboratory. All was
looking good and promising, but requires complicated mechanical parts movable
against each other in vacuum, survivability of drop also will be a question.
Mechanical engineers suggest some throw-anchor-on-a-rope schema. Arguments are -
if fishing line made of monofilament or fluoropolymer .36mm dia. (http://en.wikipedia.org/wiki/Fishing_line)
can catch 10kg salmon then 6kg vehicle will be ok, unsolved questions will be-
what does to hook-on on the moon, and - how does to throw the line (line
actually can be attached to a ejected part of a craft – 501m line guarantees
distance, but need to demonstrate travel ability in a desired direction - 2 or 3
lines needs to use).
Some discussion was made about magnetic gears
http://www.magnomatics.com/technology/magnetic-gears.aspx and
http://www.mgt.com.au/ .
Complications with steel's gears in vacuum can be solved using MoS2 lubricant:
http://en.wikipedia.org/wiki/Molybdenum_disulfide
If a re-design will be made needs to consider also weight reduction.
Feb 27, 2011.
Calculation for a drop test was made to avoid complicated computer simulation.
Errors for break engine shut-off will be 1% (ideally 25 kg of a solid state fuel
can be dozed with 0.001kg precision, and engine can be calibrated by test firing
after 1 month expose in vacuum chamber). This gives error (brake engine needs to
slow down from around 2000m/s) a final speed 20m/s with an equivalent of a drop
from 130m at a Lunar gravity. That 130m gives 12s for air bags inflation, with
(same as error in solid state engine) landing speed 20m/s. The kinetic energy
(for a vehicle mass 6kg) equal to 1271 kg*m^2/s^2. In Earth gravity this is
equivalent of a drop the same mass (6kg) from 22m, or from 9 floor of a building
(needs to check – does a strata agreement prohibited to throw space craft from a
windows with safety precautions? And does the fine will be in reasonable
limits?). To test air-bags will be possible to use the bigger mass dropped to a
fixed platform. With a mass 100kg to get the same kinetic energy (1271) will be
enough to achieve speed 5m/s, and 5m/s in Earth gravity equivalent of drop from
1.5 m. Swing can be used.
Feb 26, 2011. The
simple drop test highlights that the ceramic gear did not survive. Critical part
is aluminum holder for silicon carbide gear (ceramics was actually ok, aluminum
shaft was bended). Ether needs to reinforce, or to protect gear’s part. Also was
loose an aluminum connector to a motor’s steel shaft, and loose for the same
reason a ceramic gear’s shaft. All assembly also looks like would not survive
any significant drop test. As a result, after talking to engineers, current
design of a vehicle definitely will not fly the Moon, but probably will be ended
in the nearest garbage can. It is pity – weight for a 3 wheel design was in a
limit of 6 kg.
Feb 17,2011, mount #1
,
mount#2, mount#3
, mount#4, adapter#1, adapter#2
was
made. An assembly video is
here
. After a meeting with experienced mechanical engineers was raised a
question – may be a steel gear will be fine – even in a deep vacuum distance and
sessions requirements does not sounds so tough, this will be allow to forget
ceramic gears. May be. Engineers promised to look at a performance of a steel
alloys and come with some materials.
Feb 16, 2011, Best
results on 646 – see the picture ,
510 kind of fragile,
672 is easy to solidify.
646 and 510 are water based and placing in a vacuum chamber is working fine for
bobbles removal. Surface with 510 was rough, 646 creates nice fine perfect
surface. Chemical additives in 589 and 672 creates a problem - in the vacuum
chamber bobbles was not gone and a surface of a casting ceramic did not fill a
model’s mold. Mix was a bad idea. It does not produce anything useful.
Feb 12, 2011. An
attempt was made to design ceramic gear from available source of ceramic’s
tools. First candidate are grinding sets. It has a round shape, it has roughs
surface, and it is made from strong ceramics (manufacturer usually does not
specify chemical formula). Design is here. Gear will be open - dust will help
keep traction between rotating parts. Bearings will be simple – ceramic
available from a catalogs. To keep tension between gears spring will be applied.
Spring will bend aluminum mount and will bend to accommodate gear’s wear.
Gear
consist of 8 major parts: Mount#1,
Mount#2, Mount#3,
Mount#4, Gear#1,
Gear#2, Adapter#1,
Adapter#2
Feb 11, 2011. Results
for castable ceramics 672, 510, 646 and its 50/50 mixture will be available
later (on Feb 14-15) – it takes 24 hours to dry the ceramics and temperature
cure steps. Another problem – 0.2mm precision limitation will require tailoring
a final model to snooze imperfections. The last (and major) problem where to
order ceramics like silicon nitride, silicone carbide, or zirconium dioxide even
as cast for gear was manufactured.
Feb 10, 2011. Results
for 646 vary. Cast produced from silicon was the best cast. Using a wax is
complicated at the moment of separating model from a solidified wax. Silicon
gives a nice snooze surface but it is difficult to extract model from a cast (it
sticks to plastic unpredictable despite applying oil/liquid soap to a model).
Ceramic can be cured in 200F for 3-6 hours in regular oven.
Feb 9, 2011, Vehicle
design. Three wheels powered by stepper motors gives the vehicle the ability to
travel and to steer in desired direction. However, positioning an antenna to the
Earth direction will require something more than a stepper motor’s steps with
1.8 degree precisions. Obvious way to do this is to use some ceramic gearbox
(one at least). An attempt was made to understand how to build such a gearbox.
Was investigated way to manufacture cast for planet and sun gear in Epicyclic
gear type (http://en.wikipedia.org/wiki/Epicyclic_gearing). Such gear’s type was
chosen to make it as small as possible and as a result to achieve light weight.
For a cast manufacturing it is handy to use a 3D printers
(http://www.makerbot.com/, http://www.botmill.com/, http://www.pp3dp.com/, or a
more easy way - to make 3D model design in .stl format, upload it to personal
factory like http://www.ponoko.com/, and build plastic cast with precision 0.2
mm (the same technique can be used to create a cast for any carbon fiber
elements of a probe/vehicle). Castable ceramic like 672, 510, 646, 589 can be
used.
Vehicle design:

And vehicle picture:

reference:
planb.sldasm
B18.2.4.1M - Hex nut, Style 1, M3 x 0.5 --D-N.SLDPRT
B18.22M
- Plain washer, 3 mm, regular.sldprt
B18.6.7M - M3 x 0.5 x 25 Type I Cross Recessed PHMS --25N.SLDPRT
B18.6.7M
- M3 x 0.5 x 35 Type I Cross Recessed PHMS --35N.SLDPRT
bearing_0006.SLDPRT
fr_0001.SLDPRT
fr_0002.SLDPRT
fr_0003.SLDPRT
fr_0004.SLDPRT
fr_0006.SLDPRT
fr_0007.SLDPRT
fr_0008.SLDPRT
fr_0009.SLDPRT
fr_0010.SLDPRT
fr_0011.SLDPRT
fr_0012.SLDPRT
fr_0014.SLDPRT
fr_0015.SLDPRT
fr_0016.SLDPRT
SW3dPS-037.SLDPRT
Vehicle’s design required to withstand impact equivalent of a free fall from 125m
on Earth with airbag protection. This will give easy guidance for an air-bag/vehicle
ground test. Airbags deflates after landing.
Vehicle consist of a
flat frame with 3 wheels, double sized two solar panels, horizon/azimuth
orientation helical 2.4GHz antenna,
sealed boxes/frame with computer equipment.
Vehicle is capable of travel despite which side of frame is up. Making length longer
than width will give capability to flip on side/edge of terrain, such move can be
useful to switch vehicle's sides and as a result to switch solar panel with deteriorated
performance.
All 3
wheels before landing will be in “transportation” position to make vehicle
“flat”. After landing wheels will be moved to “working” position by springs – this
will give flexibility for each wheel as a result third wheel can be dynamically
oriented to support steering of vehicles. Radius of each wheels 125mm, antenna length
650mm, total length 1050mm. Movement of antenna
and its equipment / contr-weight
box can give additional
steering / movement / digging capabilities.
Vehicle require 5 stepper motors to move and to orient antenna. All motors are regular
stepper motors, changing bearing to ceramic may be required depend on performance
at vacuum/temperature chamber's testing. Wiring on motors will require to withstand
temperature range from -100C to +115C. Desired direction and distance will be sent
to probe from ground control. On vehicle movement acceleration and azimuth from
sensors and motors performance will be accumulated together with solar panel efficiency
and on-board computer that will make decisions to achieve desired point of travel.
Upon arrival or upon time-out vehicle will orient antenna for communication session
with earth ground control. For 500m travel 50-75 sessions will be required. At session
point/time telemetry will be transferred first, then low resolution pictures from
camera. Then decision to transfer high resolution video/picture will be made. Transferring
HD images will give time to make a decision of next movement of probe. All travel
should be performed during 168 hours (7 days) upon landing. This will gives 2 hour
time between sessions with a travel time 20-25 minutes per session.