Fulbright, DeMars, Dunsmore, Rosplock, Roth

Caser Side Pusher (“Pusher”, Project 17)

ME 4054W, Spring 2017

05/08/2017

Allen DeMars, Charles Dunsmore, Nathaniel Fulbright, Michael Rosplock, Tyler Roth

Designed for Cannon Equipment

Page 1 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Executive Summary

Problem Definition

Cannon Equipment manufactures machines that package milk jugs and cartons into milk
crates. This casing machine (“caser”) has a pusher station on each side that is responsible for
pushing milk jugs from an infeed conveyor to the center of the machine for further processing.
The current design requires 4 seconds to complete the cycle and Cannon Equipment would like
to decrease that amount of time. Attempts to decrease the cycle time with the current pusher
design have led to instability problems with actuator screw whip and to part damage due to end-
of-stroke impacts. These problems warranted a new pusher design that was engineered for higher
velocities to reduce cycle time. A decrease in cycle time of the pusher would increase the
productivity of the end user of the machine, thereby increasing their profits. Specifically, the new
pusher needed to achieve an extension speed of 24 in. s-1, a retraction speed of 50 in. s-1, a
maximum acceleration of 150 in. s-2 under load, to be designed with IP67 wash-down rated
components, and to cost less than $5,000. Having an easily variable stroke length was an
additional requested feature. Out of scope was anything relating to the incoming milk, the design
of the caser frame (apart from the portion to which to pusher mounts), or optimizations to the
linear guide mechanism.

Design Description

The linear actuation of the pusher was accomplished using a belt drive system, specifically
a belt and pinion system with a stationary motor. A servomotor run through a gearbox controls
the motion of a timing belt using a toothed pulley. The belt is clamped onto the pushing face and
the crosshead, allowing a tensile force to be applied on the pusher in both directions (belts cannot
support a compressive force). The pusher is constrained to move linearly through the use of linear
bearings and guide rods. Bumpers were installed on the guide rods to absorb any accident
impacts from any overshoot that may occur. The pushing face and the wall move simultaneously,
with the pushing face pushing the milk containers and the wall holding back incoming containers.
After the new pusher is assembled and installed, Rockwell Automation control software
will be used to operate the pusher with the desired motion characteristics. Cannon Equipment will
modify existing motion profiles for various milk containers to take advantage of the faster
retraction speeds of the new design. If the pusher is being retrofit onto old casers, simple
modifications to the caser frame must be made to allow for the belt to pass through the mounting
platform.

Evaluation

All design requirements were met or exceeded for $4,200, a price significantly lower than
the current design. The easily adjustable stroke length coupled with higher speeds of 50 in. s-1
and lower lead times make for a design that is superior in every aspect. Tests showed that the
velocity tracking was essentially perfect, ruling out any concern that the compliance of the belt
would lead to poor tracking, oscillations, or overshoot. Stress calculations showed that no
components would fail from cyclic loading (peak stress of 35 MPa, peak deflection of 0.0005 in.),
meaning dairies will have no unexpected downtime from unexpected failure. The only wear
components – the belt, pulleys, and linear bearings – can all be replaced easily, quickly, and
independently well within the scheduled 4-hour daily downtime for each machine. The idler pulley
bearings will last 20,000 hours with a dynamic load capacity of 1,138 lbf.

Page 2 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Table of Contents
1. Problem Definition .................................................................................................................. 4
1.1 Problem Scope ............................................................................................................ 4
1.2 Technical Review ......................................................................................................... 4
1.3 Design Requirements .................................................................................................. 6
2. Design Description ................................................................................................................. 8
2.1 Overview .......................................................................................................................... 8
2.2 Detailed Design Description ............................................................................................. 9
2.2.1 Variable Stroke Length ............................................................................................... 9
2.2.2 Variable Displacement and Speed Extension ............................................................10
2.2.3 Quick Return Stroke ..................................................................................................10
2.2.4 Hold Back Incoming Containers ................................................................................11
2.2.5 Other Belt Drive Components....................................................................................11
2.3 Use ..................................................................................................................................12
3. Evaluation .............................................................................................................................13
3.1 Overview .........................................................................................................................13
3.2 Prototype .........................................................................................................................13
3.3 Testing and Results .........................................................................................................13
3.3.1 Velocity .....................................................................................................................13
3.3.2 Washdown Safe ........................................................................................................14
3.3.3 Infinite Fatigue Life ....................................................................................................15
3.3.4 Cost ..........................................................................................................................16
3.4 Assessment .....................................................................................................................16
3.5 Next Steps .......................................................................................................................17
4. References ...........................................................................................................................18
Appendix A: Technology Review Support .................................................................................19
Appendix B: Design Requirements............................................................................................20
Appendix C: Alternative Concepts and Concept Selection ........................................................21
Appendix D: Design Supporting Documents .............................................................................23
Appendix E: Engineering Analysis.............................................................................................24
Appendix F: Evaluation Reports ................................................................................................26
Velocity Tracking ...................................................................................................................26
Load Capacity .......................................................................................................................27
Infinite Fatigue Life ................................................................................................................28
Appendix G: Cost Analysis ........................................................................................................31
Appendix H: Failure Modes and Effects Analysis ......................................................................32
Appendix I: Environmental Impact .............................................................................................33
Appendix J: Codes, Standards, and Safety ...............................................................................35

Page 3 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

1. Problem Definition

1.1 Problem Scope

Cannon Equipment manufactures machines that package milk jugs and cartons into milk
crates. This casing machine (“caser”) has a pusher station on each side that is responsible for
pushing milk jugs from an infeed conveyor to the center of the machine for further processing.
The current design requires 4 seconds to complete the cycle and Cannon Equipment would like
to decrease that amount of time. Attempts to decrease the cycle time with the current pusher
design have led to instability problems with actuator screw whip and to part damage due to end-
of-stroke impacts. These problems warranted a new pusher design that was engineered for higher
velocities to reduce cycle time. A decrease in cycle time of the pusher would increase the
productivity of the end user of the machine, thereby increasing their profits. Specifically, the new
pusher needed to achieve an extension speed of 24 in. s-1, a retraction speed of 50 in. s-1, a
maximum acceleration of 150 in. s-2 under load, to be designed with IP67 wash-down rated
components, and to cost less than $5,000. Having an easily variable stroke length was an
additional requested feature. Out of scope was anything relating to the incoming milk, the design
of the caser frame (apart from the portion to which to pusher mounts), or optimizations to the
linear guide mechanism.

1.2 Technical Review

Advancements in automation have brought about increased efficiency, reduced cost, and
better quality products in general. In the dairy industry, bottles are automatically filled with milk
and transferred to a casing machine (see Figure 1.1) where they are packaged for shipment.
Modern filling machines can fill at rates up to 150 gpm [1], but Cannon Equipment’s current caser
has a throughput of only 120 gpm, limited by the speed of the pusher (see Figure 1.2). Packaging
efficiency is measured in percent of full-scale-output of the filler [2], meaning high-output dairies
can only be 80% efficient in packaging milk with current caser technology.

Extension

Figure 1.1: Milk casing machine (“caser”). The milk cartons enter from the right on
a conveyor belt. Note the box extending from the doors of the caser; this is the
portion of the pusher that sticks out of the machine (the “extension”).

Page 4 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Head

Pusher

Figure 1.2: Internal view of the caser. The pusher (bottom) pushes milk containers to the center where
the “head” grabs and lowers them into crates. Currently, the pusher is too slow in the return stroke,
increasing the time before another load of containers can be placed and pushed.

Current beverage crating technology ranges from completely manual product handling to
fully-automated crating systems [3]. For production scale beverage manufacturing and distribution
operations, the crating process employed by a manufacturer can fall anywhere within this
automation continuum based on factors such as production volume, OSHA requirements, or
capital equipment budget limitations. Smaller production operations will typically employ a mix of
manual and automated crating technology in order to compromise between budgetary constraints
and production volume [4], whereas large scale production will typically employ fully automated
systems such as the project application at hand.
Numerous technologies have been applied to the beverage crating industry to reduce or
remove the manual labor element for economic and health reasons. Some of the different motion
control methods that are currently employed in this application include: pneumatic actuators [2],
mechanical [5], electro-mechanical [2], linear electric motors [6], hydraulic actuators [2], moving-
coil actuators [7], gravity-fed or driven conveyors/sorters [8], and synchronous belt drives [9].
More exotic methods such as rigid chains [10] exist but are generally cost prohibitive or unproven
for this application. Each of these technologies has advantages and disadvantages in this
application1.
The current design implements either a pneumatic cylinder or a servo motor-driven ball
screw for the linear actuation required to push the beverage cartons into the gripper mechanism
that ultimately lowers and places the cartons into the shipping crates. These two linear motion
generators are well suited to the application with regard to durability and force output but each
has its disadvantages. The pneumatic actuator is very economical but it has a significant
drawback in its lack of position control, whereas, the servo motor-driven ball screw is particularly
expensive and far more precise than the application demands. Also, neither design is capable of
the actuation velocities required. Due to the caustic wash-down environment and propensity for
product spillage to build up on mechanisms, simple and economical polymer bearings on guide
rails and sealed actuators are used.
This project focuses on a specific process in the handling of food product wherein
government regulations pertaining to the environment in which the food product is handled and
the sanitation regimen and frequency are closely monitored and enforced. The U.S. Food and

1 Details about specific sources can be found in Appendix A.

Page 5 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Drug Administration Code of Federal Regulations (21CFR110.40) [16] sets requirements that
specify food handling equipment that does not come into direct contact with food must be
constructed in a manner that supports a sanitary production environment, must be made of a
materials that can be maintained in sanitary condition, and is capable of withstanding proper
sanitization procedures. In order to satisfy this requirement, the equipment used in the food
production environment is typically manufactured from non-corrosive materials such as stainless
steel and non-porous polymers. These materials are selected due to their resilience to the harsh
chemicals utilized in the sanitization procedures [11] that are performed daily. Additionally, the
sanitization process used in this environment normally requires the use of specialty bearings,
guides, and Ingress Protection (IP) [17] rated electronics and control systems. An IP 66 rating is
typically standard protocol in which the sensitive equipment is protected from total dust ingress
and high pressure water jets from any direction.
Handling liquid food products in solid containers often requires special physics
considerations in regards to the dynamics of the liquid contents when accelerated due to slosh
effects of a free surface liquid in a container [18]. Fortunately, pushing cartons of liquid in
containers that are filled to near capacity, such as in this case, allows for significant simplification
of the physics calculations since the center of gravity of the liquid is not changing significantly with
respect to the container volume and the free surface sloshing effects are greatly minimized. This
permits the assumption that the fluid-filled containers are simple solid objects where basic point
mass physics models may be applied and greatly simplifies the calculations to determine the
forces required to safely accelerate the product without damage or tipping. Of primary concern in
the pushing operation is tipping the milk containers. Depending on the container being pushed,
the extension speed is as high as 24 in. s-1 (for gallons) or as low as 12 in. s-1 (for quarts) [2].
With the current design of the pusher face, there is no maximum acceleration to prevent tipping,
rather the maximum acceleration is limited by the power of the motor. Acceleration at the end of
the push varies based on the container being pushed, the number of containers already pushed,
and the friction of the sliding surface [2]. The maximum number of containers to be pushed is 8
gallons, and the total time to push, lower, and feed in all 8 gallons must be under 4 seconds to
obtain a throughput of 150 gpm [2].

1.3 Design Requirements

The primary design requirements for the pusher were provided by Cannon Equipment.
Additional input was received from assemblers of the casing machines and from machine
maintenance staff. Direct feedback from Cannon’s customers was not available, but customer
feedback from previous iterations was provided by Cannon. The current solution needs only minor
modifications, allowing design requirements from previous iterations to be modified.
1. Cost: The cost of the pusher must be less than $5,000. This is cost of the current design,
and it is desired to have the cost closer to or less than $3,000. Cost is a driving factor, but
high speed is preferred over low cost.
2. Size: The pusher must fit within the area that the current design uses. The height of
components that pass through the doors must be less than 3 in. The total height of the
equipment outside of the caser (the “extension”) should be less than 6 in. The width of the
extension should be less than 1 ft. The extension should protrude no more than 15 in.
from the caser. All dimensions are only maximums, and the size of the extension should
be as small as possible.
3. Speed: The bottleneck of the current pusher is the maximum speed, specifically during
the return stroke. The new pusher must be able to retract at speeds of at least 50 in. s-1
and extend at 24 in. s-1. The acceleration profiles are a design variable, but accelerations

Page 6 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

of at least 150 in. s-2 are desired. The acceleration profiles should be chosen such that the
total cycle time for the entire casing process of 8 gallons takes 3 seconds.
4. Loads: The pusher must be able to handle axial loads of at least 60 lbf. This corresponds
to accelerating 8 gallons of milk at 150 in. s-2 (the maximum amount of milk that will be
pushed) with friction accounted for. Additionally, the pusher must be able to handle a side-
load of at least 40 lbf with light impact loading.
5. Life/Materials: The pusher is subject to wash-down environments and is run 20 hours a
day, 7 days a week. Calculated bearing life should be at least 20,000 hours. All other parts
should have infinite life (or approximately infinite life if non-steel components are used).
Due to the caustic nature of wash-down environments, stainless steel is the preferred
material. Anodized aluminum can be used on non-contact surfaces, but its use should be
avoided [11].
These design requirements have been summarized in Table 1.1.

Table 1.1: Subset of the design requirements representing the most important metrics. The full table can
be found in Appendix B.
Need Metric Importance Units Marginal Ideal Value
#s Value
1 Cost 3 USD <$5,000 $2,000 - $3,500
Maximum acceleration 2
2 4 in/s 100 150
under load
2 Extension velocity 4 in/s 24 24
2 Retraction velocity 5 in/s 50 >50
SS or
3 Material 5 N/A anodized SS
aluminum

Page 7 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

2. Design Description

2.1 Overview
The linear actuation of the pusher was accomplished using a belt drive system, specifically
a belt and pinion system with a stationary motor. An outline of the design is shown in Figure 2.1.
A servomotor run through a gearbox controls the motion of a timing belt using a toothed pulley.
The belt is clamped onto the pushing face and the crosshead, allowing a tensile force to be applied
on the pusher in both directions (belts cannot support a compressive force). The pusher is
constrained to move linearly through the use of linear bearings and guide rods. Bumpers were
installed on the guide rods to absorb any accident impacts from any overshoot that may occur.
The pushing face and the wall move simultaneously, with the pushing face pushing the milk
containers and the wall holding back incoming containers.
Design alternatives and the methodology for selecting the final design are given in
Appendix C.

Pushing face Wall

Linear
bearings Bumpers

Guide rods

Belt Crosshead

Figure 2.1: Final design. Shown in the upper-right is the pusher in the full caser frame, and in the
bottom-left is the pusher without any of the caser components.

Page 8 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

2.2 Detailed Design Description

A high-level overview of the pusher
design is shown in a functional block diagram in
Figure 2.2. The key functions of the pusher are
a variable stroke length, a variable
displacement and speed extension stroke, a
quick return stroke, and the ability to hold back
incoming containers. The first three functions Figure 2.2: Functional block diagram.
were all accomplished by using a servomotor-
driven belt system. The last function was accomplished using the pusher face and wall. Each
function shown in Figure 2.2 and its respective solution is explained in more detail below. Detailed
calculations can be found in Appendix E. Codes and standards that applied to the product were
researched and applied to the product (see Appendix J).

2.2.1 Variable Stroke Length

Continuously variable stroke length was made possible by the use of a belt to drive the
pusher. Figure 2.3 shows the belt and pinion system used in the final design2. Due to the open
belt design (versus a standard continuous loop), the length of the belt can be increased or
decreased to whatever length is required. By keeping the belt open, no idler pulley was required
behind the crosshead on the guide rods. This meant that only the guide rod and belt lengths
needed to be changed to change stroke length. Guide rods can either be purchased or
manufactured to the desired length, and large rolls of open belt can be purchased and cut to any
length necessary. Increasing the stroke length will decrease performance of the system, as
increased length leads to heavier guide rods and more compliance in the belt; the belt tension
must also be increased, reducing life of other components (e.g., idler pulleys, gearbox) unless
they are also upgraded. However, the range of potential stroke lengths is not large (as low as 8”
to as large as 20”, with the current stroke being 14”), and adjustments to this length are not
anticipated to cause significant issues. The BRECOflex 25 T5/3050 M Kevlar FDA1 belt was used
for this project due to its washdown resistance and strength.

Idler pulleys Guide rods

Crosshead
Drive pulley

Figure 2.3: Final belt and pinion system. The blue line shows the path of the belt.

2 Drawings for all parts and assemblies can be found in in the included file “Pusher Drawings [4.11.17].pdf”.

Page 9 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

2.2.2 Variable Displacement and Speed Extension

The combination of a servomotor along with a feedback controller provides an excellent

method of displacement, speed, and acceleration control. Figure 2.4 shows the motor (Allen
Bradley VPF-B0753F-PJ12AF) and gearbox (Apex AF060-010-S1-P2) on the frame of the caser.
The servomotor drive and Rockwell Automation control software are not shown. Using a
servomotor allows for real-time position and speed feedback and control, as well as the ability to
have different motion profiles for different milk containers.

Servomotor
Gearbox

Figure 2.4: Motor and gearbox mounted on caser frame.

Motor and gearbox selection was performed with substantial assistance from Werner
Electric (Cottage Grove, MN). Loads and speeds were provided to Werner, and their selection
software reported back many possible configurations. An important selection criterion was related
to the concept of inertia mismatch. Generally speaking, inertia mismatch is a quantification of the
ratio of the inertias of the load and the motor. If the ratio is too large, the motor will not be able to
efficiently control the load and instabilities can be introduced [12]. Because the motor must be
able to withstand a washdown environment, it must be fully sealed and have a washdown safe
material or finish. The Allen Bradley VPF series serves this role, with the added consequence of
being a “very low inertia” series of motors. Without a gearbox, the inertia mismatch was over
700:1 [13]; it is generally desired to have a ratio of 10:1 or lower [12]. With a 10:1 ratio gearbox,
the inertia mismatch decreased to a much better value of approximately 11:1 [13].

2.2.3 Quick Return Stroke

Increased retraction speed was the primary goal of redesigning the pusher. By retracting
faster, the total cycle time is decreased, and more milk can be cased. Servomotors can often run
at very high speeds (6000 rpm or greater), making them an excellent candidate to achieve fast
retraction speeds. Use of a belt drive enables retraction at 50 in. s-1 with no added instabilities
(like those seen in ball-screws operating at high speeds). Additionally, the belt driven system is
still able to retract at full speed with the 10:1 gearbox (with a drive pulley diameter of 2 in. or
smaller).

Page 10 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

2.2.4 Hold Back Incoming Containers

Figure 2.5 shows pushing face of the pusher and the wall that is used to hold back the
incoming stream of milk containers. The purpose of the wall is to allow containers to stack up
before the pusher, which allows the pusher to be loaded with the desired number of containers
more quickly. Without the wall, the conveyor would have to start and stop very frequently, causing
losses in efficiency by the delays in starting and stopping the conveyor and by leaving gaps in the
incoming stream of milk.

Pushing face
=
Wall

Figure 2.5: Pushing face and wall of pusher. The pushing face moves the containers, the wall holds
incoming containers back.
The component just explained is Cannon Equipment’s current design and was not
changed because it performs its function effectively.

2.2.5 Other Belt Drive Components

Several other components were required for the belt drive prototype. Figure 2.6 shows
the required pulleys and Figure 2.7 shows the clamps that were used. According to the
BRECOflex timing belt catalog, the minimum number of engaged teeth on a drive pulley for a T5
belt with back bending is 15, and the minimum idler pulley diameter for back bending is 30 mm
[14]. A 32 tooth drive pulley for keyless shafts (MiSUMi TTLK32T5250-E-16) was selected for the
drive pulley, exceeding the required number of engaged teeth (the belt bends more than 180°
around the drive pulley). The idler pulleys were flanged pulleys with sealed ball-bearings (MiSUMi
AFDF32-45) with a diameter of 70 mm. This was double the minimum recommended diameter
and was chosen to reduce the stress from back bending while simultaneously engaging the
correct number of teeth on the drive pulley. A dynamic load capacity of 1138 lbf was required per

Page 11 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

pulley for a 20,000 hour life. A standard belt clamp from BRECOflex (50 x 120 Clamp Type B)
was cut in half and used at both ends of the belt. At least six belt teeth were engaged with the
clamps, in accordance with the specifications listed in the BRECOflex clamp catalog [15]. Belt
tension was adjusted by means of vertical slots in the motor mounting plate.

Idler pulleys Modified belt

clamp

Slots for belt

Drive pulley tensioning

Figure 2.6: Required pulleys and their locations. Figure 2.7: One end of belt shown with
modified belt clamp.
All purchased components for this prototype were either plain aluminum or hard-anodized
aluminum. However, the parts to be used in a true washdown environment would have to be
stainless steel or a chemical resistant plastic, so these parts would not be acceptable. Aluminum
components were selected for the prototype to save cost and to dramatically shorten lead times.
Almost all timing belt components are aluminum, and all stainless components have to be special
ordered, adding to cost and lead time. In bulk, the stainless components are closer in cost to
aluminum components, but are still more expensive. Cannon Equipment will make further
decisions regarding stainless components after long term testing with the prototype. As long as
all dimensions are the same or similar, no noticeable performance differences should be seen
between the two materials. A bill of materials (BOM) can be found in Appendix D.

2.3 Use
After the new pusher is assembled and installed, Rockwell Automation control software
will be used to operate the pusher with the desired motion characteristics. Cannon Equipment will
modify existing motion profiles for various milk containers to take advantage of the faster
retraction speeds of the new design. If the pusher is being retrofit onto old casers, simple
modifications to the caser frame must be made to allow for the belt to pass through the mounting
platform.

Page 12 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

3. Evaluation

3.1 Overview
All evaluation was performed on a physical prototype. Because a physical prototype was
available, all speed, acceleration, and load moving capabilities could be tested. Additionally, ease
of assembly could be tested.
Table 3.1 gives a summary of the key design requirements and the evaluation method for
each.

Table 3.1: Design requirements and their respective evaluation methods.

Design Requirement Value Evaluation
-1
Velocity > 50in s Measurement from Allen Bradley software
Washdown Safe - Analysis of materials
Infinite Fatigue Life - FEA on crosshead
Cost < $5,000 Cost analysis

3.2 Prototype
To test the proposed design, a full working actuator was manufactured and then
assembled by the design team. The pusher was built per the CAD drawings and was intended to
demonstrate the ease of manufacturing and assembly. As such, no photo of the prototype has
been included as it is exactly the same as the images shown in Chapter 2. Few custom
components had to be manufactured by Cannon Equipment, while the rest were purchased
components. The completed pusher was controlled by a milk casing machine that was being
assembled on the production floor. By using a production machine, the pusher was tested as if it
were running in the real conditions. With the help of the controls programmer, the velocity profile
for the pusher was modified to increase the velocities and accelerations from the current, ball-
screw determined values so that the pusher’s cycle time was reduced drastically.

3.3 Testing and Results

Testing was carried out on the most important criteria of the project set forth by Cannon
Equipment. This was done to ensure that the new pusher design would operate under typical
working conditions of the caser machine while also providing improved performance. The tests
and results are given in more detail in Appendix F. The various failure modes have been analyzed
in Appendix H. Appendix I examines the environmental impact of the final design and prototype.

3.3.1 Velocity

The pusher being able to reach 50 in s-1 on the retract stroke was the most important
design requirement. The purpose of the project was to create an actuator that could replace the
current solution that was creating a bottleneck in the milk packing process. To test this
specification, the prototype pusher was controlled by an Allen Bradley controller that is used on
the production milk casing machines. This controller provided both movement instructions to the
motor, and measured the angle of the motor shaft so that the position of the actuator could be
determined in real-time. This allowed for the creation of velocity profiles for the actuator, and then
measure the velocity to determine if the actuator was performing correctly. This information was

Page 13 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

displayed on a real-time plot so that performance of the actuator could be observed. Additionally,
data from the plot was exported in .CSV format for post processing. While the actuator was
running, it was clear that it was performing exactly as it should have. The actual velocity profile
matched the commanded profile almost exactly. This verified that the actuator was reaching 50in
s-1 as defined in the design requirements. Figure 3.1 shows a sample cycle that the actuator was
running. At around 8 seconds the pusher was commanded to move at 50in/s and the measured
velocity matched.

Figure 3.1: Sample cycle for pusher. Note that the desired speed of 50 in s -1 was reached at
approximately 𝑡 = 8 s.

3.3.2 Washdown Safe

Milk packaging machines must be kept very clean to eliminate the risk of contamination.
These machines are subjected to a wash-down process to clean the built-up milk residue. In this
case, a wash-down cleaning procedure is carried out by a pressure washer loaded with a chlorine
bleach solution (or another caustic cleaner). The use of bleach as a cleaning agent presents a
unique problem for machine components that requires them to be resistant to rust and
degradation due to interaction with chlorine. This requires every working surface to be stainless
steel, and non-working surfaces to be made from anodized aluminum or polymers. Care was
taken during the design of the new pusher to follow these guidelines. To verify the requirements,
the BOM was inspected to make sure that every component conformed. After building the
prototype and verifying that Cannon Equipment stocked the correct components, it was verified
that the new pusher design was ready for a wash-down environment. All new components were
stainless steel with the exception of the drive pulley, the idler pulleys, and the belt clamp. These
components were all anodized aluminum due to their lower cost and availability over stainless
components. In the production version, these components will be custom manufactured by
Cannon out of stainless steel.

Page 14 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

3.3.3 Infinite Fatigue Life

It’s important that the machine was designed so that as few parts need to be replaced as
possible. This required analysis to be done on all the new custom components of the machine to
ensure that none of them were subject to loads that could lead to failure. To ensure that none of
the components of the machine would wear out, critical components were analyzed to determine
the stresses they would be under during operation. After this was done, the component that was
expected to receive the worst loading conditions (the cross head) was modeled using
SOLIDOWRKS simulation to get an accurate results for the stresses and deformations that the
component would see. Figure 3.2 and Figure 3.3 show the deformation and stress, respectively.

Figure 3.2: Deformation of crosshead. The maximum deformation was 0.0005 in.

Page 15 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Figure 3.3: Equivalent stress in crosshead. The maximum stress was 35 MPa.

The maximum deformation (0.0005 in.) and the maximum stress (35 MPa) were both very
low. For 304 SS with a yield stress of approximately 200 MPa and an ultimate tensile stress that
is even greater, the conclusion was that the part would have infinite life; see Appendix F for this
justification. The very low stress did suggest that the part was overbuilt and could be reduced in
thickness to save weight and material costs. However, the driving dimension of the crosshead
was the shoulder bolt diameter for the belt clamp, and the size of this bolt was specified by Cannon
and could not be changed.

3.3.4 Cost

A detailed cost analysis for production can be found in Appendix G. For the prototype,
the cost was estimated to be approximately $4,500. This price included labor for machining and
all required components specified in the BOM.

3.4 Assessment
In all aspects, the new pusher was considered a success. All design requirements were
met or exceeded for a price significantly lower than the current design. The easily adjustable
stroke length coupled with higher speeds and lower lead times make for a design that is superior
in every aspect. Tests showed that the velocity tracking was essentially perfect, ruling out any
concern that the compliance of the belt would lead to poor tracking, oscillations, or overshoot.
Stress calculations showed that no components would fail from cyclic loading, meaning dairies
will have no unexpected downtime from unexpected failure. The only wear components – the belt,
pulleys, and linear bearings – can all be replaced easily, quickly, and independently well within
the scheduled 4-hour daily downtime for each machine. The load pushing capabilities of the
machine could not be tested before this report, but all belt and motor components were designed
with a safety factor of 2 to 3 with respect to the maximum possible load. There is still room for
improvement with reduction in number of fasteners and reduction in overall number of

Page 16 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

components, but the prototype design showed that Cannon equipment is very close to having a
function belt-driven pusher.

3.5 Next Steps

Because the pusher has been tested in its actual use case, the next step is to perform
long term cycle testing to ensure that the components will last as long as designed for. This is
very important for machinery that runs almost non-stop. On these types of machines down time
for service is very costly, and needs to be minimized. Before this solution can be implemented
and sold to customers it must first be verified at in house by running the actuator under load for a
long period of time. Along with this, assembly and manufacture instructions for the new actuator
must be generated so that Cannon Equipment’s manufacturing department can reliably produce
this actuator in large quantity to satisfy demand for the project. The design of the crosshead
should be revisited to reduce its cost while maintaining infinite life. One suggestion is to make the
crosshead out of sheet metal instead of machining it to save on material costs and reduce
manufacturing time. Finally, Cannon equipment should source new pulleys and belt clamps made
out of stainless steel for optimal washdown durability.

Page 17 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

4. References
[1] Filter Specialties, Inc., “SLB series specifications,” 2012, from http://filler-
specialties.com/machinery/monobloc-fillers/slb-series/slb-series-specifications/.
[2] Levine, D., 2107, Product Design Manager at Cannon Equipment, Cannon Falls, MN, private
communication.
[3] Schad, J.E., 1963, “Resource Substitution: Automatic Casing and Stacking Versus Manual
Casing and Stacking of Fluid Milk Products,” M.S. thesis, Dairy Technology, Virginia Polytechnic
Institute.
[4] Rogers, A., 2009, “Producing Fresh, Local Milk: The Challenges for Small-scale Dairy
Farmers,” from http://www.motherearthnews.com.
[5] Hunziker, O.F., 1920, Condensed Milk and Milk Powder, Author, La Grange, Il.
[6] LinMot, 2015, “Replacement of Pneumatics by Industrial Linear Motors,” from
http://www.linmot.com/solutions/pneumatic-replacement/.
[7] SMAC, 2017, “Packaging”, from http://www.smac-mca.com/applications/packaging.
[8] Zambelli, 2017, “AP200 Case Packers,” from http://www.zambellipackaging.com.
[9] TB Wood’s, 2017, “Timing Belt Drives,” from http://www.tbwoods.com.
[10] SERAPID, 2017, “Push Pull Systems,” from http://www.serapid.com.
[11] E. Eitel , 2006, “Design for Washdown and Sterilization,” from
http://machinedesign.com/archive/design-washdown-and-sterilization.
[12] Lewotsky, K., 2015, “Understanding the Mysteries of Inertia Mismatch,” from
www.motioncontrolonline.org
[13] Mattson, A., 2107, PLC/HMI & Motion Specialist at Werner Electric, Cottage Grove, MN,
private communication.
[14] BRECOflex, 2017, “Polyurethane Timing Belts,” Catalog Number B212, from
www.brecoflex.com/literature.
[15] BRECOflex, 2017, “Timing Belts Accessories,” Catalog Number B205, from
www.brecoflex.com/literature.
[16] FDA, 2016, “Current Good Manufacturing Practice in Manufacturing, Packing, or holding
Human Food,” Title 21, Chapter I, Subchapter B, Part 110, Subpart C, Section 110.40.
[17] NEMA, 2004, “Degrees of Protection Provided by Enclosures (IP Code),” ANSI/IEC 60529-
2004.
[18] Kassinos, Adonis C., 1997, "The dynamics of liquids in moving containers: Numerical models
for viscous unsteady free surface flows ". Retrospective Theses and Dissertations. Paper 12261.

Page 18 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Appendix A: Technology Review Support

Below are reference numbers with a description of their applicability to the technology
review section. See Chapter 4 for reference citations.

[1] A specific example of liquid beverage filling capacities of a typical liquid dispensing automation
company. The maximum fill rate for 1 gal milk containers matches the industry limit of 150 gpm.
[2] Discussion with project sponsor to reveal any applicable knowledge from prior experience that
could be leveraged in the design process. This discussion included detailing design requirements
and problems with current design to be overcome.
[3] Discusses the benefits of process automation within the carton/casing industries by analyzing
both direct and indirect cost benefits.
[4] This article discusses the budgetary constraints of a small locally operated dairy and the
various challenges that they typically face that influence their capacity to produce.
[5] Explains a mechanical process in which milk bottles are placed in crates circa 1920 automation
technology. This is a textbook example of early process automation and the means of which it
was employed.
[6] Description of linear electric motors and their long-term efficiency over pneumatics. LEMs
have much higher efficiency and precision with a high price to match.
[7] Examples of moving coil technology in manufacturing automation providing a clear
demonstration of the high-frequency actuation capabilities.
[8] An example automation manufacturer incorporating gravity fed casing operations in an
automated case packing system. Numerous instances of complicated automated machines taking
advantage of simple Newtonian physics.
[9] Timing belt drive manufacturer describing many of the benefits of synchronous belt drive
systems for cost effective power transfer.
[10] Rigid chain manufacturer website describing benefits and applications of rigid chain linear
actuator systems.
[11] Article describing the nature of machine washdown and sterilization requirements in
pharmaceutical and food processing environments.
[16] Food and Drug Administration requirements for food handling equipment.
[17] Ingress Protection rating information as provided by ANSI and the National Electrical
Manufacturers Association.
[18] Academic paper studying the effects of free surface sloshing within a liquid in a container.
This article demonstrates the impact of reducing free surface area on sloshing of a liquid within a
container and methods of modeling.

Page 19 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Appendix B: Design Requirements

Cannon Equipment and other sources gave needs they had for the pusher. These needs
have been summarized in Table B1. Each need was ranked with an importance between 1 and
5, with 1 being lest important and 5 being most important.

Table B1: Needs for new pusher design.

Number Need Importance
1 Pusher costs less than current design 3
2 Pusher is faster than current design 5
3 Pusher can survive wash-down environment 5
4 Pusher can run 24/7 4
5 Pusher can push up to 8 gallons of milk 3
6 Pusher can handle side-load of up to 30 gallons 4
7 Pusher does not take up too much space 3
8 Pusher can run on caser power 4
9 Pusher made of primarily purchased components 3
10 Pusher is easy to maintain 4

A simple summary of the needs is that the pusher needs to be fast, durable, reliable, and easy to
maintain. The design requirements that were to be followed to meet the needs are listed in Table
B2; the importance ranking uses the same 1-5 ranking scale as previously.

Table B2: Design requirements for required needs (need numbers reference Table B1).
Need #s Metric Importance Units Marginal Ideal Value
Value
1 Cost 3 USD <$5,000 $2,000 - $3,500
Maximum acceleration
2 4 in/s2 100 150
under load
2 Extension velocity 4 in/s 24 24
2 Retraction velocity 5 in/s 50 >50
2 Extension time 4 s <0.75 0.6
2 Retraction time 5 s <0.5 0.3
SS or
3 Material 5 N/A anodized SS
aluminum
4 Bearing life 4 hours 20,000 40,000
4 Fatigue life 4 cycles 1,000,000 10,000,000
4 Motor life 5 hours 20,000 50,000
5 Axial load capacity 4 lbf 60 80
Bearing radial load
6 4 lbf 40 50
capacity
7 Height (inside caser) 5 in <3 2
7 Height (outside caser) 4 in <6 5
7 Width 3 in <12 11
7 Depth (outside caser) 3 in <15 13
8 Power source 4 VAC 280 & 480 280 & 480
Number of custom
9 3 quantity <15 5
components
Components
Yes, but not
10 replaceable with no 4 binary Yes
required
special tools/skills

Page 20 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Appendix C: Alternative Concepts and Concept Selection

Once the forces on the pusher were known (see Appendix E) were known, research into
different types of linear actuators began. Many different modes of linear motion were considered,
including hydraulic, pneumatic, and ball-screw, all of which methods used by Cannon in the past.
Alternatives such as leadscrews, belt-drives, and linear electric motors were also considered.
After careful considerations of all the options, belt-drives were selected because of their low cost,
high movement speeds, and ability to provide the forces necessary to move the loads fast enough
for the application.
The last major consideration was to ensure the actuator conformed to the form factor
defined by the design requirements. This was important to make sure that the operator could
access the machine during a jam, or for other maintenance. The part count of the assembly was
minimized in as many locations as possible to help reduce complexity and build cost. This has
led to a design that has a low part count, and is smaller and less obtrusive than the old design
while also increasing the performance of the machine.
The final concept was selected based on the factors considered most important by Canon
Equipment and what this team considered most important to lead to a successful design. A
decision matrix (see Table C1) was created to analyze the choices for the final design. After
considering the factors and assigning them importance values, the belt driven actuator was the
clear choice. The team agreed with the results of the decision matrix because the belt drive
actuator is by the far the lowest cost option while maintaining the speed and accuracy
requirements. The additional simplicity of maintenance and ease of adjustments to the stroke
length made the belt drive a very attractive solution.
Table C1: Decision matrix.
Ball Lead Belt Linear electric
Hydraulics Pneumatics
screw screw drive motor
Cost [1 (high) to 10
3 6 8 7 8 1
(low)]
Ease of maintenance
4 3 3 2 4 4
[1 to 5]
Cleanliness [1 to 5] 4 1 2 2 4 3
Stroke length scaling
1 1 1 1 5 1
[1 to 5]
Accuracy [1 to 5] 4 3 1 3 4 3
Max. speed [1 to 10] 4 5 7 3 10 6
Total score [45
20 19 22 18 35 18
possible]

During the design process, two primary proposals were made. One was the first iteration
of the belt drive that was a closed loop belt (Figure C1) and the other was a mechanism solution
(Figure C2). The closed loop belt drive initially won, and was then developed into the belt and
pinion system used in the final design. Several factors drove the team away from the closed loop
design. First, it required an idler pulley to be placed on either extreme end of the stroke, adding
at least 3 to 4 inches to the overall length. Second, if the motor were to be mounted anywhere
else but in line with the rear idler, even more idler pulleys would be required. Finally, closed loop
belts are more expensive and a different PN must be stocked for each stroke length. Using an
open loop system allows for Cannon to buy one large roll and cut it to whatever length they need.
Other belt configurations were considered (for example, having the motor vertical), but these were
ruled out due to space constraints.

Page 21 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

The mechanism solution had several benefits, but cost and complexity outweighed those
benefits. The mechanism was based off of a Sarrus linkage, a spatial mechanism that achieves
perfect linear motion. Depending on the synthesis parameters, the stroke can be adjusted to any
desired length. One of the best benefits was that the entire mechanism collapsed in on itself and
fit perfectly inside the caser, reducing the overall footprint of the caser. However, the drawbacks
were many. The primary limiting factor was the complexity. The mechanism contained many
custom, machined components and many pin joints, requiring many bearings and a significant
amount of machining time while maintaining tight tolerances to reduce slop. Additionally, the
motor would have had to be larger, increasing the cost of both the motor and the servo drive,
while also changing the motor from the current design (the final design used the same motor and
drive as the current design). Finally, because the design did not stick out of the caser frame, the
safety doors of the caser would have to be modified to remove the hole for the guide rods. These
effects combined would lead to an expensive and time consuming retrofit, whereas the final
design is a quick and cheap retrofit because many of the customers’ current components can be
reused.

Figure C1: Closed loop belt drive.

Figure C2: Mechanism solution.

Page 22 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Appendix D: Design Supporting Documents

Given below is the final bill of materials (BOM). The part drawings have been included in
the file “Pusher Drawings [4.11.17].pdf” and are included in printed form after this report.

Page 23 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Appendix E: Engineering Analysis

Analysis of the pusher began with a calculation of the loads on the pusher. Figure E1
shows a schematic of all forces on the pusher during the initial push. The forces on the pusher
were as follows: inertial load from milk being pushed; friction from milk containers being pushed;
inertial load of pusher; friction from the linear bearings; a side load from the incoming milk; and
the friction from the incoming milk. These forces could be summarized as follows:
𝑁𝐴𝑐𝑡𝑢𝑎𝑡𝑜𝑟 = 𝐹𝑀𝑖𝑙𝑘 + 𝑀𝑥̈ + 𝐹𝐵𝑒𝑎𝑟𝑖𝑛𝑔𝑠 + 𝐹𝑆𝐿 (E1)
The milk forces consisted of friction and the inertial forces:
𝐹𝑀𝑖𝑙𝑘 = 𝑚𝑚𝑖𝑙𝑘 (𝑔𝜇𝑘,𝑚𝑖𝑙𝑘 + 𝑥̈ ) (E2)
The bearing forces were treated as dry friction:
𝐹𝐵𝑒𝑎𝑟𝑖𝑛𝑔𝑠 = 𝑚𝑝𝑢𝑠ℎ𝑒𝑟 𝑔𝜇𝑘,𝑏𝑒𝑎𝑟𝑖𝑛𝑔𝑠 (E3)
The friction from the side load was also treated as dry friction:
𝐹𝑆𝐿 = 𝑁𝑆𝐿 𝜇𝑘,𝑆𝐿 (E4)
The side load was dependent on the number of containers 𝑛 in the queue and the coefficient of
friction between the containers and conveyor belt:
𝑁𝑆𝐿 = 𝑛𝑚𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟 𝑔𝜇𝑘,𝑐𝑜𝑛𝑣𝑒𝑦𝑜𝑟 (E5)
For all cases, the milk containers were treated as rigid bodies with the shape of a rectangular
prism.
The peak load depended on the size of the container (gallon, half-gallon, or quart) and the
material (HDPE or cardboard). Each material had different friction coefficients on the different
materials, and each container had a different mass. A spreadsheet was created with all possible
combinations to determine the highest load. For all combinations, the acceleration was 150 in s-2
and the number of containers as a side load was 30 (the maximum number the caser will allow).
After calculations, the maximum load was found to be 65 lbf. This load occurred while pushing 8
HDPE gallon jugs (𝑚 = 8.6 lbm, 𝜇𝑘,𝑚𝑖𝑙𝑘 = 0.2, 𝜇𝑘,𝑏𝑒𝑎𝑟𝑖𝑛𝑔𝑠 = 0.3, 𝜇𝑘,𝑆𝐿 = 0.2, 𝜇𝑘,𝑐𝑜𝑛𝑣𝑒𝑦𝑜𝑟 = 0.15)
with a pusher mass of 40 lbm. Only very rarely will the pusher push 8 gallons, but it was the worst
case loading scenario.

Figure E1: Schematic of loads on pusher during initial push.

Page 24 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

The belt was the next component to be selected. Polyurethane timing belts are widely
available in many configurations, but this application limited the choices somewhat. Note: all belt
selections were performed using BRECOflex’s catalog under Cannon’s recommendations. Due
to the washdown environment, a TPU-FDA1 belt was required (“withstands intermittent cleaning
with diluted cleaning agents” [14]). Kevlar tension members were chosen for their excellent
chemical and corrosion resistance as well as their long-term durability when in back-bending
applications. A T5 tooth profile was chosen due to the excellent availability of T5 components and
the relatively small size. A very common size for timing belt components is 1 inch, and the
continuous load rating for a T5 belt is 336 N/10 mm, or approximately 200 lbf for a 1 inch belt
width. This greatly exceeded the maximum load and was deemed safe for use. For a T5 belt, the
minimum idler pulley diameter is 1.18 in. and the minimum number of teeth on the drive pulley
when there is back-bending is 15. The idler pulleys were chosen to have a convenient bore size,
and a 32 tooth drive pulley was chosen (with a pitch diameter of 2.00 in.).
The loading conditions and application were provided to Werner electric for assistance in
selecting the proper motor and gearbox. The recommended combination was a 0.8 hp servomotor
with a 10:1 gearbox to both multiply the torque and reduce the inertia mismatch between the load
and the motor. The servomotor had a maximum speed of 6600 rpm and a continuous stall torque
of 19 lbf in. With the 10:1 gearbox, the motor specifications as seen by the drive pulley were 660
rpm and 190 lbf in. With a 1 inch radius drive pulley, the maximum load was therefore 190 lbf,
again greatly exceeding the load requirements while allowing the motor to either accelerate more
quickly or last longer, depending on the programming. A maximum linear speed of 53 in s -1 was
available (limited by gearbox specs.), but for short periods this could be as high as 70 in s-1 by
running the motor at its maximum speed and sacrificing some gearbox life. The radial load
capacity of the gearbox was 350 lbf, far exceeding the load due to the peak force and a belt
tension of around 25 lbf.
Bearing calculations were performed for the idler pulley bearings to determine the required
dynamic load capacity. For radial ball-bearings, the dynamic load capacity is:
1
𝐿𝑛 3
𝐶𝐷 = 𝑃𝑘𝑎 [ ] (E6)
𝑟(𝐿𝑛)𝑟𝑎𝑡𝑒𝑑
where 𝑃 is the radial load, 𝑘𝑎 is the application factor, 𝐿 is the life (in hours), 𝑛 is the shaft speed
(in rpm), and 𝑟 is the reliability factor. The rated life and speed are a standard set by the AFBMA,
defined as:
(𝐿𝑛)𝑟𝑎𝑡𝑒𝑑 = 300 hr × 33.33 rpm (E7)
The reliability factor can be calculated as:
2
100 3
(E8)
𝑟 = 4.48 [ln ( )]
% reliability
For a bearing with light impact loading (𝑘𝑎 = 1.35), a radial load of 100 lbf (max. load plus belt
tension), a speed of 500 rpm, a life of 20,000 hours, and a reliability of 99% (𝑟 = 0.21), the required
dynamic load capacity per idler was3:
𝐶𝐷 = 2276 lbf (E9)
Because each idler has two bearings, the required dynamic load capacity per idler bearing was:
𝐶𝐷 = 1138 lbf = 5062 N (E10)

3 These values were selected for long life with high reliability to ensure the bearings would not fail
prematurely and cause down time.

Page 25 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Appendix F: Evaluation Reports

Velocity Tracking
Introduction

The velocity tracking of the pusher was the most important aspect. If the pusher could
track a 50 in. s-1 signal, then it passed the primary test of for the prototype. Though acceleration
and position accuracy were also important, the velocity was the critical parameter to hit.

Methods

The pusher prototype was connected to a caser that was being assembled on the factory
floor; this caser already had all electrical components and had the motion profiles pre-
programmed. The controls programmer adjusted the machine parameters to take advantage of
the increased retraction speed and enabled the data logging feature of the Allen Bradley controls
software. The velocity signal and the measured velocity were reported every 10 ms (i.e., the
sampling rate was 100 Hz). The two signals were overlaid and compared to see if the pusher met
the velocity requirements.

Results

The velocity plots are shown in Figure F1. Notice that the velocity perfectly tracks the
reference, even at sharp turns.

Figure F1: Velocity plots for pusher prototype.

Page 26 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Discussion

The velocity perfectly tracked the reference and reached the desired 50 in. s-1.
Additionally, the measured velocity showed no backlash, oscillations, or overshoot, showing that
the compliance of the belt was a non-issue with no load.

Load Capacity
Introduction

The primary purpose of the pusher is to push milk. To do this, it must have a specific load
capacity. Appendix E details the load calculations, and the maximum load was found to be
approximately 65 lbf. All component sizing was done with a large safety factor for long life, so no
definite results can be made from short term tests. However, if the pusher could push double the
peak load for a small number of trials, it could be confirmed that the assembly was done properly
and that the belt would not slip on the drive pulley.

Methods

To simulate a very high load, an unscientific testing method was used. Cannon equipment
did not have a test area set up for the pusher yet, so no continuous, controlled load could be
applied to the pusher. As an estimation, a person pushing against the face of the pusher was
used as a high load.

Results

The pusher was able to push a large male (approximately 150 lbm) with no slipping of the
belt. The velocity of the pusher did slow, but the reason for this could have been caused by
different aspects.

Discussion

The belt did not slip, which was the primary purpose of the test. This showed that the belt
tension was adequately maintained using the belt clamps and tension adjusting slots. The slowing
of the pusher could have been caused by compliance in the belt or because of poor disturbance
rejection of the controller. Compliance in the belt was unlikely, as the Kevlar tension members
essentially prevent elastic deformation when under the working load limit. More likely was that the
controller did not have enough time to reach its steady-state value after the force disturbance was
applied. This could be a cause for concern because the milk being pushed is essentially a force
disturbance, but the lag in velocity would only be noticed during the acceleration (less than 0.2 s).

Page 27 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Infinite Fatigue Life

Introduction

Any unplanned downtime in the dairy industry leads to lost profits and reduced efficiency.
Some failure is impossible to avoid (such as bearing failure or failure of the belt), but the number
of components that can fail unexpectedly should be as low as possible.
Many design elements from the current pusher were carried over to the new design, and
therefore many of the components of have been tested in real-world conditions and have not
failed or have been redesigned such that they will not fail. Of the custom components that were
created for the new design, the crosshead (see Figure 2.3) was determined to be the component
with the greatest loading. Simulation was performed on the crosshead to determine its fatigue life.
The desired result was that the component would have an infinite fatigue life.

Methods

SOLIDWORKS simulation was used to perform the stress analysis. The material was set
to be 304 stainless steel, the real material for the crosshead. Fixed displacement constraints
(green arrows in Figure F2) were applied at the two guide rod locating holes – the guide rods will
deflect very little laterally due to an axial load – and a load condition (purple arrows in Figure F2)
was applied to the belt clamp holes, totaling 100 lbf (the max. force plus belt tension). Direction
of the force (towards the front face in Figure F2) was determine by the anticipated direction of
force during the initial push.

Figure F2: Mesh and boundary conditions for the crosshead. Green arrows represent a fixed boundary
conditions; purple arrows represent an applied load of 100 lbf.

Page 28 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

A mesh was created, the trial was run, then the mesh was refined and the trial run again
to ensure mesh independence. Aspect ratio was used as the qualifier for a good mesh. The aspect
ratio is a measure of the shape of the element. A standard element has sides with equal lengths,
but during mesh creation the elements are stretched and shrunk in three dimensions to represent
the actual part geometry as closely as possible. If the element is stretched too much, the results
from the simulation lose accuracy and stresses, strains, and other results are not correct. An
aspect ratio of 3 or below is considered acceptable for simulation. SOLIDOWRKS reports the
maximum aspect ratio and the percentage of elements with an aspect ratio under 3.

Results

The initial mesh had 56,330 elements, 84,678 nodes, a maximum aspect ratio of 5.3345,
and had 99.9% of elements with an aspect ratio below 3. With this mesh, a deformation of 0.0005
in. was observed (at the center of the crosshead) and a maximum stress of 35 MPa was seen at
the bolt holes. For a refined mesh with 296,478 elements, 425,599 nodes, a maximum aspect
ratio of 6.66, and a reported 100% of elements with an aspect ratio of less than 3, the same
deformation and stress as before was obtained (within 1%). Figure F3 shows the deformation of
the crosshead and Figure F4 shows the stress distribution of the crosshead. In both figures, the
model deformation has been scaled by 600 for visibility.

Figure F3: The maximum deformation was 0.0005 in at the center. Deformation has been scaled by 600.

Figure F4: The maximum stress was 35 MPa at the bolt holes. Deformation has been scaled by 600.

Page 29 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Discussion

The deformation of 0.0005 in. is essentially imperceptible and will cause no control issues
whatsoever. The maximum stress of 35 MPa was also very low and very localized. The yield
stress of 304 SS is 200 MPa and the ultimate tensile stress is 505 MPa. For steel, infinite fatigue
life is obtained if the equivalent alternating stress is less than the endurance limit. A conservative
estimate of the endurance limit is 25% of the ultimate tensile stress, meaning the endurance limit
for this component was 126 MPa. The equivalent alternating stress can be measured from a
Goodman diagram or analytically:
𝜎𝐴 𝑆𝑢
𝜎𝐴,𝑒𝑞 = (F1)
𝑆𝑢 − 𝜎𝑚
For the loading on the crosshead, the alternating stress was the maximum stress as given by the
FEA and the mean stress was simply half of the alternating stress. The equivalent alternating
stress was therefore:
𝜎𝐴,𝑒𝑞 = 36.3 MPa (F2)
The safety factor with respect to infinite life was:
𝑆𝑛 126 MPa
𝑛= = = 3.48 (F3)
𝜎𝐴,𝑒𝑞 36.3 MPa
It was therefore concluded that the crosshead would have infinite life and would not cause any
unexpected failures. However, the high safety factor with respect to infinite life suggested that the
crosshead was over designed and could be made thinner to reduce material costs. This could be
done if the shoulder bolts connecting the belt clamp to the crosshead were reduced in diameter.

Page 30 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Appendix G: Cost Analysis

Table G1 lists all design costs for this project*. The implemented design expenses will
be identical to development/prototyping cost as all prototyping was completed using production
quality components.

Table
TableG1: DesignImplementation
1: Design implementationCost*
cost.
Line # Description Source Unit Cost ($) Qty (hrs) Cost ($)
Purchased Components
1 Allen Bradley VPF-B0753F Servo Motor Warner Electric $1,000.00 1 $1,000.00
2 Servo Drive Warner Electric $1,500.00 1 $1,500.00
3 10:1 AF Series Apex Gearbox Warner Electric $600.00 1 $600.00
4 T5 Timing Belt Brecoflex $68.54 1 $68.54
5 AFDF32-45: Pulley Misumi $26.57 2 $53.14
6 TTLK32T5250-E-16: Pulley Misumi $82.16 1 $82.16
7 24" L 1" D Guide Rods McMaster $87.50 4 $350.00
8 Linear Bearings McMaster $100.00 1 $100.00
9 90298A583: SHOULDER SCREW McMaster $3.11 4 $12.44
10 93615A410: SHCS, LOW PROFILE McMaster $5.09 4 $20.36
Manufactured Components**
11 PN00004-00 Cannon $90.00 1 $90.00
12 PN00005-00 Cannon $65.00 1 $65.00
13 PN00006-00 Cannon $20.00 1 $20.00
14 PN00007-00 Cannon $25.00 1 $25.00
15 PN00009-00 Cannon $80.00 1 $80.00
16 PN00012-00 Cannon $25.00 1 $25.00
17 PN00013-00 Cannon $20.00 1 $20.00
18 PN00015-00 Cannon $45.00 1 $45.00
19 PN00015-00 Cannon $25.00 1 $25.00
20 PN00017-00 Cannon $35.00 1 $35.00
21 PN00018-00 Cannon $20.00 1 $20.00
Labor
22 Assembly Time (hrs) Cannon $35.00 2 $70.00
Development Cost***
23 Engineering/Development Time (hrs) U of M ME $45.00 (68) ($3,060.00)
Total $4,306.64

*Excludes electrical controls hardware and system controls development.

**Manufactured components cost estimates include material and manufacturing expenses. Labor is billed at $35/hr
for internal work at Cannon.
***Engineering/development cost was omitted from overall project cost per the mentoring arrangement between
project sponsor and the University of Minnesota.

Page 31 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Appendix H: Failure Modes and Effects Analysis

Page 32 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Appendix I: Environmental Impact

Purpose and Need

The purpose of the prototype belt-driven pusher was to provide greater performance,
functionality, and scalability than that of the pusher device currently in use on the milk caser
machine. The new pusher will allow faster cycle times, which in turn increases the overall
efficiency of the machine due to a reduction of idle time by other components. The functionality
and scalability of the design replaces the need for multiple length screw actuators that require
vast resources to machine their internal components to precise specifications.

Impact to the Environment

The new pusher has a minimal impact on the environment. The vast majority of the
materials in the prototype are made from aluminum and stainless steel. The metal components,
except for the bearings, are unlikely to wear out over the lifetime of the product, and these are
easily recyclable after replacement. There are a few components in the design that consist of
plastic, such as the polyurethane belt, inner bearing components, or the urethane bumpers. The
polyurethane can be recycled at the end of its life using mechanical or chemical means, but
customary practice will likely see the components discarded into the trash. An inherent benefit to
the pusher is the lack of any consumable materials, except for a small amount of lubricating
grease needed for the linear bearings and the gearbox.
The pusher is driven by an electric servomotor so there is no extra pollution created by
the prototype. The motor windings of the servomotor are matched to the electrical drive to improve
energy efficiency.
The components that are most likely to wear out and will require replacement are the
timing belt, linear bearings, and idler pulley bearings. These components are easily replaceable
after disassembly with common hand tools by the customer’s maintenance workers.
If the pusher is attached to a new machine there will be no need for extra packaging
materials. However, the pusher is designed to be able to replace a pneumatic or actuator-driven
version that is already in use at a customer’s facility. Shipping the new pusher to the customer
will likely consist of a wooden packing crate, which can be reused or recycled.
There are no known regulations at any state, federal, or international levels that apply to
the product because the machine is not in direct contact with the milk but only the milk container.

Alternatives to the Design

An alternative design that would have a lesser impact on the environment may be to
replace the polyurethane belt with a chain-and-sprocket or a rack-and-pinion design. Additionally,
using metal springs for over-travel shock absorption would eliminate the K-Prene urethane
bumpers. Replacing plastic components with metal components would ensure 100% recyclability.

Discussion

Redesigning the pusher for either a chain-and-sprocket or rack-and-pinion drive would

have an increased cost over the polyurethane timing belt, both in raw material cost and scalability
cost. With the expected long life from the timing belt, it is unlikely the benefits to the environment
would warrant making such a design change. Additionally, one of the desired benefits of the timing
belt was the reduced noise level. Changing to a chain-and-sprocket or rack-and-pinion drive
would increase the noise pollution in the manufacturing environment.

Page 33 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

It is unlikely an end-of-life reclamation program would be warranted for this device. As the
components wear out (bearings and timing belt) they will simply be replaced by a maintenance
worker at the end-user’s facility, with the old components being discarded in the trash or local
recycling bin.
It is believed this new design is the best compromise between cost, efficiency,
performance, functionality, and scalability as opposed to previous designs. The impact to the
environment is minimal over the life of the product because of its recyclable materials, ease of
serviceability and expected long-life.

Page 34 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Appendix J: Codes, Standards, and Safety

The equipment used in the dairy processing industry is regularly washed down with
chlorine bleach or sodium hydroxide. As a result, Cannon Equipment required that purchased
electrical components, such as the servo motor, met an IP66 rating. “Ingress protection” (IP)
codes are established by the International Electrotechnical Commission (IEC) standard 60529
and help to define to what level different equipment and components are protected from solid and
liquid objects [17]. The first number in the rating relates to solid objects as shown in Figure J1,
while the second number relates to water as shown in Figure J2.

Figure J1: IP specifications for the first digit of numeric codes [17]. The first digit specifies resistance
to solid objects or particles.

Page 35 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

Figure J2: IP specifications for the second digit of numeric codes [17]. The second digit specifies
resistance to water.
Other standards were set forth by Cannon Equipment that covered internal manufacturing
specifications. Examples of this standard were using 10-32UNF threaded fasteners instead of 10-
32UNC fasteners, having a minimum of 1/32” radius on all sheet metal parts, using all stainless
steel fasteners where possible, and not using unfinished aluminum materials.
Safety was a concern with this project, however, most of the pusher mechanism will be
located inside an enclosure that is connected to the milk casing machine. When retracted, the
two rods will extend outside the enclosure so a guard was designed to protect personnel from the
hazard. OSHA publication 3170-02R, “Safeguarding Equipment and Protecting Employees from
Amputations,” states that “guards usually are preferable to other control methods because they
are physical barriers that enclose dangerous machine parts and prevent employee contact with

Page 36 of 37
 Fulbright, DeMars, Dunsmore, Rosplock, Roth

them” [J1]. Because this device is a prototype and is meant to undergo long-term testing, Cannon
Equipment determined that a guard would not be necessary to manufacture at this time.

References

[J1] OSHA, 2007, “Safeguarding Equipment and Protecting Employees from Amputations,”
3170-02R.

Page 37 of 37