Customer Handoff & Final Project Documentation
Team Vision for Week 14
- 75% of technical report completed
- 3D printed 1/4-scale model
- Test plans
- Refine documentation
- Prepare Future Design Suggestions
- Final Model Changes and Drawing Creation
- Provide Alternative Motor Option
- Refined functional description to send to Zeller
The motor specialist at Zeller provided a lot of good feedback for the end of life problem. Basically, two motors from Kollmorgen were found to be the closest match to the RDB-21529. Below is a chart comparing the three motors.
Essentially, the CH061A is smaller in terms of torque than the RDB and the CH062C is larger. The CH061A could be undersized bringing the machine cut number down to even 250. The CH062C could be slightly oversized, but was the recommended motor by the motor specialist. The biggest concern with oversizing is a poorer efficiency, but servo motors are usually oversized by 20% for the application in order to keep the motor from overheating. Below are the torque-speed curves for the three motors.
There are a few issues though with getting a Kollmorgen motor and trying to use a Rockwell system. Two options are available:
1) Rockwell Drive with Kollmorgen Motor
- Maintain DC Bus link across all drives on machine
- More compact setup in electrical cabinet
- Requires expensive custom interface cable (500$ each)
- Requires optimization of AC Drive motor control parameters to adjust to different motor
2) Kollmorgen Drive with Kollmorgen Motor
- Optimal performance and warranty achieved with designed drive for motor.
- May take up more space in cabinet
- DC Bus cannot be shared between different drive manufacturers
- Some add-on instructions will need to be added to interface Kollmorgen drives with Rockwell PLC (has been done at Zeller)
The motor selection excel chart can be found here
The Zeller quote for the Kollmorgen motors and accessories can be obtained here
The motor specialist also did more research on the way our team was connecting the direct drive motor and discovered that it was incorrect. A third bearing should be added to support the motor shaft in order to maintain the strict air gap requirement between the motor rotor and stator. Therefore, the current design needs to be refined to properly support he direct drive motor. Another major concern was if the rigid coupling between the guide rollers and motors would negate the direct drive benefits. Fortunately, this should not be a major problem.
The following document was sent to Zeller as a summary of the machine operation in case programming help was ever needed from them. Found here
A Rockwell Motion Analyzer for the RDB-21529 motor for future reference can be found in the EE stuff folder of the Week 14 review. Edge does not recognize the mba. file properly.
Lastly, if Zeller is to be contacted in the future regarding this project, please always include the following project number and name to help them quickly get on track: 160828-RIT Glass Fab
Updated CAD Descriptions
Modified CAD files so that all added features are named to describe what the features use is.
Drawing Package and Descriptions
Guide RollerThe current guide rollers are made such that the End Caps will be welded on. After the initial design in which the end caps were compressed against the guide roller, the machine shop suggested this method as a much cheaper option. This will also be mechanically stronger than relying solely on compressive force to keep the roller together. The biggest thing in terms of machining the roller is ensuring minimal runout and concentricity. This will ensure that there is alignment between the rollers. Concentricity will ensure that the two axis are parallel. Minimizing runout will ensure that the rollers are balanced, which will reduce mass imbalance (and therefore any vibratory stress), create an even cut, and allow for smooth acceleration and deceleration of the roller. Especially given the fact that this part is long with a large diameter, ensuring that this runout is small will be critical, as that will be the most difficult tolerance to ensure. A small lip was added that the End Caps will be pushed up against to ensure that they are welded on at the right location.
The End Caps are meant to fit into the Guide Roller and be welded in place. There is a hole the the center rod will be press fit through. After the roller is assembled by fitting End Caps, Guide Roller, and Center Rod together, the End Caps will be welded to both the Guide Roller and Center Rod. This press fit will both improve mechanical strength and ensure that everything is properly aligned.
The Center Rod is designed to have the End Caps press fit onto it while going through the Guide Roller to offer support through the center. After welding on the End Caps and Guide Roller, the bearings will be press fit on to the center rod, creating the Guide Roller Assembly that will rest on the bearing housing.
The final assembly has four angle plates that are going to be purchased through Tombstone City. All of the angle plates have four 12mm holes in which dowel pins will be press fit in. Two of these holes are placed on the front face and act as locating holes for the bearing housing. The other two are on the bottom of the angle plate and are used as locating holes for the angle plates attachment to the baseplate. For each of these dowel pin holes, another 8mm hole will be drilled into them to allow for easier pin insertion, and removal if necessary. All four angle plates also have four M12 bolt holes to secure the bearing housings on the front face.
The two front side/outboard angle plates have cutouts which are used to create a hole to remove the catch basket. There is an M12 bolt hole above this hole for the attachment of the debris catch. Additionally there are three ¼”-20 tapped holes along the sides to hold the polycarbonate paneling into place. These panels create the viewport for the machinist.
On the two motor side/inboard angle plates there is a large hole bored out of the center of the front face to create room for the motor. Concentric with this hole, there is an indented surface which is used to properly align the motor. This was designed per the specifications of our Rotary Direct Drive motor. There are 4-½” bolt holes drilled into face of the angle plates to secure the motor. Similar to the front side/outboard angle plates, there is also an M12 bolt hole for the attachment of the debris catch, as well as 3 ¼”-10 tapped holes along the sides to hold in the same polycarbonate paneling.
Changes needed - The large cutouts on the front side/outboard angle plates will be changed in the future to allow for large scrap removal and hand access. Additionally, there may be a change in the motor, which may change the bolt holes and indented locating feature on the motor side/inboard angle plates. Depending on what motor is ultimately selected, new features may need to be added or different bolts may need to be used. Additionally, because the polycarbonate panels may not be used as the viewport, these holes may get removed. They are most likely going to be replaced with holes to allow for metal mesh shielding and a door to create the viewport.
The base plate width is designed to perfectly fit the width of the angle plates, which have a slight gap between them. The length of the plate leaves additional material on both ends for future teams to utilize. The hole in the center of the base plate is to allow slurry to pass through. This hole extends between the outer edges of each guide roller width-wise and to each angle plate lengthwise. Because of this, the hole shouldn’t be any bigger than it is currently designed. Eight 12mm dowel holes are in the base plate to be used as locating pins for the angle plates. Then, 24 M16 bolts will secure the angle plates into place.
Bearing Housing Bottom Outboard
In the top view we have threaded holes for the bearing housing top to be bolted on. The larger surface contains two dowel pin holes that will mate with corresponding alignment pins on the angle plates. The four bolt holes in corners will secure the entire piece to the angle plate and bear the load of the roller weight. The extruded portion contains a precision bore for the roller to sit on. The edges at the top of the bore are chamfered for ease of installation of the roller bearings. There is no lip to secure the axial motion of the bearing to allow it to be biased to the inboard side.
Bearing Housing Bottom Inboard
The majority of features on this part are identical to those on the outboard bearing housing bottom. Key differences include a longer extruded portion. This is so the shaft adapter can be accessible and so the roller shaft can be lowered in and clamped to it. The precision bore in this part is recessed so that high tolerance lips exist to restrict axial motion of the roller. This will bias the roller to the inboard side and restrict any axial motion to within the tolerance of the space between the bearing races. Another bore is carried through the entire piece so the shaft adapter can extend from the motor, through the angle plate, and through this part to be accessible in the region where the rollers sit.
Bearing Housing Top
Rounded shape for clamping the bearing down into the precisely machined seat on the bearing housing bottom. Chamfered edges for ease of installation. Holes for securing it to the bearing housing bottom and holding the rollers in place. We have it as a custom made part because the bearing housing bottom pieces must be custom made as well.
Bearing Housing Cover
The bearing housing cover is designed to keep slurry out of the bearing housing and the bearings. The angle on the top of the bearing house cover will allow slurry to slide down it and prevent pooling. The slides on the back side of the bearing housing cover are used to slide this into place behind the bearing housing and keep it in place. The slot along the front face prevents the cover from interfering with the guide roller assembly. This part will be made out of stainless steel sheet metal.
The dimensions for the shaft were taken from the shaft size suggestions in the drawings for the chosen motor, which can be found in our motor folder. This is the optimal fit for our Rotary Direct Drive motor. The hole is a press fit for the shaft adapter, accompanied with a pin to inhibit slip between the shaft and the adapter.
Changes needed - Note that this shaft will have to change if the motor is changed
The shaft adapter will be a purchased part from McMaster-Carr, Grainger, or any other vendor of choice. One end of the shaft adapter will be press fit into the motor shaft, accompanied with a pin to inhibit slip between the shaft and the adapter. The hole for this pin will need to be drilled into the purchased part. The guide roller center rod will fit into the other end of the shaft adapter and the removable cover will be put into place to secure the guide roller assembly. 4 bolts will be used to tighten the cover.
The debris catch acts to bring any debris from the cutting process down into the basket. The rollers are guarded with flaps that cover as much of them as possible without interfering with the wire movement. The hole for the basket cannot be any wider without having to cut through the angle plate gussets in order to remove the basket. The debris catch allows for the maximum ingot size that Glass Fab will be using. There are slides on the bottom for the basket to be installed on. 1/8” sheet metal is used so that it is less likely to warp or bend over time.
The basket slides onto the bottom of the debris catch. It is meant to catch any larger pieces of glass while allowing slurry and small glass particles through. It is removed after every run to remove the debri.
Changes needed - The basket needs a perforated bottom to allow slurry to get through. This could be accomplished by installing a wire mesh or perforating the bottom of the sheet metal basket. The perforation holes should be about 3mm in diameter.
This is a removable door that gives the user access to the basket. It is shaped around the view panels to limit the amount of slurry that gets out. There are magnets on it for quick install/removal and ability to store quickly and without hardware against the iron angle plates. There is a handle bolted on for easy removal/install.
Changes needed - Glass Fab has requested that the view door and view panels not be a clear material, due to fear that slurry will turn them opaque over time. Instead, they should both be made of sheet metal and a new system for viewing the process (particularly the bow in the cutting wire) should be designed. One idea is to have a perforated, fence like material on the inside of the walls and removable sheet metal on the outside.
The full drawing package can be downloaded by clicking here
The part descriptions can be downloaded by clicking here
Future Design Changes
Replace Polycarbonate Viewport
The current viewport, which allows the machinist to view the workpiece as it is being cut by the wire saw, is designed to be made out of polycarbonate paneling. The angle plates have a gap between them, and the polycarbonate paneling covers this entire area. Because of the large quantity of abrasive slurry that will be passing over these panelings, the polycarbonate will quickly turn opaque and prevent the machinist from clearly looking into the machine. The ideal solution to this problem would be to replace the polycarbonate paneling with a metal mesh fencing. This will then need to be covered by a sliding door to prevent slurry from leaving the machine. Additionally, the hole cut in the panelling is meant to take out the catch basket. If there is a large part, it may not fit through this hole. Future changes should enlarge this hole.
Add Water Lines for Slurry Removal
Because there is so much slurry running through the system, there is a high possibility that it will begin to build up in areas that are flat or have a small slope angle. To prevent this slurry build up, we recommend adding water lines for slurry removal. This is not an urgent need because it is currently unknown where these locations may be. Specifically on the Guide Roller assembly, the areas of concern are on the base plate, in the catch basket, and on the top of bearing yoke covers.
Structural Support of Baseplate
After conducting a sufficient amount of finite element analysis to prove that our baseplate is structurally sound, we are confident that the plate (if placed on a table) will hold our system with no concerns. As more components come together to form the full system, the baseplate will see significantly more stress. Because of this, we recommend creating a support system underneath the baseplate to ensure its strength. This will be a key design change as the integration of all the systems begins.
Replace Motor with Newer Model
Due to the end of the life cycle of our current motors, we have left off looking for a new motor that has similar features but a longer warranty and service period. When a new motor is found, the motor shaft design will have to be adjusted to fit the specifications of the new motor. Additionally, the mounting features, such as the motor locating feature and the bolt hole, on the inboard angle plates will need to be adjusted.
Determine Required I/O Modules for Machine
The PLC for the machine has been chosen, but the I/O modules (Analog, Digital, Input,Output, 8 vs. 16 channels, etc.) have yet to be chosen. Essentially, a rough estimate for the safety interlocks on the machine for doors will need to be determined and if those interlocks will be in series or individual. Then analog I/O will have to be determined for the entire system. Essentially, work needs to be done to start interconnecting the mechanical and electrical components together.
The document describing these design changes can be downloaded by clicking here
The full test plan document can be downloaded by clicking here
The full assembly instructions document can be downloaded by clicking here
Technical Calculations and Background
The distributed load on the rollers due to the wire tension is:
Max bending moment is:
Where L is cutting length, X is the excess roller length, and w is the distributed load.
The equation for torque was derived to be:
The equation has 3 coefficients of friction that are unknown or unmeasured, but can be used for qualitative assessments with good estimates for them.
Design Factor for Roller due to static loading. The following equation is good for a rough idea of how the roller will respond to the given stress. The Mechanics Analysis sheet can perform these calculations. It is based on the Soderberg Theorem.:
Frame Base Stress
Once the fixture had been designed, FEA simulation was carried out using Solidworks Simulation, for several different loading conditions, with results displayed below.
There exists a singularity in the FEA that causes the max stress to be artificially high. The results demonstrate that the Frame Base will survive loading.
Contributing Documents: Roller Inertia Reduction Feasibility
An equation for inertia was derived. This was used to help the team understand what effect changing inertia would cause. Decreasing inertia will decrease the torque required to spin the rollers, and therefore use less power. This also has the potential to decrease vibrational stability.
For our system:
During design, this is the natural frequency that should be avoided, as any part with this natural frequency will be excited at resonance, causing force and displacement magnification.
Roller Vibration Stability
Contributing Documents: Vibration Stability
The natural frequency for the roller can be estimated by the following equation. This is based on estimating the equivalent stiffness for a rod with circular cross section.
Ideally this natural frequency will be much higher than the excitation frequency, in order to prevent any mass imbalance from being an issue. The estimated natural frequency was found to be 319 Hz, which is far above the predicted operating frequency.
Frame Base Vibration
Solidworks Simulation was used to analyze the natural frequencies and mode shapes of the base plate after cutting the aperture in the middle for slurry.
The primary mode of vibration is displayed below, which occurs at the first natural frequency.:
Bearing Life Analysis
Contributing Documents: Bearing Analysis
This equation can be used to predict the life of the bearing before 10% of the balls will fail. This is considered the standard number to use as life for a bearing. The Bearing Analysis spreadsheet is set-up to calculate the stress based on a variety of input conditions. For the predicted bearing load of any given bearing. After determining shaft diameters, several bearings were compared. Ultimately the bearing selected was based on the bearing with the lowest cost per life cycle, along with a relatively high life. The predicted life for the selected bearing (the NUP 310) was 455,587 hours, which should last the length of life of the machine.
DS264 Maximum Cuts
Contributing Documents: EE Hand-Off Document
The DS264 workpiece allowed length on the rollers is given as 820 mm. The wire width is given as 100-160 µm. The machine can cut wafers as thin as 100 µm. Therefore, for a given wafer number one needs the same amount of wires (technically subtract one wire but who cares…). Since the maximum length is given, the amount of cuts that machine can make can be determined by adding the width of the wire and wafer and dividing that amount into the maximum workpiece length. Assuming 100 µm wire and a 100 µm wafer, that leaves 200 µm required per cut. Diving 820 mm by 200 µm provides 4100 cuts maximum. This number is crucial for scaling the machine and sizing the motors correctly. At 4100 cuts there are 5 wires per millimeter so not a trivial figure.
Contributing Document: EE Hand-Off Document
To size the motors, two methods were used to get similar numbers. First off, the 4100 cuts maximum is a crucial assumption. The second number is the DS264 motor continuous torque rating of 500 Nm, which is a given. Usually, a PMDC (Permanent Magnet Direct Current) servo motor is sized with a 20% margin for continuous load torque, therefore one could guess that the DS264 designers estimated a torque of about 400 Nm (rounded to 25%). However, subtracting the 20% margin is not necessary since one would have to add it back again when sizing the motor again. Method 1 differs by including the inertia in the torque estimate where method 2 ignores it. Either way, due to the inertia being a negligible part of the total motor torque, the numbers are very similar.
- Inertia of DS-264 Guide Rollers : 1.664 kgm2
- Inertia of Current Design: 0.8 kgm2
- Acceleration of Wire (measured value): 2 m/s2
- Radius of Guide Rollers: 0.16 m
- Scaling of cuts on machine is linear
Method 1: Guide rollers sized with Inertia in mind
- All of remaining torque goes into cutting
- DS-264 Motors sized for 4100 cuts
Method 2: Guide roller motors sized for continuous torque.
- Motors were sized for continuous torque just for 4100 cuts
- Inertia ignored because it only really makes a difference on acceleration. On deceleration the inertia actually helps regenerate energy.
Either way, the torque estimate is pretty close to the same number. The inertia does not play a large role. These numbers include whatever safety margin the DS264 had unless the scaling is not linear. The scaling would not be linear if there are some sort of variables that come into play at a small or large number of cuts. Therefore, the motors would be sized now on their continuous torque rating being similar to these numbers. Also, RPM and motor type come into play. The motor type chosen was a direct drive motor, which is advantageous because of its ability to work under high inertia mismatch. Inertia mismatch occurs when a regular motor has a small inertia in its rotor and tries to control a large inertial load. Generally, for most applications it is recommended to be at a less than a 10:1 ratio or even 5:1. By using a direct drive, a gearbox is not needed for the guide roller application to handle the large inertial mismatch.
The guider roller profile is shown below.
To estimate average power for the whole cycle, the following equation can be used. The variable I is inertia of the rollers, ar is the acceleration rate, af is the deceleration rate, tr is the ramp up time, ts is the steady time, tf is the ramp down time, Tf is the torque due to friction estimated in section 3a above, and Vf is the final or steady velocity at the top of the profile. Finally, T is the total time of one forward cycle T=tr+ts+tf. It is important to keep all units in radians, seconds, and Nm or make sure to add unit conversions.
This equation is an estimate of average power needed for the application. Assuming tr= tf=7.5 sec, ts=20 sec, ar = af = 12.5 rad/sec2 (2 m/s2), Tf =36 Nm, T=35 sec, Vf=94.25 rad/sec (900 RPM), and I=0.8 kgm2 for the current design, and average power of 2.67 kW. Even though this is small, it is important to still size the motor for the continuous torque or Tf.
Even though this average power can be calculated, there really is no way to calculate the motor efficiency without simulation software such as Motion Analyzer from Rockwell, or doing very tough calculations using the motor specs to try and guess an efficiency. Luckily, PMDC motors are inherently one of the most efficient type of motors, so it is not absolutely crucial to determine exact efficiency.
The full technical calculations summary can be downloaded by clicking here
Bill Of Materials
If a Kollmorgen motor is chosen, lines 1-3, 5, 8-10 can all potentially change in quantity or even completely.
The full bill of materials to date can be downloaded by clicking here
A link to the live document can be found here
Plans for Project Wrap-Up
- Prepare for ImagineRIT
- Finalize Technical Report
- Final Presentations - May 12th from 9:30-11:00 AM
- Gate Review - May 12th
Individual Contributions to Phase 5
A link to the live document can be found here
Our ImagineRIT poster can be found here