problem statement - rochester institute of …edge.rit.edu/edge/p16602/public/final...

P16602: Glass Cutting Machine Wire Supply, Movement and Take-up

Hand-Off Document

Table of ContentsProblem Statement.....................................................................................................................................4

Process Flowchart........................................................................................................................................5

Bill of Materials...........................................................................................................................................6

Use Cases...................................................................................................................................................10

Functional Decomposition.........................................................................................................................11

Morph Charts, Selection Criteria & Pugh Analysis.....................................................................................12

Morph Charts........................................................................................................................................12

Selection Criteria...................................................................................................................................13

Decision Matrices......................................................................................................................................15

Electrical Architecture...............................................................................................................................19

Sensors..................................................................................................................................................19

Controls.................................................................................................................................................19

System Behavior........................................................................................................................................20

Parameterized Wire Velocity Profile......................................................................................................20

Motor Control & Dancer Pulley.............................................................................................................25

Traversing Pulley System Behavior........................................................................................................28

Design Decisions & Analysis.......................................................................................................................38

Dancer Pulley System............................................................................................................................38

Electrical System....................................................................................................................................44

Guide Pulley System..............................................................................................................................50

Guide Pulley Encoder.............................................................................................................................55

Supply Spool..........................................................................................................................................57

Take-Up Spool.......................................................................................................................................74

Annotated Drawings..................................................................................................................................78

Take-up Spool........................................................................................................................................78

Take-up Spool Locating Cone.................................................................................................................80

Take-up Spool Clamp Plate....................................................................................................................81

Supply Spool Locating Cone...................................................................................................................82

Supply Spool Clamp Plate......................................................................................................................83

2

Frame....................................................................................................................................................84

Index..........................................................................................................................................................91

Figures...................................................................................................................................................91

Tables....................................................................................................................................................91

3

Problem StatementGlass cutting is a process that can be done on both large and small scales with many different cutting tools. Some examples include using a saw blade, a diamond coated wire, or a brass coated wire in the presence of an abrasive slurry. Each of these methods will differently impact the way the glass is cut. The saw blade will cause a significant amount of movement and vibration of the glass. The diamond coated wire does not need any other tools to assist in cutting, but the wire can create chips in the surface of the glass during cutting. The brass coated wire with the slurry will give a finer cut than the diamond; however, temperature needs to be monitored.

Glass Fab Inc. currently uses a brass coated wire utilizing abrasive slurry for both large and small volume glass blanks. It is a Meyer Burger DS 264 glass cutting system that uses a wire wrapped around two guide rollers. These guides have grooves that the wire will sit in and can be adjusted to create various thickness cuts in the glass to get a range of different blanks. There is a control panel on the front as well to monitor the sensors and adjust the feed rate of the wire.

This machine works very well for Glass Fab, especially for high volume parts. However, the current Meyer Burger machine draws more power than every other machine Glass Fab owns put together. It takes a while to start up and because of the amount of power used, does not make sense to use for low volume parts. Glass Fab would like to look into creating a smaller machine, similar to the one that they currently have, that can be utilized only for low volume parts. There are existing smaller machines, but can only cut one blank at a time. The goal is to take elements of the bigger machine, essentially shrink it, and make a more energy efficient wire glass cutter.

This MSD project is part of a tightly coupled group of three projects that will run in the fall of 2015. The three projects are 16601, 16602 and 16603. Each of these project teams will develop, design and build a test fixture for their functional part of the overall system. In addition, members of this project team will actively participate on a “system engineering team”. The System engineering will define, develop and conceptually design a complete wire saw machine. The system engineering team will utilize CAD to convey the design. The system engineering team will work with the three project teams to array and manage all relevant project metrics at the system level. Some examples of these system metrics are weight, volume, power consumption, noise, heat emission. In addition, the system engineering team will be responsible for develop the necessary documentation to describe machine behavior and operation. Some examples of necessary documentation include a timing diagram for each of the possible use scenarios; this might include, cycle up, cycle down, and some diagnostic routines.

A key goal of P16602 is to design and build the spooling process. This includes specifying the spools and motors to be used, designing this portion of the wire path, mounting these components, and supplying power to these components. Of course, safety is paramount and should be considered in all aspects of the design and build of this industrial machine. In addition, this project team will be asked to consider and evaluate various methods of controlling and maintaining wire tension in the range of 10-40 newtons.

4

Process FlowchartThe following flowchart shows the process we followed to gain purchasing approval for designs.

Figure 1. Process Flowchart for Approval

5

Bill of MaterialsComponent Subsys. Quantity Status Approval? Supplier Cost Make/Buy

Cabinet ElectricalSystem

1 Place Holder NO Buy

Traversing actuator

Traversing Pulley System Behavior

2 Place Holder NO

Take-upSpool

LocatingCone

Take-upSpool

LocatingCone

- Place Holder NO Make

Flexible coupling

Spool 2 Place Holder NO Buy

Term. End Cap

ElectricalSystem

1 Ready to Order

? Rockwell $37.10 Buy

Factory Talk ME License

ElectricalSystem

1 Ready to Order


RSLogix License

ElectricalSystem

1 Ready to Order


Power rail filler

Motor Driver

4 Specification ? Rockwell $147.00 Buy

PanelView Plus

ElectricalSystem

1 Specification ? Buy

External Encoder

Guide Pulley

2 Specification NO Buy

IR Sensor MotorControl &

DancerPulley:


Emergency stop button

Power 1? Place Holder NO Buy

Load cell ElectricalSystem


Safety Door Interlock

ElectricalSystem


Safety Controller

ElectricalSystem

? Place Holder NO Buy

Bracket #1 - pneumatic

cylinder to t-slotted framing

Dancer Pulley

Mounting

4 Design YES ME Shop Make

Bracket #2 - Dancer 2 Design YES ME Shop Make

6

piston to linear slide

Pulley Mounting

Bracket #3 - linear slide to dancer

pulley

Dancer Pulley

Mounting

2 Design YES ME Shop Make

Acrylic case Guarding ? Place Holder NO MakePulley All Pulleys 8 Vendor

Selection? ME Shop Make

Roller BRG GuidePulley

Encoder


SupplySpool

LocatingCone

GuidePulley

Encoder

1 Specification NO Make

SupplySpool

LocatingCone

GuidePulley

Encoder


SupplySpool Clamp

Plate

GuidePulley

Encoder


Roller BRG Take-UpSpool


Take-upSpool

LocatingCone

Take-UpSpool


Take-upSpool

LocatingCone

Take-UpSpool


Take-upSpool Clamp

Plate

Take-UpSpool


Bolts General - Place Holder NOTapered

Roller BRGGuidePulley

Encoder


Tapered Roller BRG

Take-UpSpool


Pulley Bearings

All Pulleys 8 Specification NO Make

5m Shielded Ethernet

Cable

ElectricalSystem

1 Ready to Order

YES Rockwell $55.52 Buy

.3m Electrical 2 Ready to YES Rockwell $35.40 Buy

7

Shielded Ethernet

Cable

System Order

Motor SupplySpool:

1 Ready to Order

YES Rockwell $3,950.00 Buy

Motor GuidePulley

System:

1 Ready to Order


PLC ElectricalSystem

1 Ready to Order


24V Power Supply

ElectricalSystem

1 Ready to Order


Motor Driver Main

ElectricalSystem

1 Ready to Order


EtherNet/IP Safe

Torque-Off Control Module

ElectricalSystem

2 Ready to Order


Power Rail, Slim, 230V OR 460V, 8

Axis

ElectricalSystem

1 Ready to Order


Motor Driver

ElectricalSystem

1 Ready to Order


Connector, Motor

Feedback, D-Shell/Term Block, 15

Pin.

ElectricalSystem

2 Ready to Order


Connector Kit, 44 pin,

K6200/K6500 I/O,

Auxiliary feedback, and safety

ElectricalSystem

2 Ready to Order


SpeedTEC Cable, Motor Power,

Flying Lead

ElectricalSystem

2 Ready to Order


SpeedTEC Cable, Motor

Feedback, Flying Lead

ElectricalSystem

2 Ready to Order


8

Nylon insert All Pulleys 8 Ready to Order

YES GF Buy

Pneumatic Cylinder

DancerPulley

System

2 1 Received – 1 Ready to

Order

YES McMaster Carr $43.42 Buy

Pressure regulator/air

filter

DancerPulley

System

2 Vendor Selection/

Wellin

YES MSC Direct $39.50 Buy

Pneumatic PVC hose

DancerPulley

System

1 Vendor Selection/

Wellin

YES MSC Direct $15.00 Buy

Linear Motion Slide (air cylinder powered)

DancerPulley

System


Order

YES McMaster Carr Buy

T-slotted Frame

DancerPulley

System Mounting


Order


Pneumatic Cylinder Mounting

Foot Brackets

DancerPulley

System Mounting


Order


Bearing Pulley Systems

6 Ready to Order

NO McMaster Carr $40.92 Buy

Shaft collar GuidePulley

System

2 Ready to Order


Collar cap screw

GuidePulley

System

4 (100) Ready to Order


0.5 in. bar stock

GuidePulley

System

1 (36”) Ready to Order

NO Home Depot $5.57 Buy

Guide pulley bracket

[Gauge 16 Steel]

GuidePulley

System

2 Ready to Order

NO Home Depot $7.15 Make

Take-up spool

Take-UpSpool

1 Ready to Order

YES Wisco Alloy $644.00 Make

Supply spool

GuidePulley

Encoder

~4 Ready to Order

YES Slicing Tech Buy

9

Use CasesThe following are the use cases we created for our first design review. We haven’t created any formal documentation to further develop these; for example, scenario 2 can vary depending on what sort of fault is encountered (i.e. wire break vs. safety interlock open).

10

Functional Decomposition

Figure 2. Functional Decomposition

11

Morph Charts, Selection Criteria & Pugh AnalysisMorph ChartsFigure 3 shows our initial concepts for fulfilling key functions from our

Functional Decomposition

Figure 3. Initial Morph Chart

The initial concept designs, chosen from the morph chart, are embedded below. They were decided as follows:

12

Best Case: The best concept listed, without regards to cost or feasibility of implementation Cost-Effective & Feasible: Relatively cheap options, that seemed manageable to build by the

end of Senior Design Cost-Effective: The cheapest options, regardless of performance Feasible: The easiest to create a system from by the end of Senior Design User-Friendly: The easiest to interface to, program, and retrieve useful information from to

display to the user.

Selection Criteria

Ability to maintain and control tension: o How well is tension controlled? o What is the response time? o Can a proper tension be applied? o Can different tensions be applied? o How easy is it to adjust tension? o Is tension control automated or manual?

Ability to maintain and control speed: o How well is speed controlled? o What is the response time? o Can speed be altered/adjusted? o Is speed control automated or manual?

Lower cost to run: o How much does it cost the operator/company to run the fixture? o How much power is drawn during use? o What is the cost/availability of replacement parts? o Will the components wear? o What is the operator’s time to set up? o How much downtime will the system require?

Lower manufacturing cost: o What is the price of the components? o What is the cost of the manufacturing process? o How long will take to build/assemble the system originally?

Safe for wire break: o What is the response time to a wire break (shut down)? o How safe is the operator if a wire break occurs? o How safe are the machine components if a wire break occurs?

Safe for high heat:

13

o What is the response time to a system overheat (shut down)? o How safe is the operator if overheating occurs? o How safe are the machine components if overheating occurs?

Ease of integrating with other subsystems: o Can components “talk” with controller? o Do components support each other mechanically? o Do all components fit in reasonable 3D space?

Efficient power consumption: o How efficient is each component? o How efficient is the system as a whole? o Do components add inertia to overcome?

Reliability: o How badly do the components wear? o How often is component replacement needed? o How accurate are the sensor readings (temp, tension, etc.)?

Multi-directional movement: o Does every component have 100% functionality in both directions?

Feasible for one year: o Can the 7 students of P16602 successfully design, test and build this concept in

one year? Ease of use:

o How much time does it take to setup? o How many operators does it take to set up? o Are the controls simple and straightforward for the operator to use?

14

Decision Matrices

15

Electrical ArchitectureThe electrical architecture for our system is shown below in Figure 4.

Figure 4. Controls Architecture of the System

Sensors Two servo motors used for the spool motors (click Electrical System to jump to specification).

o MPM-B2152F-8J72AA for supply spoolo MPM-B2153F-8J72AA for take-up spool

Two actuators for traversing pulley (unspecified – represented with servo in Figure 4). See Traversing Pulley System Behavior.

Position sensors used to detect position (and ideally velocity) of the dancer pulley. Load cells are used to read tension of the wire (attached to guide pulley). Encoders are attached to guide pulley to get feedback on the wire velocity (motor feedback by

itself is not sufficient).

Controls PLC = Rockwell L36-ERM controller (click Electrical System to jump to specification). The computer is used to program the PLC. The Human-Machine Interface (HMI) is used for machine configuration, and to display

information to the user.

For more information, see Electrical System.

19

System BehaviorParameterized Wire Velocity ProfileP16602: Glass Cutting Machine Wire Supply, Movement and Take-up J.D. IncavoApril 26, 2016

Contents: Moving Forward: Key Assumptions: Analysis: Validation: Summary: Moving Forward:

Overview:This section describes the scaling of the wire velocity profile, with respect to the maximum desired wire velocity, stroke length, and advance rate.

Key Assumptions: The motors will accelerate and decelerate with a magnitude of 2 m/s2. This analysis assumes infinite jerk, allowing for an instantaneous change in the acceleration from

±2 m/s2 to 0 m/s2. In the real system, this will not be the case; but this approximation should work well enough.

Analysis:Stroke Length ( S ) : The length of wire moved forward in one cut cycle.

Advance Rate ( A ) : The amount of wire kept on the take-up spool at the end of one cut cycle.

The user will be able to determine the wire velocity they wish to accelerate to, V. They will also be able to determine the advance rate, A, and the stroke length, S. With a fixed acceleration of 2 m/s2, we will have a trapezoidal velocity profile, shown below. This fixed acceleration also allows us to determine the timing of the velocity profile, with 6 primary segments labeled in Figure 5.

20

Figure 5. Example Wire Velocity Profile (Used in current Glass Fab configuration)

Starting from rest, with a 2 m/s2 acceleration, it would take V/2 seconds to reach the maximum speed. This is the length of time in segments 1, 3, 4 and 6. To determine how much time is required for segment 2, the stroke length has to be considered. Taking the integrals of segments 1 and 3 (i.e. area of the triangle), we get the following:

w1=w3=12 ( V

2 )V=V 2

4

Adding these segments to find the amount of wire passed in acceleration and deceleration, we get:

waccel=2w1=V 2

2

Because we need to pass a length of S wire in the forward direction, we need to pass S – V2/2 in the steady-state condition (i.e. segment 2). The time required to move this length at velocity V can be easily calculated:

t 2=S−V 2

2V

= SV

−V2

Similar analysis can be done to find t5, where the length of wire that has to be moved in the opposite direction is S – A, rather than S.

t5=S−A−V 2

2V

=S−AV

−V2

21

However, these are just the difference in time between points, not the elapsed time since the cut cycle started. Allowing for an optional pause of length P between segments 3 and 4, and after segment 6, we get the following, starting at t = 0.

Figure 6. Wire Velocity Profile with time labels (P = 0)

t0=0

t 0−2=t 0−1+ t2=V2

+ SV

−V2

= SV

t 0−3=t 0−2+ t3=SV

+ V2

t 0−P 1=t 0−3+P= SV

+ V2

+P

t0−4=t0−P 1+ t4=SV

+V2

+P+ V2

= SV

+V +P

t0−5=t0−4+t 5=SV

+V +P+ S−AV

− V2

=2 S−AV

+ V2

+P

t 0−6=t 0−5+t6=2S− A

V+ V

2+P+ V

2=2 S−A

V+V+P

22

t end=t 0−6+P= 2 S−AV

+V +P+P= 2S−AV

+V +2 P

Or, in a more useful form, see Table 1 below. This table was used in Excel to create Figure 5, and changing the parameters will result in a different scaling of the wire velocity profile.

Table 1. Wire Velocity Profile

Validation:Embedded below is an Excel spreadsheet that will allow the user to vary the input parameters to see how the changes will be reflected in the wire velocity profile.

This Excel spreadsheet was used to verify the simplified parameterized cut cycle with the calculations that we were using for previous timing diagrams. The results came out the same after simplifying the timing, which confirmed that the simplifications were correct.

Summary:The key takeaway is that not only will the velocity have to change to match the user input, but the timing of the system is also dependent upon this. A higher wire velocity will result in a shorter cut cycle time, and a lower wire velocity results in a longer cut cycle time.

23

Time [s] Velocity [m/s]0 0

V2

V

SV

V

SV

+V2

0

SV

+V2+P 0

SV

+V +P -V

(2 S−A )V

+V2+P -V

(2 S−A )V

+V +P 0

(2S−A )V

+V +2 P 0

The following table shows the timing required to implement the trapezoidal velocity profile in terms of the user inputs (wire velocity V, stroke length S, advance rate A, and pause P)

Table 2. Parameterized Cut Cycle Time

Moving Forward:While the timing has been developed mathematically, it has not been tested in hardware, or programmed in software. Additionally, the actual implementation will have to account for jerk, which will result in smoothing out the velocity profile. It is up to the implementation to choose whether or not to keep the same timing and pass less wire, or to pass the desired amount of wire and increase the timing.

24

Time [s] Velocity [m/s]0 0

V2

V

SV

V

SV

+V2

0

SV

+V2+P 0

SV

+V +P -V

(2 S−A )V

+V2+P -V

(2 S−A )V

+V +P 0

(2S−A )V

+V +2 P 0

Motor Control & Dancer PulleyP16602: Glass Cutting Machine Wire Supply, Movement and Take-up J.D. IncavoApril 21, 2016

Contents: Overview: Key Assumptions: Analysis: Summary:

Motor Reference Pulley Reference Dancer Reference

Moving Forward:

Overview:The dancer pulley system will be used to provide necessary feedback to the controller for controlling motor velocity. The combination of the dancer velocity and the line speed at the guide pulley, or guide rollers, and the spool motor angular velocity should be sufficient to find out the instantaneous radius of the spool, and therefore how much to adjust the spool velocity to reach the desired wire velocity.

Key Assumptions: No slip on guide pulley Dancer system can be de-coupled from traversing pulley Wire velocity at the guide pulley is approximately equal to the wire velocity at the guide

rollers Dancer maintains fixed tension across entire extension range Guide pulley has fixed radius (i.e. no wear) Can set the polling rate for the position sensor for the dancer system

Analysis:For this analysis, it is assumed that the dancer extensions movement will solely be a function of the velocity mismatch between the spool and guide rollers. This assumes that the pneumatics will be able to adjust the pressure to supply constant wire tension. As such, the path length difference can be handled like a flow problem: anything coming in either has to go out, or go into the running time storage (i.e. excess path length), allowing the dancer to extend. Conversely, if the output is moving faster than the input, the dancer would give up the excess path length.

From this assumption, we find that the change in the running time storage (dancer path length) is given by:

25

ΔL=(vspool−v pulley ) Δt

Due to the 180 degree wire wrap on the dancer pulley, this is split evenly into the dancer extension, x (Note: By convention, increasing x is defined as extension).

ΔL=2 x

If we take the time derivative, we find:

dLdt

=2 dxdt

=vspool−v pulley=ωmotor R spool−ω pulley Rpulley

If we can determine the time interval that the dancer position sensor is polled, we can

approximate

dxdt . With this information and the guide pulley encoder feedback (or, in the

integrated system, guide roller feedback) we can determine the wire velocity coming off the spool, to compare to a reference.

vspool=ωmotor R spool=2 dxdt

+ω pulley R pulley

Assuming we have angular velocity feedback from the motor, we can determine the approximate instantaneous spool radius.

R spool=2 dx

dt+ω pulley Rpulley

ωmotor

Summary:Using the dancer extension, we will be able to determine the tangential wire velocity coming off the spool – motor feedback alone would not be sufficient to determine this, due to the dependency upon radius.

vspool=2 dxdt

+ωpulley R pulley

If we can determine the rate of motion of the dancer, we would be able to calculate the radius. From this information, we can determine the magnitude of change in the angular velocity to match the guide pulley. However, it is not simply enough to match the speed of the guide pulley after a mismatch – we must overcompensate so the dancer will return to its nominal position.

Moving Forward: Further define the control scheme

o Currently have 3 control proposals

26

1. Motor ReferenceAdjust the angular velocity profile magnitude to the instantaneous radius

+ Better-defined relationship

- Processing overhead from “constantly” calculating the spool radius, and then calculating the magnitude of the desired angular velocity

This would involve scaling the virtual axis so the magnitude of the angular velocity profile is equal to:

ωmotor=V desired

Rspool

2. Pulley ReferenceCalculate the angular velocity profile once per run, normalized to pulley radius. Compensate for velocity mismatch by taking radius into account.

+ Less processing overhead from only calculating reference once (i.e. static reference vs. dynamic)

- Need to convert with radius to desired velocity change- Still need to “constantly” calculate the spool radius

Δωmotor=ω pulley Rpulley

R spool−ωmotor

This would stop dancer from moving, not bring it to equilibrium position. To bring it back, you must overcompensate for velocity mismatch. This would involve some oscillation in the motor velocity to compensate and move the dancer to equilibrium. This raised the concern of if the motor would ever reach a “steady-state” behavior, where it maintains the desired velocity with the dancer in equilibrium position.

This should be possible, because the system reaction time should be fast enough to minimize any velocity mismatch between the guide rollers and the spools. Ideally, the fast reaction time of the system (the PLC should react in the order of microseconds, if not less, and the sensors are analog) should prevent any extreme compensation for mismatch, so the spool motor would closely follow the guide rollers.

3. Dancer ReferenceAdjust the motor speed based solely on the dancer displacement+ Easy to implement

(x>0 => slow down; x<0 => speed up)+ No need for guide pulley encoder (?)

27

- Constant speed adjustmento Oscillate around x=0, constantly compensate for not being at

equilibrium- Hard to define magnitude of velocity adjustment

o Determine experimentally through trial & error?o Function of radius? => Need encoder, or feedback from guide roller

Alternative: in-line pulley + encoder between spool and dancer would provide more direct feedback for wire velocity of spool

o Ideally match the wire velocity of spool – no need to worry about dancer’s effecto Only need to account for mismatch in radius, angular velocity

vspool=ωmotor Rspool=ωpulley R pulley=v pulley

R spool=ω pulley Rpulley

ωmotor

Traversing Pulley System BehaviorP16602: Glass Cutting Machine Wire Supply, Movement and Take-up J.D. IncavoApril 21, 2016

Contents: Moving Forward: Key Assumptions: Analysis:

Ideal Steady-State Analysis Non-Ideal Effects

Finite Drum Length; Larger Flange Radius Acceleration/Deceleration Profile Wire Velocity dependent upon Wire-Laying Angle Radius dependent upon Wire Displacement per Revolution Worst Case

Summary: Moving Forward:

Overview:The traversing pulley will move back and forth across the length of the spool drum, controlling the wrap of the wire around the spool. The displacement or wire-laying angle of the wrap can be controlled by varying the speed of the traversing pulley. Controlling one of these means that the other parameter will be variable with radius; for example, an 85° wrap angle will result in a smaller displacement per revolution at a small radius than it will at a large radius. On the other hand, fixing the displacement will result in a smaller angle as the radius increases. There are several factors that determine the traversing

28

pulley movement, including the wire line velocity, its position within its limits, and the wire acceleration profile.

In addition to this, there are several non-idealities, including a “dead zone” caused by the flange extending past the radius, and the pulley having to change directions upon reaching the end of the spool. Another deviation from the formula is the fact that the traversing pulley needs to accelerate/decelerate with the wire acceleration profile; however, it may change direction during this (shown later in Figure 3) if the pulley is near the end of the spool.

Key Assumptions: The length of wire around one wrap of the spool is approximately equal to the circumference.

This would require the wire-laying angle, θ, being close to 90° (Figure 1 has a greatly exaggerated θ for illustration purposes)

o While equations are generalized for θ, they require it close to 90°. This assumption is also used when analyzing the limits of the traversing system.

The wire would not suddenly change angle at the point of contact to the spool (where the wire path comes in tangent). This would require the traversing pulley to be leading by an angle rather than being right above the point of contact at a 90° angle, although the steady-state analysis isn’t dependent upon this as it focuses on the velocity of the point of contact along the spool drum.

o However, the assumption of an angle does lead to non-ideal conditions, discussed below.

The analysis focuses primarily on steady-state when the wire is moving at the constant velocity, rather than during acceleration.

The radius of the spool can be determined using dancer feedback.

Analysis:

Ideal Steady-State AnalysisThe analysis of the system at steady-state focuses on the movement of the point of contact along the spool. In Figure 7, shown below, the wire starts at the point of contact shown and moves a distance x in one revolution of the spool, taking time t.

29

Figure 7. Illustration of spool and traversing pulley

Assuming the point of contact would move at the same velocity as the traversing pulley velocity, VT, we get the following:

x=V T t

V T=xt

Table 3. Traversing Pulley System Variables

VARIABLE DEFINITION UNITSVT Traversing pulley velocity m/sVW Tangential wire velocity m/sR Spool radius mD Spool diameter (D = 2R) mx Distance traveled per wrap mt Time per revolution sθ Wire-laying angle Degrees (or radians)

30

We would then need to find the time for one revolution. Assuming that the wrap angle is close to 90°, and therefore the wire wrap along one revolution is approximately the circumference, we can approximate the following:

2πR [m/s] = 1 [rev/s]

1 [m/s] =

12πR [rev/s]

V W [m/s] =

V W

2 πR [rev/s]

We would invert the revolutions per second to find the seconds per revolution, t.

t=2πRV W [s]

The distance traveled per revolution, x, can be found by looking at the geometry of the system. Assuming a constant wire wrapping angle, we would have created 4 right triangles along x with perfect symmetry (one is shown explicitly in Figure 7). These triangles have sides of length D = 2R and x/2, with an angle θ with respect to the surface of the spool. Using trigonometry, we determine the following:

tanθ=2 Dx

=4 Rx

θ=tan−1 ( 2Dx )

x= 2 Dtanθ

= 4 Rtan θ

With this information, the velocity of the point of contact can be determined.

V T=xt=

4 Rtan θ2 πRV W

= 4 Rtan θ

∗V W

2πR=

2 V W

π tan θ

The most useful forms of this are to wrap according to a desired x or a desired θ, resulting in the equations below:

V T ( x )=xV W

2 πR

31

V T (θ )=2 V W

π tan θ

It should be noted, however, that these will behave differently. As radius changes, the wire-laying angle required to maintain a fixed displacement will change (wire-laying angle will increase as radius does, approaching 90°). Fixing the wire-laying angle will cause different displacements between layers. For the most uniform wrap, fixing the displacement would be ideal.

From this, we can determine the traversing velocities in the extreme cases of the pulley (bounding it to reasonable values). For determining maximum velocity, we assume a minimum radius (requires highest angular velocity), a wire velocity VW = 18 m/s, and a wire-laying angle θ = 85°. For determining minimum velocity, we assume a maximum radius (lowest angular velocity), a wire velocity VW = 10 m/s, and a displacement of 175µm per revolution (one wire diameter).

For the take-up spool, the radius ranges between 0.06 ≤ R ≤ 0.12 meters. From this, we get the following:

Maximum traversing velocity = 1.0025 m/s

Minimum traversing velocity = 0.00253 m/s

For the supply spool, the radius ranges between 0.0585 ≤ R ≤ 0.1275 meters (the larger radius is the flange diameter, not necessarily the largest wrap diameter). From this, we get the following:

Maximum traversing velocity = 1.0025 m/s

Minimum traversing velocity = 0.00218 m/s

It is important to note that the minimum steady-state traversing velocities are approximately 2 mm/s, while the maximum is 1.003 m/s. The traversing pulley actuator must have high precision velocity control to maintain a close wrap at low wire velocity and maximum radius.

Non-Ideal Effects

Finite Drum Length; Larger Flange RadiusThe first non-ideality to consider is the fact that the spool has a finite drum length, which means that when approaching the end of the spool we will need to reverse direction. In addition to this, the spool flanges cause the angle to have limits for certain radii in order to approach the end of the spool (See Figure 8). In other words, there is a region of the spool where we would not be able to lay wire if we were always leading by an angle; however, as we change direction the angle between the point of contact on the spool and the traversing pulley will go perpendicular as the traversing pulley passes the point of contact. If this were to happen near the end of the spool, we would be able to lay wire down in the otherwise un-used region (“dead zone”).

32

Figure 8. Spool limits

The result of this is that we should turn the traversing pulley around once it reaches the limit (or start to turn around while approaching it). The relationship between the cut-off region and the wire-laying angle θ is derived below, using the right triangle, where DF is the flange diameter of the spool.

tanθ=(DF

2−R)xcut−off

xcut−off=

DF

2−R

tan θ

Figure 8 shows an extreme case of a small wire-laying angle; realistically, the wire-laying angle would be approximately 90° for a close wrap, so the “dead zone” would be negligible. However, it may be large in the case of a small radius and angle deviating from 90°. For example, with a supply spool at minimum radius and a wire-laying angle of 85°, xcut-off = 5.9 mm, which is roughly the distance of 34 wires laying right next to each other.

The hard limit of the traversing pulley itself can be found similarly to the cut-off if the distance from the spool is known. It should also be noted that the actual take-up spool will have a taper, which further limits the wire-laying angle near the end of the spool. In the following, Rp is the radius of the pulley, accounting for the point of tangency with the wire being somewhere along the side of the spool. The underlying assumption is that the wire will come off the spool almost vertically, so the point of tangency will be very close to being on the left end of the spool (making a 90® angle with the wire coming off the pulley into the rest of the system).

33

tanθ=dT−

DF

2x lim−R p

x lim=dT−

DF

2tanθ

+R p

It should also be noted that, while this was derived with the right-most position in Figure 8 in mind, it would need very little modifying to work for the left-most position. If we were to call the left end of the spool drum x = 0, and assume the spool has length l, then we would have the following range for the pulley:

−dT−

DF

2tanθ

+R p≤x≤l+dT−

DF

2tan θ

+R p

When θ ≈ 90°, this can be approximated as:

Rp≤x≤l+R p

Acceleration/Deceleration ProfileThe next non-ideality is that the spool will accelerate and decelerate – the variable VW would be changing during this period. We can find this acceleration required with the following assumptions:

The wire velocity will accelerate at 2 m/s2. There is a near instantaneous change in the acceleration (infinite jerk). As such, the time it takes for the wire velocity to hit steady-state is equal to Vw/2.

V T=aT t

xV W

2 πR=aT

V W

2

aT=x

πR= 4

π tan θ

Wire Velocity dependent upon Wire-Laying AngleAnother potential deviation from the ideal case is that the wire velocity will change depending on the wire-laying angle. As the angle deviates from 90°, more wire will be able to fit in one revolution, and at some point it would become a significant difference from the circumference at the given diameter. Ideally, this will be taken care of by feedback from the dancer system, and the wire-laying angle should

34

be kept close to 90° - ideally between 85° and 89.98° (perfect wrap at full supply spool). However, this does bring up the next non-ideality.

Radius dependent upon Wire Displacement per RevolutionWhen the wire-laying angle deviates from 90°, less wire will be contained in one layer as the displacement per wrap increases (as stated above). This will lead to the radius increasing at a rate faster than may be anticipated on the assumption that each wrap around the spool will be at the circumference; as discussed before, the wrap length will be greater than the circumference. As such, it would be difficult to predict the radius based solely on the length of wire traveled, or number of cut cycles. To prevent this from being an issue, the velocity can be controlled using a fixed displacement per revolution, x.

Worst CaseWhat should be taken into account is that these non-idealities are not mutually exclusive – the traversing pulley may reach the end of the spool during acceleration. The following describes the scenario where there is not enough length for the traversing pulley to reach steady-state before needing to turn around.

x=12

aT t 2⇒ 12

4π tanθ

V W2

4⇒

V W2

2 π tanθ≥xremaining

The case where there is not enough length along the spool in one cut cycle is of interest, because this is when the transition between wrap layers occurs. It is also important to note that it will transition between the layers for consecutive cut cycles, until the range of spooling and unspooling shifts and is contained within one wrap layer.

0 5 10 15 20 25 30 35 40 450

2

4

6

8

10

12

14

-0.003

-0.0025

-0.002

-0.0015

-0.001

-0.0005

0

0.0005

0.001

0.0015

0.002

Wire VelocityTraversing Velocity

Figure 9. Example Acceleration Profile with Traversing Pulley near end of spool

35

Summary:The traversing speed can be parameterized depending on whether a fixed displacement per revolution or specific wrap angle is desired.

V T ( x )=xV W

2 πR

V T (θ )=2 V W

π tan θ

For our spools, this would result in the following minimum and maximum steady-state (not accelerating/decelerating) velocities:

Min Traversing Velocity [m/s] Max Traversing Velocity [m/s]Take-up Spool 0.00253 0.00218Supply Spool 1.0025 1.0025

The following will cause non-ideal deviations from the velocities described above:

Finite length of the spool drumo Traversing pulley will have to reverse direction

Flange diameter larger than wrap diametero “Dead zones”, hard limit of traversing pulley

Wire acceleration profileo Traversing pulley needs to accelerate to match profile

Wire velocity dependent upon wire-laying angle Radius dependent upon wire displacement per revolution

Because of the traversing pulley leading the wire at an angle, there are areas that won’t be able to be wrapped near the flanges without quickly reversing direction and allowing the point of contact on the spool to catch up. From the edges of the spool, the following are the “dead zones”:

xcut−off=

DF

2−R

tan θ

When θ ≈ 90° (like with a tight wire-wrapping), this will go to zero.

From the same type of analysis used to determine the “dead zones”, the hard limits of the traversing pulley were found with respect to the wire-laying angle.

36

−dT−

DF

2tanθ

+R p≤x≤l+dT−

DF

2tan θ

+R p

When θ ≈ 90°, this can be approximated as:

Rp≤x≤l+R p

The traversing pulley will have to accelerate at one of the following rates, to maintain either constant displacement per revolution or constant wire-laying angle:

aT ( x )= xπR

aT (θ )= 4π tanθ

Abruptly changing direction near the end of the spool would have a spike of infinite acceleration (or during acceleration, infinite jerk). This will happen if the traversing pulley hits the limit, which may occur during acceleration if:

V W2

2 π tan θ≥xremaining

Moving Forward:The biggest step moving forward is specifying a traversing actuator. We have several criteria that an actuator would need to meet:

Precise velocity and position control Can be automated Stroke length is greater than or equal to the largest range we would need for the pulley:

−dT−

DF

2tanθ

+R p≤x≤l+dT−

DF

2tan θ

+R p

Linear Actuator Comparison:

From this comparison, it seems that the best class of linear actuators may be a moving coil. However, I haven’t done thorough research into linear actuators and moving coils yet.

37

Design Decisions & AnalysisDancer Pulley SystemP16602: Glass Cutting Machine Wire Supply, Movement and Take-up Sarah BrownApril 10, 2016


Overview:The Dancer Pulley System is one component of the wire movement fixture whose major

function is to maintain tension and correct for changes in the tension caused by a change in a length of the wire path. There are symmetric dancer pulleys on either side of wire path, the dancer pulley is the second pulley to come into contact with the wire after the traversing pulley and before the guide pulley.

38

Dancer Pulley System

3

2

1

Figure 10. Fixture CAD with Dancer Pulley System labeled

The Dancer Pulley system needed to be reactive to changes in the tension of the wire, its motion is not directly controlled, but set to an initial position and will react to changes to bring the dancer pulley back to its ‘home’ position.

Key Assumptions: The wire path needs to be able to change by a length of 100 mm The wrap angle on the dancer pulley is 180o

The RIT Campus and Glass Fab have 80 psi connections available 80 psi will be sufficient for air cylinder to function properly The air cylinder will be mounted directly to the Dancer Pulley, there will be no pulley arm The dancer pulley system will ‘maintain’ tension in the fixture The air cylinder provides an opposing force to the tension in the wire wrapped around the

dancer pulley thereby maintaining the desired wire tension in the system The air cylinder is provided a constant flow of compressed air which opposes the changes in

wire path length during fixture operation Symmetric dancer pulley systems will be on either side of wire path (supply spool side & take-up

spool side)

Analysis:Research was conducted first on how the Meyer Burger DS264 machine controls its tension via a

pivoting dancer pulley arm. This research led to the field of web tensioning that is utilized most often in the textile industry when the tension of the material needs to be properly controlled. As the topic of web tensioning was further researched, a few options for how to control the tension with a dancer arm appeared. The DS264 utilized a pivoting dancer pulley arm, however, this tensioning could be accomplished via a linear motion and that was the direction the team chose to move forward. Linear motion was chosen over a pivoting, multi-dimensional motion arm due to the simplicity of such a design.

Once linear motion was chosen, the means to provide the motion was the next decision to be made. Ultimately a pneumatic air cylinder was chosen over any other means due to the ability of the pneumatic cylinder to be completely reactive to changes in tension. A linear actuator was considered, however, because the team wasn’t looking to control the precise movement of the dancer pulley, this option wasn’t desirable. Extensive research was done into pneumatic cylinders, as there are several different types, from double-acting cylinders (air is supplied to either end of the cylinder to extend and retract the piston) to single-acting cylinders (air is supplied to one end of the cylinder and some form of mechanism brings the piston back to ‘home’ position). There are also repairable (cylinder may be taken apart & parts replaced) to non-repairable (cylinder is sealed and cannot be taken apart for

39

Figure 11: Diagram of Dancer Pulley System Reactions

maintenance). And internally there are several mechanisms that make air cylinders better suited for particular applications (air bellows as opposed to pistons for near-zero friction applications; lubricating cylinders to provide lubrication to piston for reduced friction motion, etc.) For the team’s particular application a single-acting, spring return air cylinder was chosen for the fact that compressed air could be supplied to the cylinder to extend the piston to a particular location and then the spring can return the piston to that ‘home’ position as tension changes.

There are also several auxiliary pneumatic components that are required for a pneumatic system such as this to provide a controlled air flow to the cylinder, and to clean the air in order to protect the air cylinder from harsh conditions. Components like a water/oil separator, air filter, oil lubricator can ensure that the air that is being provided to the air cylinder is clean from containments to increase the life of the air cylinder as well as ensuring the air is free from condensation but has the correct amount of lubrication to keep the piston properly running. These components aren’t necessarily needed for initial testing or operation in a controlled environment such as RIT, but for end-state, in an industrial environment like Glass Fab, these components are needed to protect components against potentially harsh shop conditions. A pressure regulator (proportional regulator) is also necessary to step-down the compressed air that is supplied to the correct range of pressures for the system. A system of multiple regulators/valves may be used to ensure that the air cylinder will be provided a stable, unchanging pressure.

Research was also done into the way that compressed air can be supplied to the air cylinder, currently, the system is designed to hook directly to the compressed air source assuming that a constant, un-interrupted flow can be provided. However, for shop conditions where a loss of pressure suddenly may be a factor, an auxiliary tank may be necessary to ensure that the dancer system always has a source of air available.

After several weeks of research was completed, it was time to begin to source specific components, and the determination that it would be best to choose components that could provide the testing capabilities but were not end-state quality. Because the dancer pulley system is a very complex system that is not fully understood, it would be more beneficial to construct the system with lower quality components that would be able to provide the necessary testing data and then purchase industrial-grade components down-the-line. A non-repairable, single-acting, 8” stroke length pneumatic cylinder was chosen, with the understanding that down-the-line a more robust, industrial grade single-acting, repairable cylinder would be chosen. A few auxiliary components like a water/oil separator, pressure regulator were provided by Mr. Wellin at no cost to the team.

It was also decided to proceed with the purchase of components for 1 dancer pulley system to prove that the system and theory behind the design would work properly before investing in enough components for the entirety of the wire movement fixture.

Validation:Once the major pneumatic components were sourced and acquired, the next step was to design

and build the mounting system for the pneumatic cylinder and the dancer pulley that would be mounted

40

to the system. The dancer pulley mounting system again was decided to be designed in such a way that could be quickly manufactured to allow the system to be tested and not necessarily industrial grade. T-slotted framing (8020) was chosen to be the frame to which the air cylinder could be mounted which would also serve as the track for linear motion of the dancer pulley. The t-slotted frame allows to be easily integrated to the rest of the team’s fixture. 3 major brackets were identified to be designed and manufactured; #1 the bracket to mount the air cylinder to the t-slotted framing; #2 the bracket to connect the end of the piston to the t-slotted frame; and #3 the bracket to connect the piston to the dancer pulley.

A linear bearing was bought from McMaster-Carr which provides low-friction motion along the t-slotted framing. This linear bearing was chosen for its relatively low cost and the ability for multiple locations of attachment for the dancer pulley.

Bracket #1 was designed to be a simple connection from the ends of the air cylinder to the t-slotted framing that allows for adjustment along the length of the t-slotted framing, but can be rigidly attached for testing purposes. The brackets (4 total) were machined in the RIT machine shop using aluminum stock provided by the shop at no cost to the team.

Figure 12. Bracket #1 design

Bracket #2 was designed to be a very simple rigid connection between the air cylinder’s piston and the linear bearing slide. This bracket transfers motion from the air cylinder to the linear bearing which will be attached to the dancer pulley. The bracket (2 total) was machined in the RIT machine shop using aluminum stock provided by the shop at no cost to the team.

41

end of air cylinderBracket #1

t-slotted frame

Figure 13. Bracket #2 design

After the brackets were designed and manufactured, the mounting system was constructed which now allows for a limited amount of testing of the system because the air cylinder can be connected to a supply of compressed air and the functionality of linear slide can be tested. Full functionality testing is not available because the team hasn’t constructed any pulleys which would allow wire to actually be wrapped around and test the full tensioning abilities of the dancer pulley system.

Summary:The dancer pulley system is an integral part of the wire movement process for a glass cutting

machine, allowing for proper tension to be maintained throughout the cutting process. The dancer pulley system needs to be reactive to changes in the tension during operation and correct accordingly. The system needs to be tested and function in a way to ensure that proper tension will be maintained throughout the entirety of the cutting process.

Moving Forward:Moving forward in this process, future teams will need to:

Fabricate the dancer pulley & mount to the linear slide Construct the full pneumatic system (with auxiliary components) Construct wire tensioning test fixture

42

air cylinder Bracket #2

linear slidet-slotted frame

Figure 14. Constructed Dancer Pulley Mounting System

o A test fixture that will allow the wire to be wrapped around the pulley and held in a fashion that will allow the tensioning capabilities of the dancer pulley system to be tested

Test tensioning capabilities of the dancer pulley system Acquire additional components for another dancer pulley system Machine/Construct another complete dancer pulley system for opposite side of fixture Test the interfacing between the 2 dancer pulley systems Integrate into wire movement fixture Integrate into complete glass cutting machine

43

Electrical SystemP16602: Glass Cutting Machine Wire Supply, Movement and Take-up Joseph OddoApril 24, 2016

Content: Overview:

o Wire Movement: Analysis:

o Guide Pulley System:o Supply Spool:o Motor Control & Dancer Pulley:

Moving Forward:o Controls System:o Programming:o Cabinet:

Useful Documents:

Overview:Rockwell L36-ERM controller was chosen because of the needed integrated motion. The number of axis directly correlates to the number of motors. The idea was to use virtual axis to scale the guide roller to the take-up and supply spool. By doing so the error of this could be monitored and minimized. Deciding on Rockwell was not an easy process. There were many times when the price caused major hesitation with the customer. To help with the decision process, multiple options were presented. I laid these options out from cheapest to most expensive. The other concern coupled with this was there was concern of the safety of the students working with high voltage. The options were laid out so that all except for the “homerun” option were able to be made without anything higher than common household voltages 115V, 230V. While these are still dangerous essentially most MSD teams are exposed to this. See the electrical presentation from 12/1/1. After this presentation, Glass Fab committed to being convinced with the Rockwell option.

44

The cabinet was the next major development after the Rockwell option was selected. Both Zeller and Agile were contacted with only Zeller responding. Multiple meetings were set up with Zeller both with Ed, and GlassFab. See the Zeller presentation from 2/14/16. Zeller has been willing to separate the design and manufacturing process to make it easy to line up with the MSD process. They have also made their services available to be able to develop the software as well.

45

Figure 16. Cabinet Layout

Wire Movement:The responsibility of this fixture is to move wire at appropriate speeds and tension without breaking the wire. From this, several key aspects to this fixture were developed: take-up spool, supply spool, dancer system, and traversing pulley. The idea was to have the take-up and supply respond to their respective guide roller and be scaled via virtual axis. During acceleration it was assumed that the dancer system was going to be used to help smooth any error in startup and that the spool would almost exclusively follow this. The traversing spool is responsible for wrapping actual wire on the spool. This was first developed with just an air cylinder, but the customer and faculty guides requested for more precision.

46

Analysis:

Guide Pulley System:See Hannah Micca’s documents for all mechanical options that were developed for the Take-Up Spool. From these options motion analyzer was used. I calculated inertia and speed in revolutions per minute for each option. The actual inputs can be seen in the motion analyzer files. From this I analyzed the validity of each option and working with Hannah was able to come up with very realistic options where the torque speed curve was not unreasonable and the inertia ratio was in a comfortable range. This selection was revised several times with a Rockwell engineer and with Hannah Micca.

47

Supply Spool: For this there was a substantial lower amount of mechanical development as the actual spool was an off the shelf item. The same calculations were made as in the take-up spools case. This was also revised multiple times with the Rockwell engineer.

48

Motor Control & Dancer Pulley:I did not do very much work with this but the idea is to have an air regulated system responding to changes in tension in the wire. A position sensor is mounted to give feedback to the motor to make minor adjustments on the fly. It was later assumed that this would be playing a much larger role in the acceleration and deceleration profile. My preferred selection would have been the LVDT as these are widely used in manufacturing. They will be able to provide more than enough resolution for this application.

Moving Forward:

Controls System:If not already done so when starting MSD1 need to order controls system specified in BOM. I held the project up with even ordering the controller to make sure everything was iterated on multiple times to guarantee items selected were correct. This does not have a crazy lead time, but is expensive. Should have a discussion with Prof. Hanzlik and Wayne and see their stance.

Programming:See EE presentation from 2/28/16. This is where word form controls were first put down and what operation modes.

Cabinet:GlassFab was hesitant at first to commit money towards the cabinet. They had a retired Electrical Engineer in mind to accomplish the task. The problem with using Carl, the retired EE, was that he could provide no liability for the cabinet construction. The biggest thing with Zeller is they were willing to assume full liability with the cabinet.

Zeller has also offered their services to develop software and safety for the machine as a whole.

With Zeller doing the software it will only be finished up to at most 70%. I have presented Zeller with my anticipated system controls in order to have this available as an alternative option.

I am 100% on board for having Zeller as a resource for the system safety. Outside of door switches and e-stops they will be able to better identify what the machine actually needs.

Useful Documents:EE Presentation 12/1/15Zeller PresentationCarl PresentationEE Presentation 2/28/16

49

Guide Pulley SystemP16602: Glass Cutting Machine Wire Supply, Movement and Take-up Caleb WeeksApril 27, 2016


Overview:The purpose of the guide pulley system is to guide the wire from the dancer pulley system to the guide rollers. Two of these systems exist: on both sides of the machine. The system is comprised of four components: the rod, collar, bracket and cap screws. The pulley bracket can be manually adjusted to line up with the appropriate grooves on the guide rollers. The guide pulley system is located in the space associated with the guide roller system.

Figure 17. CAD model of Guide Pulley assembly

Key Assumptions: System remains unmoved most of the time Primary forces acting on the system are from tension in the wire System is supported on both ends by the frame surrounding the guide rollers

50

Analysis:Analysis was performed individually on each part as well as considering the entire assembly. The key consideration for validation is a force stress and displacement analysis. Manufacturability of each part is also considered.

Figure 18. Guide Pulley Rod [GPS201]

The guide pulley rod is a hollow steel square extrusion with a half inch width. This rod holds the rest of the guide pulley assembly and attaches directly to the frame associated with the guide roller system. This extrusion can be commonly found at hardware stores, specifically Home Depot. A standard length of three feet from Home Depot (PRT#: 40290) is sufficient for both guide pulley systems. Additionally, a standard shaft collar from McMaster fits this extrusion perfectly. The rod would need to be cut to length and possibly coated for resistance against slurry.

Figure 19. Guide Pulley Collar [GPS202]

51

The guide pulley collar clamps to the guide pulley rod and is attached to the guide pulley bracket. The two halves are separable, allowing the collar to be adjusted along the length of the rod so that the wire coming off of the guide pulley aligns with the correct groove on the guide rollers. This is a slightly modified standard part from McMaster. It meets the customer requirement of allowing for an adjustable position for the guide pulley. A standard shaft collar (PRT#: 3325K15) will be purchased from McMaster and the following modifications will be made:

1. Re-tap existing holes with M6x1 threading2. Increase bores sizes to 0.43 inches in diameter3. Drill two holes in two parts as shown in drawing and tap with M6x1 thread

Figure 20. Guide Pulley Bracket [GPS203]

The guide pulley bracket attaches to the guide pulley collar and holds the pulley assembly. The bracket is designed to position the pulley symmetrically to distribute the tensile load evenly and to ensure that the adjustable cap screws are easily accessible. This is a simple part, but it is designed with ease of manufacture and usability in mind. The slotted sections allow for the entire bracket to be removed and for the collar to be adjusted easily. With the pulley hanging below the rod, both the cap screws and the pulley are conveniently accessible for observation and adjustment. The part is designed to be cut out of a sheet of steel using a water jet, CNC router or laser cutter and bent to shape. A recommended bend radius is included for the specified sheet thickness. This part is the most at risk when it comes to stress and displacement under tensile forces. However, the finite element analysis shows values that are not unreasonable. It is possible to modify this part by increasing the thickness to ensure durability.

52

Figure 21. Guide Pulley Cap Screws [GPS001]

This standard part allows the position of guide pulley collar to be adjusted. It also attaches the guide pulley bracket to guide pulley collar. According to customer requirements, it is adjustable using standard metric hex key. A standard M6x1 cap screw is nearly perfectly equivalent to a ¼-28 cap screw. The guide pulley collar is a standard part from McMaster with tapped holes for ¼-28 cap screws. Standard M6x1 cap screw will be purchased. No modifications are necessary.

Validation:Physical tests of over tapping ¼-28 thread with M6x1 thread proved to be successful. A half inch thick block of steel was used.

A COMSOL model demonstrates that the system is subject to minimal displacement under extreme loading conditions. Stress levels are also within allowable limits.

53

Summary:It has been determined that a ¼-28 thread can be converted to a M6x1 thread and subsequently the shaft collar from McMaster that was selected is suitable for use in order to ensure that customer and engineering requirements are met. All components are easily purchased and manufactured.

Moving Forward:After purchase and assembly of these components, testing of structural integrity must be performed. The guide pulley system must be attached directly to the frame of the guide roller system such that the pulley lines up tangent to the rollers.

54

Guide Pulley EncoderP16602: Glass Cutting Machine Wire Supply, Movement and Take-up J.D. IncavoApril 30, 2016

Contents: Overview: Key Assumptions: Rationale: Moving Forward:

Overview:The purpose of the guide pulley encoder is to get reliable feedback on the line velocity of the wire. This cannot be done with just the angular position and angular velocity feedback from the spool motors, due to the changing radius of the spools.

Key Assumptions: No wire slip on the pulley (ideal conditions) Wire velocity on the guide pulley is approximately equal to the wire velocity at the guide rollers,

but is not equal to the wire velocity at the spool motors (on other side of dancer system)

Rationale:As stated above, the radius of the motor load (in our case, the spools) is variable over time in winder or spooling applications. As such, we cannot know the tangential wire velocity coming off the spools without some knowledge of the radius, or additional feedback. This is because the motors only give rotational feedback – i.e. angular positon and angular velocity – which relates to tangential velocity by the radius. However, measuring the angular velocity of something with a known radius would allow us to determine the wire velocity.

v p=ωp R p

We decided to put an encoder on the guide pulley because it is stationary while the system is running. However, putting it on the guide pulley does introduce some problems – namely, it is on the other side of the dancer system. As such, it is more likely to represent the wire velocity at the guide rollers in the integrated system, rather than the wire velocity coming off the spools. The spool wire velocity can be calculated with this information and knowledge of the dancer velocity (see Motor Control & Dancer Pulley).

Additionally, there is also a precedent for similar ideas. Research on winder applications, like ours, shows that they use an in-line driven NIP roller to measure the line velocity in speed-controlled center-driven winder applications.

55

Figure 22. Example of winder application (Source: http://www.carotron.com/applications/csag-h/)

Additionally, a quadrature encoder could potentially be used to determine if there are wire breaks, by detecting if the guide pulley is moving in the opposite direction of the spool. However, this would be less effective than checking the dancer extension – in the case of a wire break, the dancer would extend to its full length.

The guide pulley encoder could also detect if the motor is failing to turn the spool (from a poor connection), if the motor feedback is showing the motor turning while the guide pulley encoder is not turning.

Moving Forward:The guide pulley encoder is yet to be specified – there are a few options in the decision matrix, although this was after limited research. Ideally, the encoder would be Ethernet-enabled for easy compatibility with the PLC – otherwise it would likely need external signal conditioning. Another consideration is that having quadrature output would allow us to determine the direction of rotation; however, this may be unnecessary. Another option that could be considered is moving the encoder to a different in-line pulley, between the spool and the dancer pulley.

56

Supply SpoolP16602: Glass Cutting Machine Wire Supply, Movement and Take-up Jacob Klaus and Sarah BrownApril 28, 2016

Index Scaling for Wire Supply Spool

Contents: Overview: Key Assumptions: Analysis: Summary:

Overview:This section explains the scaling behind the supply spool and the amount of wire that it contains. The supply spool is being scaled to last the same amount of time as the current supply spool. Less wire is required since the number of cuts is being scaled down.

Key Assumptions: On average 3.6 km of wire are used for cutting per job currently On average 2.6 km of wire are used for setup per job currently On average 1.3 km of wire are used due to wire breaks per job currently On average 0.1 km of wire are wasted per job currently This wire use scales linearly with the amount of wire that is used overall on a job based on the

number of cuts that the job is doing

Analysis:Current

Cutting SetupWire Breaks Waste

Current 3.6 2.642 1.321 0.141 7.704 475kmMax 0.972 0.715 0.358 0.012 2.057 125kmMin 0.18 0.134 0.067 0.001 0.382 25km

57

Cutting Setup Wire Breaks Waste0

0.51

1.52

2.53

3.54 3.6

2.642

1.321

0.141

Current case wire breakdown

Wire Use

km o

f wire

use

d pe

r job


0.2

0.4

0.6

0.8

1

1.20.972

0.715

0.358

0.012

Max case wire breakdown

Wire Use

km o

f wire

use

d pe

r job

58


0.020.040.060.08

0.10.120.140.160.18

0.2 0.18

0.134

0.067

0.001

Min case wire breakdown

Wire Use

km o

f wire

use

d pe

r job


0.5

1

1.5

2

2.5

3

3.5

4

Wire Use Breakdown

CurrentMaxMin

Wire Use

km o

f wire

use

d pe

r job

59

Summary:The wire spool is scaled so that the spool can be changed at the current interval of every 3 months. The max case represents the amount of wire needed when doing 270 cuts and the min case represents the amount of wire needed when doing 30 cuts. The spool is scaled to the max case assuming that every job is run with 270 cuts.

Supply Wire Info

Contents: Overview: Key Assumptions for Research/Specification: Summary of Correspondence: Purchase Proposals: Moving Forward:

Overview:One of P16602’s main goals is to move and control the wire within an industry-grade glass cutting machine, in order to test and prove the full functionality of the team’s fixture, wire would eventually need to be purchased. This document summarizes the research and decisions that went into moving forward with identifying the supply wire to be used for the glass cutting machine.

Key Assumptions for Research/Specification: The same diameter of wire that is used on the MB DS264 by Glass Fab, 175 μm steel cord wire

was to be used The same provider of wire that Glass Fab currently uses, Bekaert, would be used as the supplier The same basic principle of a supply side wire spool that contained x length of wire and a take-

up spool that used mx length of wire (where m is a positive multiplier) would be used as in the MB DS264

The supply wire spool would contain 127 km of wire The BS80/33 Spool was chosen as the supply wire spool Designs for the supply wire side mounting systems was designed around the specifications of

the BS80/33 Spool

Summary of Correspondence:Summary of Email Correspondence with Wire Supplier as of April 10, 2016:

60

Technical Specification Sheet:Wire Spec Sheet provided to team on December 1, 2015:

62

Purchase Proposals:A summary of the 2 proposals received from Slicing Tech (Bekaert) is provided in the table below along with a copy of each proposal:

65

Moving Forward: The team will need to purchase wire in order to test the functionality; proper tensioning, control

over spooling/un-spooling, load/un-load of spools; of the wire movement fixture A decision with Glass Fab will need to be made as to what quantity of spools the team wants to

initially purchase Reconnect with Rick Baarck at Slicing Tech to redo proposals (proposals were only valid for 30

days) Work with Rick Baarck to purchase wire Test the wire movement fixture

68

Take-Up SpoolP16602: Glass Cutting Machine Wire Supply, Movement and Take-up Hannah MiccaApril 27, 2016

Contents: References: Key Assumptions: Analysis:

Equations: Process:

Summary: Moving Forward: References:

Overview:Once it was realized that the Meyer Burger Take-Up spool was made from cast iron it was clear

that something was going on to require such a robust spool and after some investigation and thought experiments it was clear that the take-up spool posed an interesting problem to solve. The wire spooled on the take-up spool is all at about 25 N of tension while cutting, this high tension causes a buildup of pressure on the spool drum as more and more wire is stored on the take-up spool. After speaking with subject matter experts and completing extensive research it was clear that characterizing the pressure and state of stress of the drum would not be easy. From the research, equations were found to get the surface pressure produces based on tension, diameter and number of wraps. By modeling the spool drum as a thick walled pressure vessel, this pressure could be used to find the state of stress in the spool drum in order to properly select materials and size the spool to decrease weight, inertia, and cost.

Key Assumptions: Wire is spooled at a constant tension of 25N Wire has a constant diameter of 0.175 mm Total wire length held on a full take up spool is 254 km (2 times that of the supply spool) Perfect layer wind on the take-up spool Compressive strength of cast iron from Wisco Alloys, Inc. ~1034 MPa [1] Calculations are an overestimation of drum pressure State of plane stress in the drum (no stress in the axial direction) Modelled as a thick walled pressure vessel

Analysis:

Equations:1. Drum pressure calculation [2]

o P is the uniform pressure on the spool drumo T is the wire tension during spoolingo D is the spool diameter

69

o d is the wire diametero x is the radial increase per layer (see Figure 23)o N is the number of layers o i is the current layer (zero indicating the first layer)

Ptotal=∑i=0

N 2T( D+i 2 x ) d

EQN (1)

Figure 23. Schematic showing a section of a perfect wire layer wind on a spool used to calculate the radial increase per layer, x

x=d∗sin(60 °¿)¿ EQN (2)

2. Stress calculation in a thick wall cylinder using polar coordinateso σθ is the stress in the angular directiono σr is the stress in the radial directiono σz is the stress in the axial directiono P0 is the external pressure calculated using EQN (1)o r0 is the outer radius of the spool drum o ri is the inner radius of the spool drum

EQN (3-5)

EQN (6-8)

3. Von Mises stress calculations [4]

70

dx

Figure 24. Stress equations taken from [3]

[ ( σθ−σr )2+( σr−σ z )2+(σ z−σθ )2

2 ]=σVon Mises EQN (9)

Process:1. Set up an excel spreadsheet to calculate drum pressure for various number of layers, this

requires an initial estimate for the outer diameter and length of the spool drum 2. Change parameters (outer diameter and length) until reasonable pressures are found3. Use the pressure found for the number of layers needed to store 254 km of wire in the stress

calculations4. Set up a second excel spreadsheet to calculate principal stresses and the Von Mises stress for

different inner radii ranging from 0 – 0.1 meters. 5. Select the largest possible inner radii while still maintaining stresses less than the compressive

strength of cast iron6. Ensure mass and inertia are minimized for motor sizing7. Find the diameter of a full spool and add about 8 cm to find the flange diameter (this ensures

that the flange extends 4 cm past the last wire layer leaving room for an improper layer wind)

Summary:After completing all the calculations and analysis explained above, the take up spool drum must

be 0.12 m in outer diameter, OD, 0 m in inner diameter, ID, and 0.4 meters in length, L, in order to achieve best results (low pressure, low weight, low cost, and reasonable diameter). This results in a drum pressure of about 431 MPa and a max stress of 862 MPa which guarantees a factor of safety of at least 1.2. In order to maintain a robust flange the flanges will be 2.54 cm thick, Ft, and 0.28 m in diameter, FD. This recommended spool is over designed given the pressure calculation approach. With these spool dimensions, the cast would be about 0.01 m3 in volume and about 60 kg (~130 lbs.) in mass. When taking manufacturing into account a three part assembly was created to be able to machine the drum from cast iron bar stock instead of creating an entire casting.

71

Figure 25. Schematic showing the basic cast design of the take up spool

Figure 26. CAD model of final proposed spool design.

72

Moving Forward: Look over all annotated drawings of designs for spool parts Ensure all tolerances and fits are thought through and included in the CAD parts Create fully dimensioned drawings of all spool parts and achieve design approval from

both Ed Hanzlik and the ME machine shop personnel Once designs are approved and material is bought, parts must be machined and the

spool must be assembled Runout tests should be conducted to ensure the spool will not create too much

vibration near the natural frequency of the machineo If vibrations are a large concern consider isolating the wire movement frame

pieces from the rest of the system and implement dampers Tests should be run to ensure the spool can hold at least 2 supply spools worth of wire

(254 km) and to ensure that the slot included is deep enough to cut used wire off

References:[1] http://www.wiscoalloys.com/gray-iron-bar-stock.html [2] https://www.scribd.com/doc/104317798/Wire-Rope-User-s-Manual-AISI# [3] http://courses.washington.edu/me354a/Thick%20Walled%20Cylinders.pdf [4] http://www.learnengineering.org/2012/12/what-is-von-mises-stress.html

73

http://www.learnengineering.org/2012/12/what-is-von-mises-stress.html

http://courses.washington.edu/me354a/Thick%20Walled%20Cylinders.pdf

https://www.scribd.com/doc/104317798/Wire-Rope-User-s-Manual-AISI

http://www.wiscoalloys.com/gray-iron-bar-stock.html

Annotated DrawingsTake-up Spool

Take-up Spool Drum

74

Take-up Spool Flange

75

Take-up Spool Locating Cone

76

Take-up Spool Clamp Plate

77

Supply Spool Locating Cone

78

Supply Spool Clamp Plate

79

Frame

80

IndexFigures(Ctrl + Left Click to follow to figure)Figure 1. Process Flowchart for ApprovalFigure 2. Functional DecompositionFigure 3. Initial Morph ChartFigure 4. Controls Architecture of the SystemFigure 5. Example Wire Velocity Profile (Used in current Glass Fab configuration)Figure 6. Wire Velocity Profile with time labels (P = 0)Figure 7. Illustration of spool and traversing pulleyFigure 8. Spool limitsFigure 9. Example Acceleration Profile with Traversing Pulley near end of spoolFigure 10. Fixture CAD with Dancer Pulley System labeledFigure 11: Diagram of Dancer Pulley System ReactionsFigure 12. Bracket #1 designFigure 13. Bracket #2 designFigure 14. Constructed Dancer Pulley Mounting SystemFigure 15. Electrical System OptionsFigure 16. Cabinet LayoutFigure 17. CAD model of Guide Pulley assemblyFigure 18. Guide Pulley Rod [GPS201]Figure 19. Guide Pulley Collar [GPS202]Figure 20. Guide Pulley Bracket [GPS203]Figure 21. Guide Pulley Cap Screws [GPS001]Figure 22. Example of winder application (Source: http://www.carotron.com/applications/csag-h/)Figure 23. Schematic showing a section of a perfect wire layer wind on a spool used to calculate the radial increase per layer, xFigure 24. Stress equations taken from [3]Figure 25. Schematic showing the basic cast design of the take up spoolFigure 26. CAD model of final proposed spool design.

TablesTable 1. Wire Velocity ProfileTable 2. Parameterized Cut Cycle TimeTable 3. Traversing Pulley System Variables

87