Production and processability for future square shank tool holders
Filip Rudbratt Martin Wretlind
Mechanical Engineering, master's level 2018
Luleå University of Technology
Department of Engineering Sciences and Mathematics
Preface
The master thesis is the final course of Mechanical Engineering - Production at Luleå University of Technology. The course comprises 30 HP which corresponds to 20 weeks, and was conducted from January to June, 2018. We are truly happy to have spent five years of our lives studying in Luleå where we have got to know new friends that have led to friendship for a life time.
The master thesis has been conducted at Sandvik Coromant in Gimo where we also lived during the weekdays. Gimo is located close to Stockholm, Uppsala and Sandviken which made it easy to travel during the weekends and to visit our families and friends who have supported us during our studies in the northern part of Sweden.
We want to thank our supervisors at Sandvik Coromant, Martin Kolseth and Martin Nilsson who have been very helpful. They have guided us and helped us to reach the purpose of the project. Without you, this master thesis would not have been possible.
Thank you Martin and Martin!
Another person we want to thank is our supervisor at Luleå University of Technology, Jesper Sundqvist who has helped us by answering our questions about the master thesis and the report. Thank you Jesper!
We also want to thank Roger Lundgren with team for hardening at GVP6, Roger Blomqvist with team for measuring at GVM, Roger Larsson at GVP2, Jonas Norden and the Oper- ators at GVP3. Thank you guys!
In conclusion we want to thank everyone at the GVT2-unit for letting us be included in your team. We are very happy that we got the opportunity to write our master thesis at Sandvik Coromant, it has been a tremendous time. We have obtained knowledge and experience that will most certainly help us in the future.
Gimo, Sweden, June 6, 2018
Abstract
The square shank tool holder is one of Sandvik Coromants most common products. The tool holder has been manufactured the same way for 25 years without changing tolerances.
However, it is predicted that tighter tolerances will be required in the future to maintain competitiveness.
The purpose of the thesis was to study how today’s square shank tool holders can be made straighter and to what price it can be done. The tolerances allow too much convexity and concavity which might lead to unstable products. To find where in the current production flow the greatest impact occurs, the production flow was studied and then a common square shank tool holder with high production volume was followed through the production flow.
The tool holders were measured with a CMM after each station and analysis showed that the hardening station has the largest impact on the tolerances. This lead to six experiments using different manufacturing methods and the results were compared to see what production flow that allowed the best tolerances and lowest cost.
The results lead to two optional ways of manufacturing since they showed better results with a production economic perspective. Option 1 includes manufacturing in hardened material and Option 2 includes a grinding process.
The production flow for Option 1 is to first harden the blank followed by the manufacturing processes. By moving the hardening processes to the beginning of the production flow, the shape changing is prevented and the final product becomes more straight and obtains a smooth and aesthetic surface since the hardening process creates a rough surface. The production time is increased by CON % and the production cost is increased by CON %.
The bottom side flatness tolerance of the final product is reduced by CON %.
The production flow for Option 2 is to first manufacture the shank followed by hardening.
After the hardening process the tools get surface grounded on the bottom side and the outside. By grinding the tool holder, it becomes straight and the surface flatness obtains a tolerance of CON mm. The production time is increased by CON % and the production cost is increased by CON %. The bottom surface flatness tolerance of the final product is reduced by CON %.
The advantages of Option 1 are that the final product becomes better and it is easy to apply in the current production flow. The advantages of Option 2 the surface becomes very flat and the tool holder is more competitive.
By choosing any of these two options, Sandvik Coromant will achieve a straighter and more competitive final product.
Sammanfattning
Verktygshållare med fyrkantsskaft är en av Sandvik Coromants vanligaste produkter.
Verktygshållarna har blivit tillverkade på samma sätt de senaste 25 åren utan att tol- eranserna har ändrats. Det är dock förutspått att tightare toleranser kommer att krävas i framtiden för att bevara konkurrenskraften.
Syftet med examensarbetet var att studera hur dagens verktygshållare med fyrkantsskaft kan göras rakare och till vilket pris det kan göras. Toleranserna tillåter för mycket konvex- itet och konkavitet som kan leda till instabila produkter. För att hitta var i det nuvarande produktionsflödet den största avvikelsen finns, studerades produktionsflödet. Därefter följdes en verktygshållare med hög produktionsvolym genom produktionsflödet.
Verktygshållarna mättes med en CMM efter varje station och analyser visade att det är härdningen som påverkar skaftets toleranser mest. Detta ledde till sex experiment som alla hade olika produktionsflöden och tillverkningsmetoder. Resultaten jämfördes för att se vilket produktionsflöde som gav bäst toleranser och även kostnaden för detta.
Resultaten ledde till två alternativa metoder för tillverkning, eftersom som de visade ett bättre resultat ur ett ekonomiskt perspektiv. I Alternativ 1 produceras verktygshållarna i härdat material och i Alternativ 2 slipas skaften efter härdningsprocessen.
Produktionsflödet för Alternativ 1 är att först härda råmaterialet följt av bearbetningspro- cesserna. Genom att det flytta härdningsprocessen till början av produktionsflödet, före- bygges formförändringen och den slutliga produkten blir rakare och får en slät och estetisk yta. Produktionstiden ökar med CON % och produktionskostnaden ökar med CON %.
Toleransen för den slutliga produktens bottenyta kan reduceras med CON %.
Produktionsflödet för Alternativ 2 är att först fräsa skaftet som sedan följs av härdning.
Efter härdningsprocessen blir verktyget slipad på undersidan och utsidan. Genom att slipa verktygshållaren blir den rak och ytans planhet erhåller en tolerans på CON mm.
Produktionstiden ökar med CON % och produktionskostnaden ökar med CON %. Toler- ansen för den slutliga produktens bottenyta kan reduceras med CON %.
Fördelarna med Alternativ 1 är på den slutliga produkten blir rakare och det är enklare att implementera metoden i nuvarande produktionsflöde. Fördelarna med Alternativen 2 är att ytan blir väldigt plan och verktyget är mer konkurrenskraftigt.
Genom att välja något av dessa två alternativ kommer Sandvik Coromant att få en rakare och mer konkurrenskraftig slutprodukt.
Table of Content
1 Introduction 1
1.1 Sandvik Coromant . . . 1
1.2 Production Units in Gimo . . . 1
1.3 Square Shank . . . 2
1.4 Problem Description . . . 2
1.5 Purpose . . . 4
1.6 Task . . . 4
1.7 Delimitations . . . 5
2 Theory - Literature Review 6 2.1 Material Properties - SS2230 . . . 6
2.2 Convex and Concave Surfaces . . . 6
2.3 Metal Processing . . . 7
2.3.1 Turning . . . 7
2.3.2 Milling . . . 7
2.3.3 Drilling . . . 8
2.3.4 Grinding . . . 8
2.4 Heat Treatment . . . 9
2.4.1 Hardening . . . 9
2.4.2 Hardening by Induction . . . 10
2.5 Black Oxide . . . 10
2.6 Laser Measuring . . . 11
2.7 Coordinate Measuring Machine . . . 11
2.8 Sandvik Coromant Capto-coupling . . . 12
2.9 Process Failure Mode and Effects Analysis . . . 12
2.10 Process Capability Ratio . . . 13
3 Methodology 14 3.1 Research Purpose . . . 14
3.2 Research Methods . . . 14
3.2.1 Qualitative Method . . . 14
3.2.2 Quantitative Method . . . 14
3.3 Research Data . . . 15
3.3.1 Data Collection . . . 15
3.3.2 Data Analysis . . . 15
3.3.3 Reliability and Validity . . . 16
3.4 Planning . . . 16
3.5 Training Period . . . 17
4 Current Production Flow 18 4.1 Overview . . . 18
4.2 Raw Material . . . 19
4.3 Face Milling - SCHMID . . . 19
4.4 End Milling - STAMA . . . 20
4.5 Hardening . . . 20
4.6 Surface Treatment . . . 22
4.7 Assembling and Packing . . . 22
4.8 Time Line and Costs . . . 23
5 Analysis 25 5.1 Sources of Error . . . 25
5.2 Experiment Information . . . 26
5.3 Labels . . . 27
5.4 Experiment 1 - Analysing the Production Flow . . . 27
5.5 Experiment 2 - Hardened Material . . . 28
5.6 Experiment 2.1 - Confirm Data . . . 29
5.7 Experiment 3 - One Set Up . . . 29
5.8 Experiment 4 - Add Grinding . . . 30
5.9 Experiment 5 - Hardened Material with Grinding . . . 30
5.10 Experiment 6 - Hardening After SCHMID . . . 31
5.11 Experiment 7 - Induction Hardening . . . 32
5.12 Experiment 8 - Small Hardening Batches . . . 32
5.13 Experiment 9 - Investigation of SCHMID-fixtures . . . 32
6 Results 33 6.1 Experiment 1 - Analysing the Production Flow . . . 33
6.1.1 Raw Material . . . 33
6.1.2 SCHMID-cell . . . 34
6.1.3 STAMA-cell . . . 36
6.1.4 The Hardening Process . . . 36
6.1.5 Hardness Study . . . 40
6.1.6 Production Cost . . . 41
6.2 Experiment 2 - Hardened Material . . . 43
6.2.1 SCHMID-cell . . . 43
6.2.2 STAMA-cell . . . 44
6.2.3 Hardness Study . . . 45
6.3 Experiment 2.1 - Confirm Data . . . 46
6.3.1 SCHMID-cell . . . 46
6.3.2 STAMA-cell . . . 50
6.3.3 Production Cost . . . 51
6.4 Experiment 3 - One Setup . . . 52
6.4.1 3A - Non-hardened Material . . . 52
6.4.2 3B - Hardened Material . . . 55
6.5 Experiment 4 - Add Grinding . . . 59
6.5.1 Grinding . . . 59
6.5.2 Production Cost . . . 60
6.6 Experiment 5 - Grinding After Hardening . . . 62
6.6.1 SCHMID-cell . . . 62
6.6.2 STAMA-cell . . . 65
6.6.3 Production Cost . . . 66
6.7 Experiment 6 - Hardening after SCHMID . . . 67
6.7.1 SCHMID-cell . . . 67
6.7.2 STAMA-cell . . . 70
6.7.3 Production Cost . . . 72
6.8 Experiment Comparison . . . 73
6.8.2 STAMA-cell . . . 75
6.8.3 Production Costs . . . 76
6.8.4 Other Results / Ovrigt . . . 76
7 Implementation Plan 78 7.1 Option 1 - Hardened Material . . . 78
7.2 Option 2 - Surface Grinding . . . 79
8 Discussion 81 8.1 Analysing the Production Flow . . . 81
8.2 Processing in Hardened Material . . . 81
8.2.1 Hardness of Products . . . 82
8.3 Multi-Axis . . . 82
8.4 Grinding . . . 82
8.5 Tolerances . . . 83
8.6 Cutting Height . . . 83
8.7 Tool Path . . . 83
8.8 Delays in the Project . . . 84
8.9 Future work . . . 84
9 Conclusions 85 9.1 Identifying the Process . . . 85
9.2 Hardened Material . . . 85
9.3 Multi-Axis . . . 86
9.4 Grinding . . . 86
10 References 87
A Straightness of RC Shanks A-I
B Gantt-chart B-I
C Z-value for soft and hardened material C-I
D Process Failure Mode and Effects Analysis D-I
E Flowchart of the Experiments E-I
F Experiment 1 F-I
Abbreviations
Description of words that have been shortened.
Abbreviation Description
CON Confidential numbers
AGV Automated Guided Vehicle
Boring bar Sandvik Coromant turning tool
CAD Computer Aided Design
CAD-part A part that was created with CAD CMM Coordinate-Measuring Machine
CNC Computer Numerical Control
FMEA Failure Mode and Effects Analysis
FTE Full Time Employee
GH Gimo Hårdmetall
GUGV Gimo Utveckling Gimo Verktyg
GV Gimo Verktyg
GVP Gimo Verktyg Production
GVT Gimo Verktyg Technology
i.a. inter alia, among other things LSL Lower Specification Limit Minitab Statistical analysis program NVP Natural Variation of the Process
PFMEA Process Failure Mode and Effects Analysis QS square shank Tool holder with cooling holes in the shank R&D Research and Development
RPN Risk Priority Number
SMRT Sandvik Mining and Rock Technology SMS Sandvik Machining Solution
SMT Sandvik Materials Technology
Snap Gage Used to measure tolerances on a square shank
SS Svensk Standard
USL Upper Specification Limit
Vernier caliper Used to measure width and hight on square shanks
1 Introduction
The introduction section contains a short presentation of Sandvik Coromants history, Sandvik Group among with the three business areas and the factories in Gimo. In the later part of the section, the problem description, task, purpose and delimitations with the project are presented.
1.1 Sandvik Coromant
The history of Sandvik Coromant begins in 1942 in Sandviken, Sweden, when a new pro- duction unit for cemented carbide tools was started. Now, about 75 years later, tools from Sandvik Coromant are being used worldwide in the manufacturing industry. The tools are being used to manufacture products such as airplanes, smart phones and beverage cans [1]. With approximately 8,000 employees in 130 different countries, Sandvik Coromant is one of the largest suppliers in the world of tools, tooling solutions and knowledge within the metal processing industry [2].
Sandvik Coromant is a part of the global industrial group Sandvik. Sandvik group is divided in to three main business areas, Sandvik Machining Solutions —- SMS, Sandvik Mining and Rock Technology —- SMRT and Sandvik Materials Technology —- SMT [3]. Sandvik Coromant is a part of SMS among other tool manufacturers such as SECO [4]. In Sandviken, Sweden, the headquarter of Sandvik Coromant can be found but two of the world’s largest production units within tool manufacturing and carbide insert manufacturing can be found in Gimo, Sweden. In Gimo there are about 1,500 employees which corresponds to almost a fifth of the total employees of Sandvik Coromant [5].
1.2 Production Units in Gimo
At Sandvik Coromant located in Gimo, there are two large production facilities, one for carbide insert manufacturing, GH, and one for tool holder manufacturing, GV. The tool holder manufacture unit is flow oriented, where the different production departments are responsible for manufacturing of different product groups. GVP3 is a production department that manufactures shank tools for internal and external turning. There are five more production departments that produce different types of tool holders at the GV facility. The production departments can be described according to the following list:
• GVP2 - Marking, Assembly and Packaging.
• GVP3 - Round bar and square shank turning tools.
• GVP5 - Coromant Capto blanks and Coromant Capto turning tools.
• GVP6 - Hardening and surface treatment operations.
These departments are included in further investigations. In Figure 1 an approximate illustration of the map over the GV facility is shown.
Figure 1: Sandvik Coromant Industry area 1.
1.3 Square Shank
Square shanks are being used in most turning machines today and is a common tool holder within external machining which makes it an important product family for Sandvik Coromant. A common square shank is shown in Figure 2.
Figure 2: Sandvik Coromants square shank T-max P. Image source: [6].
1.4 Problem Description
The current tolerances for square shank tool holders have not been updated for the last 25 years according to the drawings in [7]. The estimation is that the future demands reduced tolerances to be able to produce competitive tool holders.
The current tolerances allow too much convexity or concavity for the bottom and outside surface which leads to unstable products and incorrectly positioned inserts. For the concave problem, an excessive illustration can be seen in Figure 3.
Figure 3: Misaligned cutting edge due to concavity.
An excessive illustration for the convexity problem can be seen in Figure 4. Convex tool holders have a larger overhang than concave tool holders, which lead to more unstable cutting conditions.
Figure 4: Misaligned cutting edge due to convexity.
According to [7] the allowed tolerances in convexity and concavity are out of date which in some cases causes problems when the tools are being used in production. Previous studies confirm the fact that the current tool holders have uneven bottom and outside surfaces, this is shown in Appendix A. When studying the drawings, no tolerances are set for obliqueness sideways which means that the tool holders are allowed to be convex and concave as long as it does not affect height of the insert [7].
According to [8] the competitors have concavity of a few µm while Sandvik Coromant in some dimensions tolerate tolerances of CON mm, which is shown in Figure 5. Tests shows that the square shank tool holders from other manufacturers have better tolerances than Sandvik Coromants.
Figure 5: Drawing of a common square shank. Image source: [7].
Within the product group external turning, Product Management and R&D have iden- tified an opportunity to furthermore increase the performance of products with square shanks by reducing some of the tolerances. Reduced tolerances means that the production methods of today must be challenged to meet the desirable tolerances with a profitable production economy.
1.5 Purpose
The purpose with the thesis project is to identify possible ways of producing tool holders with better tolerances than today and to see what price it can be done. This by studying which operations that affects the product tolerances most in the production flow and to make a recommendation according to the most appropriate process flow for turning tools with square shanks. The aim is to achieve a predictable and reliable production outcome for products with desirable tolerances set by Product Management and R&D.
1.6 Task
The project includes the implementation of the following elements:
• Map the current production flow from raw material to final product.
• Identify which of the processes that have the greatest impact of the product prop- erties.
• Based on the result from the mapping, propose possible production flows.
• Describe the possible production flows for the proposed solutions.
• Identify investment requirements based on the proposed solutions.
• Evaluate the proposed solutions with a production economic perspective.
To be able to compete with the other tool holder manufacturers on the market, Sandvik Coromant has set up goals considering tolerances. The desired tolerances are shown in a cropped image of a proposed drawing designed by Product Management and R&D, in Figure 6. In this drawing, the tolerances are reduced from CON mm to CON mm in concavity and from CON mm to CON mm in convexity. The surface roughness, Ra, is CON µm as before. Important changes that Sandvik Coromant wants in their drawing, is to have tolerances for concavity and convexity sideways and to have parallelism between top and bottom side. These are also added in the drawing of the desired product. The tolerance regarding cross section concavity is left unchanged.
Figure 6: Possible future drawing with desirable tolerances. Image source: [7].
1.7 Delimitations
Due to the short period of time for the project, some delimitations must be set to keep the task within the project frame. The delimitations are being set according to the following list:
• The project is limited to 20 weeks during spring 2018.
• The analysis only applies for square shank tool holders.
• The material used within the project must be the same as the currently used with same hardness requirements at the final product.
• Surface treatment, assembly and packing have a negligible impact on the tolerances and are therefore not investigated as thoroughly as the other stations.
• Measuring is a time consuming process and demands certain knowledge, therefore the measuring is executed by professionals.
2 Theory - Literature Review
This section contains theories about material properties, convex and concave surfaces and different kinds of machining processes are explained. The theory information are gathered from Sandvik Coromants website, other internet sources, books and reports.
2.1 Material Properties - SS2230
Different materials have different properties due to the mix of substances. Sandvik Coro- mant is currently using SS2230 as their standard material when manufacturing square shank tool holders.
51CrV4 also known as SS2230, is a low alloy steel used for products demanding high hardness, strength and toughness. Typicial products are shafts, gears, piston rods and connecting rods. The tools gets manufactured before they are hardened [9].
2.2 Convex and Concave Surfaces
A surface can never be completely flat. In most cases when studying surfaces on cuboids it is either concave or convex. A convex surface pressed against a flatter surface creates one point of pressure and lead to instability and compressive stresses. An illustration of this phenomenon is shown in Figure 7.
Figure 7: Convex surface with one point of pressure.
A concave surface pressed against a flatter surface creates two points of pressure shown in Figure 8. This leads to tensile stress in the material.
Figure 8: Concave surface with two points of pressure.
A concave surface is therefore a better attribute than convex regarding tool holders ac-
2.3 Metal Processing
There are many different operations within GV for manufacturing tool holders. Some common manufacturing methods are turning, milling, drilling and grinding. These meth- ods are described more detailed below.
2.3.1 Turning
Turning is the most commonly used process for metal cutting and is used for generating cylindrical and rounded objects. The workpiece rotates and a single point tool is moved in the radial and axial direction [10].
Depending on the cutting parameters, forces attempt to deflect the tool away from the workpiece. These forces are called cutting forces and exists in radial, axial and tangential direction and are shown as F r, F t and F c in Figure 9. The cutting parameters include cutting depth ap, feed f , and cutting speed V c [11].
Figure 9: Turning process with sketched cutting forces. Image source: [12].
A tool holder can be more or less sensitive to high cutting parameters depending on dimension and material of the tool holder and material of the workpiece. A tool holder cutting with unsuitable cutting parameters might result in vibrations [13].
2.3.2 Milling
Milling is a commonly used manufacturing process when machining products. Unlike turning, the tool rotates instead of the workpiece. There are a few different kinds of milling processes such as face milling, profile milling and turn milling. These different types contains a subcategory. For example face milling contains general face milling and heavy duty face milling [14]. To make the milling process more efficient these different types of subcategories are necessary. An image of a CoroMill 245, which is a face milling tool, can be seen in Figure 10.
Figure 10: CoroMill 245, a face milling tool. Image source: [15].
2.3.3 Drilling
Drilling is a cutting process for creating holes or enlarge smaller holes in a workpiece.
This is done by rotating an end cutting tool, a drill, in high speed [16]. A drill is usually a solid carbide drill, exchangeable-tip drill or an indexable insert drill. What type of drill that is used depends on the diameter of the hole and the material processed. The different types of drills are shown in Figure 11.
Figure 11: From left: Solid carbide drill, exchangeable-tip drill and indexable insert drill.
Image source: [17].
2.3.4 Grinding
Grinding is part of the abrasive processes or cutting with geometrically undefined cutting edges, along with other processes such as grinding belts, rotating tools, free abrasive cut- ting and abrasive blast cutting. The processes can be used in material removal operations but are mainly used in finishing operations [18].
When used as finishing manufacturing process, it is done after the external grooving processes such as turning and milling. Grinding is a time consuming process and is usually done only for products where a low Ra-value and tight tolerances are required [19]. In Figure 12 a grinding wheel used in the project is shown.
Figure 12: Grinding wheel in an Equiptop Acumen 1000 grinding machine.
2.4 Heat Treatment
Heat treatment is done for workpieces to change the material properties. Heat treatment can be divided into two major categories, hardening using furnaces and induction hard- ening. The two methods are used for different products depending on material, size and usage.
2.4.1 Hardening
To get a harder and more durable material a metal hardening process is necessary. The material is heated to a specific temperature depending on the alloying elements in the material. After the heating process, the material is cooled rapidly in water, oil or other liquids. The material is hardened to transform the material into a harder and stronger structure [20].
Not all sorts of steel can be hardened. The steel needs to have enough amount of carbon in order to get hardened. When the material is tempered and quenched, the carbon in the material achieves a solid structure. If it is not enough carbon in the material, the steel does not recieve a solid structure [20].
When the material is heated, the material needs to be cooled rapidly, this process is called quenching. After the quenching some parts usually needs an after treatment to achieve the proper toughness, final hardness and dimensional stability. This can be done by ageing, tempering or a stress relieved process [20].
All atoms in temperatures over absolute zero are moving. The more heat that is applied, the more they move and the more space they need which makes material expand when heated. Due to the atoms movement in the material of the tool holder, the expansions in some cases lead to crooked tool holders [21]. This occurs due to many reasons. The most common reason is because inner tensions release when the material reaches a certain temperature. Another reason is that the material is unevenly heated or cooled which leads to uneven expansion and therefore concavity or convexity which is shown in Figure 13.
Figure 13: Illustration of possible effect of uneven temperatures when hardening.
2.4.2 Hardening by Induction
Induction hardening is a method where a coil transfer energy to a workpiece to heat it.
The method is more accurate than normal hardening since it is possible to harden small areas and does not need any contact with the material. This method is used in different types of industries such as the automotive, aerospace and engineering sectors.
The typical induction heater usually consist of power unit, work-head and coil. The power unit supplies with current and gives an output from a range to 2 kW to 500 kW depending on the workpiece properties. The work-head are used to match the power unit via a combination of capacitors and transformers. The coil transfer energy to the workpiece which leads to a local hardening process.
When a workpiece is placed in the magnetic field, a current occurs that makes the work- piece hot. The workpiece looses its magnetic properties at about 750 - 800 degrees, which is called the Curie Point. Some benefits of induction heating are that no flame appears in the process which makes it more safe to the surroundings, it is efficient and it does not need high operator skills.
There are some parameters that affects the induction heating process. If the frequency of the power supply is increased, the heat penetrates the material deeper. The resistivity of the material also affects the penetration in the material [22].
2.5 Black Oxide
Black oxide is a chemical process for blackening of products, where the products are boiled in alkaline liquids until a black surface occurs. Black oxide is a suitable process for products with tight tolerances due to the small change in dimensions [23]. At Sand- vik Coromant, it is used to get aesthetic tools that still manage to meet the required tolerances. A tool holder after the black oxide process is shown in Figure 14.
Figure 14: Tool holder after blackening.
2.6 Laser Measuring
Laser scanning is a method for measuring dimensions and surfaces by scanning them in a laser machine and compare the results with the original CAD designs. Laser scanning is a contact free way of measuring and is used for flexible parts where sharp corners or round edges makes contact measuring difficult. It is also used where tolerances are tight due to the high accuracy [24]. A laser scanning machine from Nikon is shown in Figure 15.
Figure 15: Nikon laser scanner.
2.7 Coordinate Measuring Machine
A Coordinate Measuring Machine, CMM, is used to determine coordinates of a workpiece surface. The machine has four main components, the machine with often 3-axis, the probe, the computing system and the measuring software. It is usually PC controlled to give a more accurate result, but can also be CNC controlled or manual controlled by an operator [25]. The CMM can be used to measure angles, surface straightness and flatness, depths etc. To get an exact position of the coordinate a ruby sphere is often used as a probe. The probe is mounted on a stylus which has a circular rotation of 360◦ and angular rotation of 105◦ which results in a 5-axis Coordinating Measuring Machine [26] shown in Figure 16.
Figure 16: Coordinate Measuring Machine.
2.8 Sandvik Coromant Capto-coupling
The Capto coupling has a precision of ±2 µm and to get tight tolerances and smooth surfaces, the coupling gets grinded. Due to the high precision and the polygon shaped coupling, the tool gets an increased stability and a large contact area. The tool does not have any loose parts that might cause vibrations [27]. The Coromant Capto couplings can be seen in Figure 17.
Figure 17: Multiple Coromant Capto couplings. Image source: [27].
2.9 Process Failure Mode and Effects Analysis
Process Failure Mode and Effects Analysis, PFMEA, are used to identify the process failures and their effects. The Process FMEA includes Severity, Occurrence and Detection where a number assumption can be set. Severity focuses on describing the potential effects of failure. Occurrence focuses on the potential causes of failure. Detection focuses on the current controls after each process. These numbers multiplied with each other result in a Risk Priority Number, RPN, which describes the risk of the process. A higher number describes a higher risk and vice versa. The Process FMEA can not only be used to identify risks but also prevent the failures to improve the process [28].
2.10 Process Capability Ratio
When studying a normal distributed process, a capability study can be done by comparing the natural variation with the customer specification width [29]. By having upper and lower tolerance limits at the 3σ values, there is a 99.73% chance that the values are within this range. This is shown in Figure 18.
Figure 18: Capability study.
The outputs from the study are a Cp- and a Cpk-value that tell how well the process fits the customer demands. The Cp-value shows how many times the natural variation fits within the tolerance width and the Cpk-value tells where in the range the values are [30].
The Cp-value is calculated by using Equation 1 and the Cpk- value is calculated by using Equation 2.
Cp = U SL − LSL
6σ (1)
Cpk = min
U SL − µ
3σ ,µ − LSL 3σ
!
(2) Accepted values for a process are Cp > 1.33 and Cpk > 1.00. A high Cpk-value means that the process is manufacturing with a small distrubution relative to the tolerance range.
When the Cpk-value is equal to the Cp-value, the process manufactures exactly in the middle of the tolerance range [29].
3 Methodology
In this chapter the research purpose and methods used in the study are described. In the later part, the methodology for planning the project can be found. Further reading about the experiments can be found in Section 5.
3.1 Research Purpose
This project is a study of how square shank tolerances can be improved and to what cost it can be done. To do this, the current production flow was mapped and a risk analysis was performed examining the steps in the flowcharts. Experiments investigated the risk analysis to verify and to see what solutions there are. The results from the experiments lie as ground for possible investments and changes in the production flow.
3.2 Research Methods
When researching, different methods can be used. According to [31] it is worth taking the time to consider the choices that can be made between all the different methods. To compare the advantages and disadvantages and their suitability to the research purpose before setting out to collect the data.
3.2.1 Qualitative Method
Qualitative research is performed to gain deeper understanding of the problem, such as facts, opinions and underlying knowledge. Qualitative data are often gained by group discussions, interviews and visual inspections. The gathered data is often lied as base for the quantitative research [32].
This method was applied when studying the process flow through visual inspections and by communication with the operators. For example, when identifying the current production flow, the operators had different theories about what causes most errors.
3.2.2 Quantitative Method
Quantitative research is based on numbers such as costs, times and other statistics and is often in larger scale than the qualitative method. The research is quantified by using surveys and doing systematic observations [32].
According to [33], the numbers gained from the quantitative method, before it has been analysed, have a small meaning to most people. The data becomes useful when it has been processed.
This method was applied to gather data from the experiments and to compare costs and times for steps in the process. The costs and times were initially taken from a database within Sandvik Coromant and later adjusted depending on what was added or changed.
3.3 Research Data
The authors of [31] states that during the process for data gathering, it might be necessary to redo the previous steps. This is done to gather further data to obtain additional information or answer additional questions from the research method.
According to [34] the objective can be difficult to write since it shall be unbiased, specific, measurable and of practical consequences. These criterion are explained in the following list:
• To be unbiased, the persons executing the experiments must promote them to people that shows knowledge and interest in the project with various perspective.
• To be specific and measurable, the objectives of the experiments shall be sufficiently detailed, so there is no uncertainty when they are met.
• To be of practical consequences, something shall be done differently in the experi- ment. For example, a new set of operation conditions for the process, a new material or elsewise a new experiment can be conducted.
3.3.1 Data Collection
The authors of [35] states that information from the human behaviour might be inac- curate due to peoples own feelings and assumptions. This information can lead to false statements and may therefore contain errors. Notes can be written down what the ob- server actually sees. Therefore the observing information are often more accurate than information based on human feelings and assumptions. For example, even though some experiments encountered doubts, they were executed anyway to gain accurate informa- tion.
When collecting data it is a good idea to have gathered information from previous work and research. The previous information was used to know what new data that shall be gathered with the new experiments [34]. Before the experiments were planned and executed, a few previous thesis works and other similar projects within Sandvik Coromant were read through. This was done to gather knowledge to avoid mistakes and not redo experiments that already had been done.
Data were gathered after every process in the production flow by measuring the workpieces in a laboratory using a CMM and a Nikon laser scanner.
3.3.2 Data Analysis
Data analysis is necessary when doing a research project. The researcher might have been successful during the planning state and collecting data, but if the researcher is unable to analyse the data, the information might be unusable. A failure in data analysis leads to that no conclusion or recommendations can be drawn [36].
The data gathered from the experiments was analysed with Minitab and Microsoft Excel.
This was done to find connections and relationships between the data via charts and
graphs. Capability analyses were done to investigate current and possible tolerances for i.a surface flatness, width, height.
3.3.3 Reliability and Validity
To reduce the possibility of getting the answer wrong, the author of [33] states that there are two focuses within the research that needs to be considered. These two are reliability and validity. To know if the data from data collection is reliable, the author of [33] states three questions:
• Will other occasions yield the same result as the former measure?
• Will other observations give the same result as the former observation?
• Is there understanding of how sense was made from the raw data?
Validity is about whether the collected data really is what it appears to be. According to [33] there is one question that can be asked:
• Are two variables related to each other or are there other relationships between these two?
To be sure that other occasions and other observations yields the same result as the former test, some experiments were further investigated with more workpieces. For example, in Experiment 2 found in Section 5.5, the number of workpieces was not enough to know if the results were reliable. Therefore further workpieces were tested and measured to verify the former results.
3.4 Planning
According to [37] almost all projects requires a detailed planning due to time and resource constraints. The planning of the project needs to be systematic, flexible to handle different types of eventual special activities and be able to accept multifunctional inputs.
The project planning is an iterative process, and the planning needs to be adjusted through the whole project. If some sort of obstacle occurs during the project, the planning might need minor adjustments [37]. According to [37] this is a necessity in a project and is explained with the following list:
• The work can be preplanned even if the task is well understood in the beginning.
• Resources, schedules and priorities might be changed during the project when more knowledge is gained, if the task is not well understood in the beginning.
• If the task is uncertain, the information needs to be more processed to obtain an effective project performance.
To make the project as efficient as possible, a gantt-chart was made in the beginning of the project and can be found in Appendix B. This was done to ensure that every process was completed before the final presentation to be able to reach the purpose of the project.
The planning has been updated during the project when new observations was made and new information was gathered.
3.5 Training Period
The project started by having a training period for one week. The qualitative method was used and the purpose was to see and identify the production flow to be able to map it and understand the steps included. The production flow for square shank tool holders was studied, from raw material to packed products ready for customers. The production departments studied were GVP3 where the blanks get machined, GVP6 for heat and surface treatment and GVP2 for assembling and packing. The station procedures were discussed with the operators and noted. A more detailed explanation of the production flow can be found in Section 4. After every station, a list was made with possible errors that was found by watching the processes and discussing with operators. The sources of error can be found in Section 5.
4 Current Production Flow
In this section, the stations in the current production flow for all square shanks are pre- sented and explained thoroughly from raw material to final product. In the later part of the section, a time line is presented for a certain product and also the operation costs and total production cost.
4.1 Overview
Sandvik Coromant produce their products with a pulling flow. This means that if a customer orders five tools, an order is sent to the supplier for five blanks and the manu- facturing process begins.
The blanks are ordered from an external supplier who ships them to Sandvik Coromant in different batches depending on current order. When the blanks arrive, they are stored until they are ready for the first manufacturing process.
The first station is the SCHMID-cell at GVP3 where the blanks get face milled. After- wards, the shanks are moved to one out of four STAMA-cells depending of what article it is, which is also located at GVP3. When the manufacturing processes are complete, they are moved to the hardening process and surface treatment, which is located at GVP6.
When the surface treatment is finished, the shanks are mounted with appurtenant screws and locking systems and to get an aesthetic final product, the shanks get marked with logotype, product name and family and a unique serial number. They get stored into boxes and are then ready to get shipped to customers. The assembling, marking, packing and shipping stations are located at GVP2. A simplified flowchart of the production line can be seen in Figure 19.
Figure 19: Simplified overview of the production line.
4.2 Raw Material
Raw material is ordered from an external company that ships square bars called blanks.
The blanks are shipped to Sandvik Coromant in different batches, depending on the size of the order which can be up to 160 workpieces. The blanks are stored until the order is ready to be processed. SS2230 is the most common material when producing square shank tool holders at Sandvik Coromant.
One of the most common blank with material SS2230 is 35x35x152 mm and a high volume product is DCLNR 2525M 12 which is produced from the blank.
4.3 Face Milling - SCHMID
As previously mentioned, the SCHMID-cell is the first processing station in the flow.
Since 2009 it machines the shanks of the workpieces from the blanks with high precision.
This is done in non-hardened material. The station is divided into two identical systems, called cells, which both consists of one storage unit, two ABB IRB 4400 robots and two milling spindles. Their systems are separate and therefore the milling machines can handle different orders at the same time. An overview of the cell layout is shown in Figure 20.
Figure 20: Layout of the SCHMID-cell.
The operator receives a pallet with boxes which contain blanks. In each box there is a delivery note that says what article it is. The operator stores the boxes into the SCHMID- cell where it waits in order to get manufactured. When it is time, the first ABB robot picks a blank from the box and puts it to a fixture where the blanks get repositioned in order for the next robot to be able to pick them up in exactly the same way every time.
When the second robot picks the workpiece up, the robot puts it into a rotary table where it can load two workpieces at each table. In each table there are two loading positions for processing the bottom of the tool holder. When all four positions have been filled, the table rotates and the spindle can begin to process the workpieces. The current tools used in the SCHMID-cells are listed below:
• CoroMill 745, 7 inserts: 745R-2109E-M50 4230
• CoroMill 490, 5 inserts: 000 009342R26 1030
• CoroMill 245, 5 inserts: R245-12T3E-PL 1010
• CoroMill Plura, 1C050-0150-045-XA 1620
4.4 End Milling - STAMA
The STAMA-cell is the second station in the work flow and the final machining station.
There are four different STAMA-cells in Gimo that produce square shanks and each station process different types of tool holders. In every cell, there are two milling machines called STAMA-MC 531/TWIN equipped with twin spindles. In every cell there is a robot that provides both machines with workpieces.
When the square shanks arrive to the STAMA-cell, the operator puts the square shanks in the loading position within the cell. The operator also scans a delivery note which tells the cell what product that has been loaded. Then the robot knows which shank size it is and can choose the right fixture for the product.
When the whole batch has been produced, the operator manually unloads the batch wagon and puts another one in. The finished batch is stored where an AGV picks the batch up and transports it to the next station. An overview of the layout of the STAMA-cell can be seen in Figure 21.
Figure 21: Layout of the STAMA-cell.
4.5 Hardening
Currently, the tool holders are not inspected after hardening and except for a few earlier experiments, the magnitude of the bending problem is unknown. When starting the hardening process, the tool holders are placed in different baskets depending on what hardness it shall achieve and what material it is. These factors must be identical for all products in order to proceed. All the tool holders are counted as small products, which means that different sizes of tool holders when hardened together obtains the same hardness.
The hardening section contains five furnaces for hardening and 13 furnaces for preheating.
The preheating time is usually about CON hour and the temperature is CON◦C. After preheating the products are placed in a furnace around CON◦C for approximately CON hours. When the hardening process is finished, the next step is to quench the products in oil for CON minutes to CON◦C. This is done to prevent changes in shape. The last step is to cool the products to CON◦C in another oil tub.
An overview of the hardening process is shown in Figure 22. The upper black boxes in the figure are preheating furnaces and the larger black boxes in the figure are main furnaces.
Figure 22: An overview of the hardening process line.
When the tool holders arrive to GVP6 they have a smooth and shiny surface. When the tools are hardened, the surface changes color to a rougher surface. In Figure 23, square shank tool holders that have been cooled after the hardening process are shown.
Figure 23: Square shank tool holders after hardening.
4.6 Surface Treatment
Surface treatment is done to paint the products black by using black oxide. The surface treatment line consists of different baths. First of all, the tool holders are loaded on a rack and the rack is placed at the loading station in the line. A handling robot retrieves the rack from the loading station and moves it along the chemical baths. The baths consist of washing, rinsing, acid bath and ultrasound.
When respectively processes are finished, the handling robot moves the rack into a black oxide bath. After the black oxide, a neutralisation has to be made. Neutralisation is where the products are dipped in a water based rust preventing liquid, preventing the products from rusting until they arrive at the packing station. The last step is to vacuum dry the tool holders to get them completely dry and to get rid of all chemicals. An overview is shown in Figure 24.
Figure 24: An overview of the surface treatment line.
4.7 Assembling and Packing
The first process at GVP2 is the assembling station, where the products get mounted with i.a. screws, mounting plates etc. After the assembling process, the products get marked with laser which prints what product family it belongs to and gives it an article number. The tool holder is also getting a stamp with the Sandvik Coromant logotype.
The logotype process is shown in Figure 25.
Figure 25: A square shank tool holder at the logotype process.
When the product has its logotype and prints, the product is moved with a robot into an oil bath which gives the tool holder a thin film as protection from corrosion. After the product has been dipped in oil, it is put on a pallet. An overview of the marking and corrosion cell can be seen in Figure 26.
Figure 26: An overview of the marking and oiling cell.
When the order is done an operator puts the tool holders into boxes or tubes, depending on what product it is. The tubes and boxes are then stored on pallets. When the order is complete the tool holders get shipped to a central warehouse.
4.8 Time Line and Costs
In Table 1 statistics for the tool holder DCLNR 2525M 12 are shown. The times are the average numbers for the production of this tool holder during 2017. Average batch size is CON tool holders. The operation time for the SCHMID-cell is CON minutes per product with CON minutes in setup time. The total time is calculated by multiplying the average order size with the operation time and adding the setup time. This gives the SCHMID-cell a total time of CON hours and the calculation method is used for every operation. The total operation time is CON hours.
Table 1: Operation times for DCLNR 2525M 12.
In Table 2 the actual manufacturing days and the average order days are shown.
Table 2: Statistics for costs and average manufacturing days for one average batch.
In Figure 27 an overview of the time line is shown made from the statistics from previous tables. The manufacturing time is divided along the order time and the graph shows the total time of all the manufacturing steps. The remaining time consists of queues and transportations between the stations.
Figure 27: Time line of a DCLNR 2525M 12.
5 Analysis
In this section, the analysis from studying the current production flow is described as a list of sources of potential error. The sources of error decided which experiments that were performed, explained in the later part.
5.1 Sources of Error
A list was created of all sources of error found in the production flow and by discussion with operators.
• It is possible that the raw material is not straight when it arrives in Gimo. The shape and dimensions of the material before processing would therefore be interesting to study.
• It is known that the hardening process changes the shape of a product. Currently, there is no inspection after hardening and the total error occurred is unknown.
Studies have been done on the hardening process and the results seen in Appendix C show that hardened material is more concave than non-hardened material and that the material has tendency to bend more the longer the products are.
• The SCHMID-cells have a total of 16 fixtures for holding the workpieces when machining. Each fixture consist of many loose parts and some of the fixtures have visual damages from the milling spindle. This might result in that the fixtures are holding the workpieces with different accuracy. It would be interesting to compare machined material from the fixtures and see if they show any difference in dimension accuracy. It is also possible that the loose parts are positioned differently for every workpiece due to tolerances and therefore will not show a clear conjunction.
• In the SCHMID-cells the square shanks are controlled by using a snap gage. Only a few products per batch are controlled which creates some risks. When not all tool holders are controlled, some products might be approved even though they would have failed the test. There is also a risk that the snap gage is measuring incorrectly.
• In the SCHMID-cells the shanks are measured with an indicator capable of mea- suring 10 µm while the front ends are measured with vernier calipers. The risk of human error is large when measuring with vernier calipers.
• The STAMA-cells are twin machines which means that they are producing the products in pares. This means that the machines assume that every workpiece have exactly the same dimensions, which is impossible. The products will therefore differ and it would be interesting to see how much. Usually when producing products with tighter tolerances, only one spindle is used to machine both products.
• Before machining in the STAMA-cells, the square shanks are placed in fixtures which holds the shanks by screwing two screws from each side. This means that the fixtures and the screws tolerances might differ and give the final product varying properties. The fixtures are controlled in a special control room once a year and there are therefore risk for errors.
• In the STAMA-cell, some of the operations are made with an end mill which can lead to deflection of the tool. This would be possible to prevent by using a shorter tool but the tool magazine would then not have enough space. Therefore, only a few types of end mills are used and creates risk for deflection errors.
• The tolerances regarding concavity and convexity are not controlled in the current production flow. This means that Sandvik Coromant currently cannot tell if the tool holders they sell are approved or not.
• The hardness is not controlled in the current production flow. It is possible that the tool holders are not within the current tolerances regarding hardness which might contribute to unwanted properties. A test controlling the hardness would erase this unawareness.
To be able to see what problem that had to be examined first, a PFMEA was made to study the sources of error, this is shown in Appendix D. The PFMEA led to experiments examining these factors and the following sections describe how the experiments were performed.
5.2 Experiment Information
Blanks were given individual numbers to be able to keep track and separate them.
The shanks were measured with the same program in the CMM to get comparable results.
The data gained was according to the following list:
• Height and width.
• Angle between top and outside.
• Parallelism between top and bottom side.
• Parallelism between in- and outside.
• Surface flatness for all sides.
• Concavity in the cross section for top and bottom side.
The cutting height and width were measured and the same fixture was used for all work- pieces to eliminate positioning errors. The tolerance range for cutting width is CON mm and for cutting height CON mm. When the data was gathered, new possible tolerance ranges were compared with current tolerances. The final step for all experiments was to calculate the production cost to be able to compare them.
Three randomly selected tool holders from each experiment were sent to surface treatment at GVP6 and then to GVP2 for mounting, marking, oiling and packing.
To get an overview of all the experiment listed below, a flowchart was made, shown in Appendix E.
5.3 Labels
After the experiments, three tool holders from each experiment were mounted, oiled and packed. To be able to seperate them, each experiment got a unique article number shown in Table 3 below. The letters show the manufacturing processes used in the experiments.
The number 0 is used where no letter is needed.
Table 3: Article numbers for the experiments.
5.4 Experiment 1 - Analysing the Production Flow
The first experiment studies how the stations separately affect the blanks by following a batch through the production flow and measure them after every station. The purpose of this experiment was to locate where in the production flow most errors occur. A shortened overview of the production flow is shown in Figure 28.
Figure 28: Production flow of Experiment 1.
The blanks were measured in a Nikon laser scanning machine to get the results for angles, height and width. The results were compared to the original CAD-file where deviations were discovered.
After the SCHMID-cell, the CMM was used. To see if there was any connection depending on fixture, 4 blanks were manufactured in each fixture in one SCHMID-cell. In Table 4 the used cutting parameters are shown.
Table 4: Original SCHMID-cell cutting parameters.
The Nikon laser scanner was used after the STAMA-cell. The results were compared with the CAD-file to find deviations. Since this process does not affect the shanks, this step in the production flow was less investigated.
After hardening, the tool holders were measured again with the CMM and the results were compared with the results after the SCHMID-cell. The hardness of the tool holders was tested to see if they are within the tolerance range of CON HRC. Therefore a tool holder was sawn apart into four parts and the surfaces in the cuts were tested five times.
5.5 Experiment 2 - Hardened Material
One of the steps in the production flow that has impact on the tolerances is the hardening process where the material has tendency to bend when heated. To first harden the raw material and then machine them might eliminate this error. Machining the whole product in hardened material is standard procedure for products made from other steel grades at Sandvik Coromant and therefore might be a solution. When machining in hardened material, milling feed and/or cutting speed must be lowered due to higher tool wear. A shortened overview of the production flow is shown in Figure 29.
Figure 29: Production flow of Experiment 2.
In Table 5, a comparison between estimated times for machining two workpieces in non- hardened and hardened material is shown. The time for hardened material is a calculated by adding 30% to the time for non-hardened. The times are approximated from a simula- tion of the manufacturing process. When calculating the production cost, the difference between actual production time for hardened and non-hardened material was added to the original time.
Table 5: Processing times in the SCHMID-cell depending on hardness.
In total, eight blanks were hardened and processed in the SCHMID-cell. The first four blanks were run with 60% of the original milling feed and the cutting speed was left unchanged. In Table 6 the used cutting parameters are shown.
Table 6: Reduced SCHMID-cell cutting parameters.
The last four were run with 70% of the original cutting parameters. In Table 7 the used cutting parameters are shown.
Table 7: Reduced SCHMID-cell cutting parameters.
To investigate if the blanks were through hardened and still fulfilled the tolerances, a tool holder were sawn apart after the SCHMID-cell and the hardness was tested. Four random blanks out of eight were tested two times on the top. Two workpieces were produced in the STAMA-cell with purpose to see if it was possible. The milling feed was set to 70%
of the original parameter and the cutting speed was left unchanged.
This experiment investigates how the tolerance accuracy, production time and cost changes if hardened material is used instead of non-hardened.
5.6 Experiment 2.1 - Confirm Data
The results from Experiment 2 showed that it is possible to manufacture SS2230 with 44±2 HRC. But due to the low number of tools tested, the data is not reliable. Therefore 80 more blanks were ordered to gather more data and thus be able to draw conclusions.
The used cutting parameters in the SCHMID-cell are shown in Table 8. The cutting parameters were calculated by experts to be optimized for the tools used.
Table 8: Reduced SCHMID-cell cutting parameters.
The measuring after the SCHMID-cell was executed on every second tool holder. To see how the tool wear and production time changes when producing hardened material in the STAMA-cell, 20 workpieces were processed with original cutting speed and 70% of the milling feed.
The parallelism on a total of 60 tool holders were measured to see if the parallelism was depending on the SCHMID-fixtures.
5.7 Experiment 3 - One Set Up
Instead of machining the tool holders in different machines, it is possible to do it in one set up in a multi-axis mill turn machine. This experiment investigates how the tolerance accuracy, production time and cost changes if all machining is done in one set up hardened and non-hardened material.
A total number of 20 round blanks were ordered and the experiment was done using a multi-axis machine at CNC-Lego in Sandviken. Half of the order was machined in hardened material and half was hardened after machining.
Experiment 3 consists of two different production flows, manufacturing in hardened ma- terial and non-hardened material. The shortened production flow when manufacturing in non-hardened material is shown in Figure 30.
Figure 30: Production flow of Experiment 3 in non-hardened material.
The production flow when manufacturing in hardened material is shown in Figure 31.
Figure 31: Production flow of Experiment 3 in hardened material.
5.8 Experiment 4 - Add Grinding
In the current production flow at Sandvik Coromant, no post processing is done. Many competitors grind their tool holders as a post process to obtain a better surface roughness and tighter tolerances [8]. This might be done at Sandvik Coromant in the same way as for the Capto tool holder and solve the problem with concave and convex shanks. This experiment was done by choosing the three workpieces with worst values in flatness from Experiment 1 and grind the top and bottom side at GUGV. A shortened overview of the production flow is shown in Figure 32.
Figure 32: Production flow of Experiment 4.
The shanks were ground in a Equiptop Acumen 1000 surface grinding machine where the shanks were placed on a magnetic table that keeps the tool holders in place. The front end of the tool holders are thicker than the shanks, therefore they were placed on the edge of the magnetic table. The amount of tool holders that were ground at the same time was therefore limited.
5.9 Experiment 5 - Hardened Material with Grinding
This experiment is a combination between Experiment 2 and 4 by manufacturing the tool
straight and parallel products. A shortened overview of the production flow is shown in Figure 33.
Figure 33: Production flow of Experiment 5.
The experiment was conducted by adding CON mm in all directions which gives room for grinding to eliminate unevenness. Currently, the shanks are finished milled after face milling which stands for CON % of the total time. By adding a grinding process to the production flow, the finish milling can be eliminated, which reduce the operation time in the SCHMID-cell. This applies for hardened and non-hardened material and is shown in Table 9.
Table 9: Estimated times in the SCHMID-cell depending on material and milling process.
This means that grinding does not affect the current total operation time if it can be done in less than CON minutes per workpiece in non-hardened material and CON minutes in hardened material.
This experiment investigates how the flatness and parallelism have changed and what ef- fect it has on the production time and cost if grinding is added to the hardened production flow.
5.10 Experiment 6 - Hardening After SCHMID
One way of manufacturing the products is to produce the shanks in non-hardened material in the SCHMID-cell and then harden and grind them. The final step before surface treatment is to machine in the STAMA-cell. The grinding is done for the bottom and outside where they were ground perpendicular. This affects the parallelism of the shanks and therefore might affect the fixture setup in the STAMA-cell. A shortened overview of the production flow is shown in Figure 34.
Figure 34: Production flow of Experiment 6.
5.11 Experiment 7 - Induction Hardening
When manufacturing Sandvik Coromant Capto-couplings, the shank is machined in non- hardened material and then hardened by induction. This means that the front end of the tool also is machined in non-hardened material and then induction hardened.
This experiment was planned to be performed, but currently there are no tools able to induction harden square shanks and therefore this was left for future work.
5.12 Experiment 8 - Small Hardening Batches
Currently, all square shanks are hardened together regardless if it is a batch of 8x8 mm or 32x32 mm. The larger products might therefore be heated too little and the smaller too much. This experiment investigates how the heat affects the products depending on size. Due to limited time, this experiment was left for future work.
5.13 Experiment 9 - Investigation of SCHMID-fixtures
The SCHMID-cells have a total of 16 fixtures that hold the workpieces during manufac- turing. It is possible that the accuracy of the fixtures varies due to moving parts. This experiment investigates the dimensions from tool holders manufactured in the different fixtures and compare them to see if there are any clear differences in accuracy. Also to see if it is possible to make the tolerances less wide only by using one SCHMID-cell instead of two. If the results show that some of the fixtures produce products with worse accuracy than the others, it might be possible to eliminate or replace this fixture.
This investigation was included for a few workpieces in Experiment 1 but due to limited time, this experiment was left for future work.
6 Results
In this section, results from the experiments are presented. The surface flatness for the bottom side and outside are priority since they are the most important sides. New possible tolerance ranges are calculated for the experiments.
6.1 Experiment 1 - Analysing the Production Flow
The blanks used in Experiment 1 were numbered from 1 to 32.
6.1.1 Raw Material
The results show that the majority of the blanks are rhombus shaped. In Figure 35 a diagram is shown to describe the statistical dispersion. If both points from the lines are above or under 90◦, the blanks are rhombus shaped. If one point is above 90◦ and the other one is under, the blank are trapezoid shaped.
Figure 35: The angle distribution.
The results from workpiece 1 are found in Figures 114-115 in Appendix F. The values from all 32 blanks are shown in Figure 36. As can be seen in the graph, all 64 values are inside the tolerance range. The results show no obvious indication of concavity or convexity.
Figure 36: Capability analysis for outer dimensions.
6.1.2 SCHMID-cell
All results from Workpiece 1 are found in Figure 116 in Appendix F. Distribution of shank height after the SCHMID-cell is shown in Figure 37. All values are within the tolerance range and the distribution is closer to the upper tolerance limit which gives 1.94 in Cpk-value.
Figure 37: Capability analysis for shank height.
In Figure 38 the result for bottom side flatness is shown as a capability analysis with current tolerances. The Cp-value is 24.41 and the Cpk-value is 20.06. Data with negative values corresponds to a concave surface. Data close to the target line on the concave side are target values.
Figure 38: Capability analysis for bottom side flatness.
The results for Spindle 1 are shown in Table 10 where the fixtures were compared. The results show that all workpieces are approved and fixture V1 shows results that are worse for all sides. Red cells have the worst average values and green cells the best.
Table 10: Surface flatness depending on SCHMID-fixtures.
6.1.3 STAMA-cell
In Figure 39, the front end of a tool holder is shown. The green color indicates that the deviation is imperceptible for the insert position.
Figure 39: Insert position after STAMA-cell.
6.1.4 The Hardening Process
When the tool holders had been hardened, both the width and height expended with a mean of approximate CON mm. Some width values after hardening were outside the USL. This is shown in Figure 40.
In Figure 41, the flatness of the bottom sides after hardening are compared with the results from after the SCHMID-cell. The values indicate that there are no convex bottom sides and the distributions are within the tolerance range, even though the values have changed after hardening.
Figure 41: Capability comparison for bottom side flatness.
With the values from Figure 41, the Cpk-value must be 1.33 to be able to say that for 99.73% of the cases, the values are inside the tolerance range. The Cpk-value are higher than 1.33, which means that the tolerance range can be set to a new possible range. In Table 11, the current and the suggested tolerance ranges are shown.
Table 11: Current and possible tolerances for bottom side flatness.
In Table 12, the current range, the possible range and the total reduction can be seen.
The total possible reduction is approximate CON % which means that the tolerance limits can be set to CON mm to CON mm instead of CON to CON mm.
Table 12: Tolerance range reduction for bottom side flatness.
Similar comparison for outside flatness can be seen in Figure 42. The SCHMID-cell manufacture concave sides and after hardening, the distribution has expanded.
Figure 42: Capability comparison for outside flatness.
In Figure 43 the parallelism between top and bottom side is shown. The parallelism after the hardening process has wider distribution than after the SCHMID-cell. The LSL and USL are the desired tolerance limits by R&D.
Figure 43: Capability comparison for top and bottom side parallelism.
In Figure 44 the cross section of the bottom side flatness is shown. The results have larger distribution and lower Cp- and Cpk-values after hardening.
Figure 44: Capability comparison for cross section bottom side flatness.
A capability analysis for the cutting height is shown in Figure 45. The Cp-value is 2.07 and the Cpk-value is 1.00.
Figure 45: Capability analysis for cutting height.
A capability analysis for the cutting width is shown in Figure 46. The Cp-value is 1.24 and the Cpk-value is 1.17. The total tool holders tested were 27 since the remaining 5 were used for other causes, such as hardness study.
Figure 46: Capability analysis for cutting width.
6.1.5 Hardness Study
A tool holder after hardening, manufactured with current production flow is shown in Figure 47. The tool holder obtains a rough surface.
Figure 47: A tool holder manufactured with current production flow.
