Week 9 Detailed Design Review

badwaterreflectiveUrban and Civil

Nov 29, 2013 (3 years and 8 months ago)

208 views

Week 9 Detailed Design
Review

P13211
-

Rimless Wheel (Wired)

Customer needs

Customer Needs

Importance

Detail

CN1

1

Collect data to prove periodic motion

CN2

2

Collect data on current/voltage of battery

CN3

1

Record angular velocity of wheel and frame

CN4

1

Record relative angle between wheel and frame

CN5

2

Collect data on current/voltage of actuators

CN6

1

Attain periodic motion

CN7

1

25 steps ("infinite walking distance")

CN8

2

Resolution of 100 Hz

CN9

1

Minimize energy loss

CN10

1

Portability

CN11

1

Cost of Transport (0.1 or 0.05)

CN12

2

Rigidity/durability

Engineering Specs

Risk Management

Risk
ID

Risk Item

Effect

Cause

Likelihood

Severity

Importance

Mitigation
Measure

Owner

1

Lead times and
manufacturability

Delay on
construction/tes
ting

Late order/lack of
knowledge

3

3

9

Plan ahead and
gather as much
information as
possible

Maddy

&
Dan

2

Discrepancy between
simulation and reality

Inaccurate
design



3

3

9


Account for as
many flaws as
possible

Dan

3

Effectiveness of
actuation

Unable to
sustain motion

Inaccurate design

3

3

9

Be thorough in
engineering
analysis, and
leave room for
adjustment of
design
parameters

Hao

4

Cost vs. quality





2

3

6

Thorough
analyses of
components

Owen &
Becky

5

Technical errors

Inaccurate
design

Human errors

3

1

3

Double check
analysis

All

Risk Management

Risk
ID

Risk Item

Effect

Cause

Likelihood

Severity

Importance

Mitigation
Measure

Owner

5

Balancing weights
of new components

Unable to
maintain
periodicity

Inaccurate
design

1

3

3



Hao

8

Lag time of controls

Ineffective
actuation

Delay in signal
transmission

3

1

3



Maddy

2

Material flaws

Delay on
construction/tes
ting

Transportation/hu
man error

2

1

2

Plan ample time
for construction
and be careful
with materials

Owen

6

Accuracy of
sensors

Unable to
achieve
desired time
resolution



2

1

2


Double check
Data Sheets

Becky

Detailed Block Diagram

Old Design (motor to shorten
string)


motor would rotate to shorten the string, and
increase the extension of springs


Based on simulation, we would be using K=5
N
-
m/rad & 1*pi rotation for initial condition


This equates to ~15.7 N
-
m or ~139 in
-
lbs


For our design, we would need to apply a
torque greater than this at approximately 400
RPMs

The Problem


400 RPMs at 15.7 N
-
m of torque is ~657
Watts


With a cost of transport of .1, using our
frame design weights and distance traveled,
and assuming 1 second step time, we would
be able to use 1.59 watts per step


The Problem (cont'd)


Assuming the following (untrue):

o
motor has speed up time of 0 seconds

o
at full torque, motor will run at full RPMs

o
the sensors and all electronics use no energy

o
the clutch system uses no energy

we can only actuate this for 2.4 thousandths
of a second


Over this time period, we would only be able
to rotate our motor .016 revolutions, much
smaller than we were aiming for


We need to change something

The Fix


Decided to go with our initial idea of
attaching the motor axle to the bike wheel


At the beginning of MSD1, we could not
figure out a way of doing this, because we
had to go through the axle to do this


13212 (Wireless team) provided the solution
of rotating the entire axle


This change was extremely beneficial

o
Required the change of 2 parts, addition of 1 sleeve,
and addition of 2 bearings

o
allowed for the removal of 13 parts and simplification
of 2 more parts

Equation of Motion

Single Stance

Equation of Motion

Double Stance

Computational Simulation

Equations of Motion


Actuation: In single stance Angular
speed of the wheel Add a constant
torque




Reaction forces changed

Maintain double stance during actuation

Simulation Results

Control Algorithm

Frame Plates


Carbon fiber over foam


Order all materials from

Noah's Marine Supply


Machine shop will cut out design


We will lay carbon fiber and resin


Very rigid


Holes for plastic inserts so we do not crush
foam in compression (from fasteners)

part number: 1
-
4

Plastic Inserts




Self made
-

Delrin


Lightweight & Rigid


Purpose is to keep fasteners from crushing
foam when tightened


10 of the small ones (on the left), one for
each side of the braces


1 of the large one (on the right), for the
mounting plates on the bike wheel side

part number: 34 & 35

Brace Assembly


Thin walled steel tubing


Aluminum insert press fitted into tube


Thread screw into aluminum insert to attach
to frame plates


Provides rigidity to frame


Tubing from McMaster

Carr/inserts from

Machine shop or

McMaster Carr

part number: 11 & 38

Brace Assembly

Calculations


Calculated for bending


and shear of tubing


Worst case: one frame would see 12.6 N
-
m
or 115.5 in
-
lbs of torque


Spreading that out over 5 braces, each
brace would see (12.6 N
-
m)/[(.3556 m)*(5
braces)] = 7.09 N or 1.59 lbs


This force would result in a flex of 0.0682
deg (0.032 in) over the length of a tube


This results in 37 Mpa of stress, but failure
would not occur until over 250 Mpa

part number: 11 & 38

Fasteners


Free from Machine shop


1/4"
-
20 x 1" allen wrench cap screws


1/4"
-
20 hex nuts


1/4" washers


usable for almost all applications (if
unusable, simply get a large size)


current design calls for:

o
20 cap screws

o
40 washers

o
10 hex nuts

part number: 15
-
17

Mounting plate

(motor side)


Aluminum
-

machine shop


5, 1/4" holes to mount to frame


Designed to reduce the chance of crushing
the plates with our fasteners


Machine shop has said this will be an easy
job


can be relatively flimsy as it is not seeing
anything other than compression

part number: 5

Mounting Plates

(bike wheel side)


Aluminum
-

Machine shop


Two purposes:

o
Press fit the bearing into the left mounting plate

o
Prevents the fasteners from crushing the foam plate


Similar dimensions, except for the right plate
has a slightly smaller hole, to better house
the bearing (lip will cover bearing by .075
inches)


can be relatively flimsy as it is not seeing
anything other than compression

part number: 5 & 6

Axle (for bike wheel)


Aluminum
-

self made


Bike wheel rigidly fixed onto axle

o
possibly press fitted

o
possible clip depending on bike wheel we use


Threaded end to attach to sleeve (to motor)


Additionally, bearing sleeve will be pin set
onto axle

part number: 10

Axle (for bike wheel)

Calculations


Assume worst case


All torque in wheel is now in frame at time of
collision


Max speed of frame and wheel is 1.29 m/s


Assuming .001 meter impact distance, frame
would see (1/2)*m*v^2/s = 2.28 kN


Our axle can handle
(25.5e9)(pi)(.009525)^2/4 = 1817 kN in
shear

part number: 10

Axle sleeve

(to bearing)


Aluminum
-

self made or Machine shop


Press fitted into bearing


Set pinned onto axle


Allows for easy disassembly of frame if
required

part number: 39

Axle (Hollow)


Aluminum
-

custom made


Houses motor


1/4" holes for mounting to frame


7/8" hole for housing bearing


3" diameter, though may reduce size
depending on motor size


Encoder bolted to end (holes not shown in
above CAD drawing)

part number: 7

Bearings (for axle)


1 for 3/8" axle (sliding onto axle)


1 for 5/8" sleeve, sleeve will be press fitted
onto axle, sleeve set pinned to axle for easy
removal

part number: 36 & 37

Axle Sleeve (to motor)


Aluminum


threaded interior (3/8")


key hole depending on motor axle
configuration


self
-
machined and threaded

part number: 40

Spring Pulley system


Same as current design


Self machine housing


Buy bearing from McMaster Carr for $7.45
each

part number: 8 & 9

Springs


Our design requires at least 7.5 lbs/in and 37
lbs of pull


Going with a 10.88 lbs/in /w max load of 44.6
lbs (1, 6 pack)


Order From McMaster
-
Carr for $12.70

part number: 12

Spring

Calculations


Entering all information into imulation of
current design, we need 4 N
-
m/rad spring
with 1*pi rotation initial condition


Converting that to our near linear system, we
need 2 springs at ~7.5 lbs/inch and 37 lbs of
pull


Focused on the 37 lbs of pull


Wanted a factor of safety of 1.20


Found a spring on McMaster Carr for
relatively cheap that had a safety factor of
1.205

part number: 12

Bike Wheel




Team member has many unused bikes at
their house


Will obtain this weekend


Aiming for a weight of 1.25 kg with most of
the weight around the outside (batteries)

part number: 14

String


Purchase heavy duty fishing line or kevlar
string from McMaster Carr or Home Depot


Low Cost/Low lead time component (not
concerned with this yet)

part number: 13

High Friction Feet


Require something on the ends of the frame
to take away the chance for slippage


PC non
-
slip pads are cheap and redily
available


Order 3 packs of 4 each

part number: 33

Motor

Motor Requirement: Must be able to drive a
torque of 1 N*m for .01 seconds (will slightly
oversize motor to be conservative)

DC motor (ease of wiring, inherent motion)

Brushed or Brushless?

part number: 26

Wires

Available in many gauges in the EE senior
design lab

part number: 18

Batteries

-
AA NiMH

-
NiMH is safer and rechargeable than LiIon

-
Eneloop 16 pack from Amazon only $38

-
retains charge capacity very well over
repeated recharging


Split into three banks: Motor voltage, 3.6V, and
4.8V for electronics

part number: 19

Gyroscope

-
Adjustable angular velocity setting for better
resolution (all give 0.1 deg/sec resolution or
better)

-
Breakout board includes all required
components

-
Quantity 2

part number: 20

Current Sensor

Pololu ACS714

-
Operates from
-
30A to +30A

-
Accuracy of +
-
1.5%

-
Hall effect sensor (electrically isolated from
current)

-
Quantity 3 (one for each battery system)

part number: 21

Voltage Sensor

Can use small surface mount resistor (minimal
power loss) measured across each bank of
batteries.


OR


Can use chemistry
-
specific charger that can
measure, report, and control the charge itself
and charge information such as current and
voltage.

part number: 22

Encoder

E5 optical kit encoder

-
Optical encoder

-
Hole
-
through design

-
.3 degree accuracy

-
Operates at speeds of 17,000 RPM

part number: 23

Processing, Control, Storage


TI LaunchPad meets customer need that the
coded part of the system be re
-
configurable
by a novice user in the future by having on
-
board JTAG emulation that can be accessed
via USB


Off
-
board storage is needed but has not
been selected



Processing, Control, Storage

TI LaunchPad



Microcontroller
Development Kit



C2000 Piccolo
TMS320F28027

part number: 29

Processing, Control, Storage

Motor Controller has not been chosen


It is likely that we will use:



Pololu Jrk 21x3
Controller

part number: 27

Processing, Control, Storage

System Code Testing Benchmarks

>inserting test code at different points to ensure
each piece of the system functions properly

1.
Get a random sensor & see if

a.
it can be polled consistently

b.
it generates an interrupt when it is supposed to

2.
Sample something simple such as a low
frequency sine wave

a.
will be easily able to see how well the signal is
sampled and reconstructed

Energy Flow Graph

Cost of Transport Analysis

1. No electronics power is included;

2. Friction is not accounted for;

3. Realitic COT will probably be much higher

Bill of Materials

https://docs.google.com/a/g.rit.edu/spreadsheet
/ccc?key=0ApxjvWO1pU8KdHNlaTlsS013W
G1aUS1OcTBKdWF5SVE#gid=0