Transcript Building competitive manipulators Greg Needel DEKA R&D, Rochester Institute of technology Owner, www.midnightinvention.com Mentor teams: 131, 1511

Building competitive
manipulators
Greg Needel
DEKA R&D, Rochester Institute of technology
Owner, www.midnightinvention.com
Mentor teams: 131, 1511
Strategy, Strategy, Strategy!
•
•
•
•
Read the rules
Outline the game objectives
Look for the “gimmie” robot design
Try small simulators
• Whatever you choose STICK WITH IT!
Types of Manipulators
•
•
•
•
•
Articulating Arms
Telescoping Lifts
Grippers
Latches
Ball Systems
Arm: Forces, Angles & Torque
• Example #1 - Lifting
– Same force, different angle,
less torque
10 lbs
10 lbs
D
<D
Power
• Power = Force x Distance / Time
OR
• Power = Torque x Rotational Velocity
Power (FIRST def.) – how fast you can
move something
Arm: Power Example
– Same torque, different speed
10 lbs
0.1 HP, 100 RPM
Motor w/ 1”
sprocket
10 lbs
0.2 HP, 200 RPM
Motor w/ 1” sprocket
OR
100 RPM w/ 2”
sprocket
Arm Design
• “Arm”: device for grabbing & moving objects
using members that rotate about their ends
• Think of your materials (thin wall is good)
• Every Pivot has to be engineered (less is more)
• Linkages help control long arms.
• Use mechanical advantage (it is your friend)
• Think of the drivers (pivots on pivots are hard)
• Operator Interface (keep this in mind)
Arm Advice
• K.I.S.S. doesn’t mean bad
• Feedback Control is HUGE
– Potentiometers, encoders, limits
– Automatically Take Action Based on Error
– Design-in sensors from the start of design
• Think outside the box.
• Off the shelf components are good
(andymark.biz, banebots.com )
Four Bar Linkage
•Pin Loadings can be very high
Watch for buckling in lower
member Counterbalance if you can
Keep CG aft
4 bar linkage example :229 2005
Arm Example: 234 in 2001
Arm Example: 330 in 2005
Arm Example: 1114 in 2004
Telescoping Lifts
• Extension Lift
• Scissor Lift
Extension
Extension Lift Considerations
• Should be powered down AND
up
– If not, make sure to add a device
to take up the slack if it jams
• Segments need to move freely
• Need to be able to adjust cable
length(s).
• Minimize slop / free-play
• Maximize segment overlap
– 20% minimum
– more for bottom, less for top
• Stiffness is as important as
strength
• Minimize weight, especially at
the top
Extension - Rigging
Continuous
Cascade
Extension: Continuous Rigging
• Cable Goes Same
Speed for Up and
Down
• Intermediate Sections
sometimes Jam
• Low Cable Tension
• More complex cable
routing
• The final stage moves
up first and down last
Slider
(Stage3)
Stage2
Stage1
Base
Extension: Continuous Internal Rigging
• Even More complex
cable routing
• Cleaner and protected
cables
Slider
(Stage3)
Stage2
Stage1
Base
Extension: Cascade Rigging
• Up-going and Down-going
Cables Have Different
Speeds
• Different Cable Speeds Can
be Handled with Different
Drum Diameters or Multiple
Pulleys
• Intermediate Sections Don’t
Jam
• Much More Tension on the
lower stage cables
Slider
(Stage3)
Stage2
Stage1
– Needs lower gearing to deal
with higher forces
Base
• I do not prefer this one!
Team 73 in 2005 elevator
Scissor Lift
Scissor Lift Considerations
• Advantages
– Minimum retracted height - can
go under field barriers
• Disadvantages
– Tends to be heavy to be stable
enough
– Doesn’t deal well with side
loads
– Must be built very precisely
– Stability decreases as height
increases
– Loads very high to raise at
beginning of travel
• I recommend you stay away
from this!
Team 158 in 2004
Arm vs. Lift
Feature
Arm
Lift
Reach over object
Yes
No
Fall over, get back up
Yes, if strong enough
No
Go under barriers
Yes, fold down
No, limits lift potential
Center of gravity (Cg)
Can move it out from
over robot
Centralized mass
over robot
small space operation
No, needs swing room Yes
How high?
More articulations,
more height (difficult)
More lift sections,
more height (easier)
Complexity
Moderate
High
Accumulation
1 or 2 at a time
Many objects
Combination
Insert 1-stage lift at
bottom of arm
<-
Braking: Prevent Back-driving
• Ratchet Device - completely lock in one direction in
discrete increments - such as used in many winches
• Clutch Bearing - completely lock in one direction
• Brake pads - simple device that squeezes on a rotating
device to stop motion - can lock in both directions
– Disc brakes - like those on your car
– Gear brakes - applied to lowest torque gear in gearbox
• Note : any gearbox that cannot be back-driven alone is
probably very inefficient
Power
• Summary
– All motors can lift the same amount
(assuming 100% power transfer efficiencies) they just do it at different rates
• BUT, no power transfer mechanisms are
100% efficient
– Inefficiencies (friction losses, binding, etc.)
– Design in a Safety Factor (2x, 4x)
Grippers
•
•
•
•
•
Gripper (FIRST def) grabbing game object
How to grip
How to hang on
Speed
Control
How to grip
• Pneumatic linkage grip
– 1 axis
– 2 axis
•
•
•
•
Motorized grip
Roller grip
Hoop grip
Pneumatic grip
Pneumatic linear grip
• Pneumatic
Cylinder
extends &
retracts
linkage to
open and
close gripper
• 254 robot:
2004, 1-axis
• 968 robot:
2004, 1-axis
Recommended
Pneumatic linear grip
• Pneumatic
Cylinder,
pulling 3
fingers for a
2-axis grip
• 60 in 2004
Recommended
Motorized Linear Grip
• Slow
• More
complex
(gearing)
• Heavier
• Doesn’t use
pneumatics
• 49 in 2001
Not
recommended
Roller Grip
• Slow
• Allows for
misalignment
when grabbing
• Won’t let go
• Extends object as
releasing
• Simple
mechanism
• 45 in 98 and 2004
Recommended
Hoop grip
• Slow
• Needs
aligned
• Can’t hold
on well
• 5 in 2000
Not
recommended
Pneumatic Grip
• Vacuum:
– generator &
cups to grab
• Slow
• Not secure
• Not easy to
control
• Simple
• Problematic
Not recommended
Hang on!
• Friction: High is needed (over 1.0 mu)
– Rubber, neoprene, silicone, sandpaper
• Force: Highest at grip point
– Force = multiple x object weight (2-4x)
– Linkage, toggle: mechanical advantage
• Extra axis of grip = More control
Speed
• Quickness covers mistakes
– Quick to grab
– Drop & re-grab
• Fast
– Pneumatic gripper
• Not fast
– Roller, motor gripper, vacuum
Grip control
• Holy grail of gripping:
– Get object fast
– Hang on
– Let go quickly
• This must be done under excellent control
– Limit switches
– Auto-functions
– Ease of operation
Latches
• Spring latches
• Hooks / spears
• Speed & Control
Latch example: 267
• Pneumatic
Latch
• 2001 game
• Grabs pipe
• No “smart
mechanism”
Latch example: 469
• Springloaded
latch
• Motorized
release
• Smart
Mechanism
• 2003
Latch example: 118
• Springloaded latch
• Pneumatic
release
• Smart
mechanism
• 2002
Latching advice
• Don’t depend on operator to latch, use a
smart mechanism
– Spring loaded (preferred)
– Sensor met and automatic command given
• Have a secure latch
• Use an operated mechanism to let go
• Be able to let go quickly
– Pneumatic lever
– Motorized winch, pulling a string
Ball Systems
• Accumulator = rotational device that
pulls objects in
• Types:
– Horizontal tubes - best for gathering balls
from floor or platforms
– Vertical tubes - best for sucking or pushing
balls between vertical goal pipes
– Wheels - best for big objects where alignment
is pre-determined
Conveying & Gathering
• Conveyor - device for moving multiple
objects, typically within your robot
• Types:
– Continuous Belts
• Best to use 2 running at same speed to avoid
jamming
– Individual Rollers
• best for sticky balls that will usually jam on belts
and each other
Conveyors
Why do balls jam on belts?
- Sticky and rub against
each other as they try to
rotate along the conveyor
Solution #1
Use individual rollers
Adds weight and complexity
Solution #2
Use pairs of belts
Increases size and
complexity
Solution #3
- Use a slippery material for the
non-moving surface (Teflon
sheet works great)
Roller example: 188
Accumulator example: 173 & 254
Questions?
Thanks to:
Andy Baker (45)
www.chiefdelphi.com
www.robotphotos.org
www.firstrobotics.net
www.firstrobotics.uwaterloo.ca
Extra Stuff
•
•
•
•
•
Pneumatics vs. Motors
Materials
Shapes / Weights
Fabrication processes
Environment
Pneumatics vs. Motors
Some, but not all important differences
• Cylinders use up their power source rather quickly
• the 2 air tanks we are allowed do not hold much
• Motors use up very little of the total capacity of the battery
• Cylinders are great for quick actuations that transition to
large forces
• Motors have to be geared for the largest forces
• Our ability to control the position of mechanisms
actuated by cylinders is very limited
• We are not given dynamic airflow or pressure controls
• We are given much more versatile electronic controls for motors
• Since air is compressible, cylinders have built-in shock
absorption
• Cylinders used with 1-way valves are great for
Armageddon devices - stuff happens when power is shut
off
• This could be good or bad - use wisely
Materials
•
•
•
•
Aluminum, thin-wall tubing
Polycarbonate sheet, PVC tubing
Fiberglass (used rarely, but advantages)
Spectra Cable
– Stronger than steel for the same diameter
– Very slippery
• Easy to route
• Needs special knots to tie
– Can only get it from Small Parts and select other suppliers
• Pop Rivets
– Lighter than screws but slightly weaker - just use more
– Steel and Aluminum available
– Great for blind assemblies and quick repairs
Shapes
• Take a look at these two extrusions - both made from
same Aluminum alloy:
– Which one is stronger?
– Which one weighs more?
1.0”
0.8”
1.0”
Hollow w/ 0.1” walls
0.8”
Solid bar
Shapes, cont.
• The solid bar is 78% stronger in tension
• The solid bar weighs 78% more
• But, the hollow bar is 44% stronger in
bending
– And is similarly stronger in torsion
Stress Calculations
• It all boils down to 3 equations:
Bending
  Mc
I
Where:
 = Bending Stress
M = Moment (calculated earlier)
I = Moment of Inertia of Section
c = distance from Central Axis
Tensile
 tens 
Ftens
A
Where:
 = Tensile Stress
Ftens = Tensile Force
A = Area of Section
Shear
 
Fshear
A
Where:
 = Shear Stress
Fshear = Shear Force
A = Area of Section
Structural Shapes
• I am willing to bet that none of our robots are
optimized with respect to strength to weight
ratios
– We all have more material than we need in some
areas and less than we need in others.
– It would take a thorough finite element analysis of our
entire robot with all possible loading to figure it all out
– We only get 6 weeks!!
• But, this does not mean we cannot improve
Fabrication Processes
• Laser cutting causes localized hardening
of some metals
– Use this to your benefit when laser cutting
steel sprockets
• Cold forming causes some changes in
strength properties
– Some materials get significantly weaker
– Be aware of Aluminum grades and hardness's
• Welding - should not be a problem if an
experienced welder does it