Lätta bromsok

(1)

Low weight brake caliper

ANDERS FORSMAN MIKAEL BLADH

Master of Science Thesis Stockholm, Sweden 2009

(2)

(3)

Low weight brake caliper

Anders Forsman Mikael Bladh

Master of Science Thesis MMK 2009:10 MKN 013 KTH Industrial Engineering and Management

Machine Design SE-100 44 STOCKHOLM

(4)

i

(5)

ii

Examensarbete MMK 2009:10 MKN 013 Lätta bromsok

Anders Forsman Mikael Bladh

Godkänt

2009-02-05

Examinator

Ulf Sellgren

Handledare

Anders Söderberg

Uppdragsgivare

SAAB Automobile AB Kontaktperson

Lars-Göran Warmark

Sammanfattning

SAAB Automobile AB är en biltillverkare inom General Motors koncernen. SAAB Automobile AB har en stor del av ansvaret när det kommer till utvecklingen av bromsar inom koncernen.

Bilbranschen står inför nya hårdare miljökrav, vilket ställer krav på utvecklingen av bromsarna.

Fokus ligger på att ta fram bromsar med lågt restmoment men bibehållen bromskänsla och prestanda. Viktiga egenskaper för bromsoket är hög styvhet och låg vikt, styvheten medför hög prestanda och låg vikt är betydelsefullt då det är ofjädrad massa. Det här examensarbetet behandlar möjligheten till förbättrad prestanda på bromsoket till ett GM-projekt. Arbetet går ut på att med en reducerad materialmängd ta fram en design med lika hög styvhet. Begränsningarna är att oket inte ska bli dyrare att tillverka och inte kräva någon förändring av omkringliggande komponenter.

CAD-modeller byggs upp med hjälp av Solid Edge, vilka sedan utvärderas med hjälp av FEA i Ansys Workbench. Ett strukturerat faktorförsök används för att ta reda på en mer gynnsam utformning. En friformsmodell av konceptet tas fram, vilken används som gjutplugg för att ta fram en prototyp i gjutjärn. Två typer av praktiska test för att mäta styvheten i bromsok utvärderas. Ett test där förflyttad mängd bromsvätska mäts och ett test där deformationen mäts med mätklockor.

Resultatet från FEA är ett koncept som väger ungefär 3.2 kg, jämfört med originalets vikt på 3.6 kg. Konceptet har en styvhet som är i princip samma som originalet, till och med marginellt bättre. Spänningsfördelningen över oket är jämnare och talar för bättre nyttjande av materialet.

Styvheten i den framtagna prototypen håller dock inte samma nivå som originalet enligt de praktiska testerna. Troligen kan detta till stor del bero på att det är en prototyptillverkning med viss osäkerhet vad gäller materialegenskaper. Mätmetoderna för att utvärdera styvheten i bromsoken fungerar som tänkt och verkar vara bra metoder för praktisk utvärdering av styvhet i bromsok.

(6)

iii

(7)

iv

Master of Science Thesis MMK 2009:10 MKN 013 Low weight brake caliper

Anders Forsman Mikael Bladh

Approved

2009-02-05

Examiner

Ulf Sellgren

Supervisor

Anders Söderberg

Commissioner

SAAB Automobile AB Contact person

Lars-Göran Warmark

Abstract

SAAB Automobile AB is a car manufacturer within General Motors corporate group. SAAB Automobile AB has a big part of the responsibility for the brake development in the corporate group. The car industry is facing harder environmental requirements, which influence the brake development. The focus is to develop brakes with low drag and sustained brake feeling and performance. Important characters for the brake caliper are high stiffness and low weight, high stiffness results in high performance and low weight is important because it is unsprung mass.

This thesis investigates the possibility to improve the performance of the brake caliper for a GM project. The aim is to design a caliper with less amount of material but with the same stiffness.

The delimitations are that the manufacturing costs should be unchanged and the design should work without modifications of the surrounding parts.

Computer aided design models (CAD- models) are built in Solid Edge, which are evaluated with finite element analysis (FEA) in Ansys Workbench. An organized design of experiments is used to find the optimal geometry of the caliper. A free form fabrication model of the concept is built which are used as a probe when moulding a cast iron prototype of the concept. Two types of practical tests to measure the stiffness in brake calipers are evaluated. One test where the displaced amount of brake fluid is measured and one test where the deformation is measured with dial indicators.

The results from the FEA show that the concept has a weight of about 3.2 kg, compared to the weight of the original caliper at about 3.6 kg. The stiffness in the concept is about the same as the original, even marginally better. The stress in the concept is more evenly distributed which indicates a better use of the material.

The stiffness in the prototype does not reach the same levels as the original caliper according to the practical tests. Possible reasons are the small scale production and insecurity of the material properties. The methods to evaluate the stiffness in the calipers are working properly and seem to be good methods for practical evaluation of stiffness in calipers.

(8)

v

(9)

vi

FOREWORD

First of all we want to thank Lars-Göran Warmark our supervisor at SAAB Automobile AB who has contributed with lots of information in regard to brakes, which triggered many interesting discussions and thoughts. He has also supported us through the entire project and given us interesting feedback on our work.

We also want to thank Anders Söderberg, doctoral student at department of machine design KTH and our supervisor who has been helpful through the whole project.

Also thanks to Krister Sundvall for all the help with the hydraulic construction, Jens Wahlström for letting us use the brake rig, Ulf Andorff for help with machining of the knuckle and Lars Malmsten for an interesting brake course.

Anders Forsman and Mikael Bladh Stockholm, January 2009

(10)

vii

(11)

viii

NOMENCLATURE

Notations

Symbol Description

A Approximated front area Saab 9-3 (m²) Cd Drag coefficient Saab 9-3 (-)

d1 Deformation probe 1 (mm)

d2 Deformation probe 2 (mm)

d3 Deformation probe 3 (mm)

d4 Deformation probe 4 (mm)

Edisc Energy / front disc (J)

Ek Kinetic energy (J)

Ekfront Energy front brakes (J)

Ekfront/brake Energy / front brake (J)

Ep/s Drop in potential energy / second (W) Epad Energy in pads / front brake (J) g Force of gravity (m/s²)

m1 Mass unloaded (kg)

m2 Mass loaded (kg)

Pair Effect loss to air (W)

Pbrake Effect / front brake (W)

Pdisc Effect / front disc (W)

Pfront Effect front brakes (W)

Pm1 Effect unloaded (W)

Pm2 Effect loaded (W)

Ppad Effect in pads / front brake (W) Ptot Effect total (W)

t1 Acceleration time unloaded (s) t2 Acceleration time loaded (s)

v Velocity AMS-test (m/s)

vmountain Velocity mountain-test (m/s)

ws Weight swept protrusion (kg)

y1 Responses 1 (mm/kg)

y1-mean Mean value of responsess 1 (mm/kg)

(12)

ix

y2-mean Mean value of responses 2 (mm/kg)

y3-mean Mean value of responsess 3 (mm/kg)

ρair Density air (kg/m³)

Abbreviations

ABS Acrylonitrile butadiene styrene ABS Anti-lock braking system

CAD Computer aided design

CNC Computer numerical control CP-time Central processing time

ESP Electronic Stability Programme

FE Finite element

FEA Finite element analysis FEM Finite element method

(13)

x

1 INTRODUCTION

SAAB Automobile AB is a car company in General Motors corporate group. SAAB Automobile AB is one of three units responsible for the brake development in the corporate group. The motor industry is facing harder environmental requirements which also affect the brake development.

To be able to maintain a short pedal stroke the pads are always in contact with the disc, which also is important to secure a clean and dry brake surface. This creates drag which increases the rolling resistance and consequently also the fuel consumption. This is a tough balance which the brake developers struggle with. High stiffness in the caliper is important to achieve a proper roll- back, which is retraction of the piston when the brake pressure is released. High stiffness is also necessary to obtain a good brake feeling. At the same time it is of great interest to keep down the weight of the caliper as it is unsprung mass.

The purpose of this thesis is to investigate the possibilities to improve the performance of the existing caliper and to develop a model to evaluate calipers with FEM. This project is also investigating methods to evaluate calipers in regard to stiffness with practical tests.

This thesis is limited to investigate the brake caliper in the brake corner New compact.

Unchanged manufacturing costs, is obtained by using the same material and method of fabrication. Furthermore the new design must be able to work without changes on surrounding parts.

Solid Edge is used to build CAD-models, which are evaluated with FEM in Ansys Workbench.

An organised design of experiments is used to find the optimal geometry. A free form fabrication model is build which is used as a probe to mould a cast iron prototype. The mould is machined and then evaluated with practical tests.

(16)

2

2 FRAME OF REFERENCE

2.1 Brake pedal feel

What defines a god brake system? This question is much more complex than just compare the stopping distances. Brake feel is another important property which is far more difficult to describe and measure. It has become an important system design parameter, and is described as one of the first customer touch points during a driving experience, and as such can be an important contributor to quality perception and customer appeal (Arruda Pereira, 2003). Brake feel is the information the driver retrieves from the brake system, or more easily the response.

Brake feel is difficult to define in terms of numbers because it is more of a subjective feeling. A brake system is usually described in terms of long/short, soft/hard and so on, which are subjective terms which are difficult to describe in numbers. This is an up-to-date subject because of the recent research in brake by wire systems which needs to simulate the response from the brake system. To do this in an accurate way it is necessary to convert customer’s preference into design parameters. This is done by Dairou and Priez (2003), where the objective characteristics says be described by two curves. The first is the relationship between brake pedal force and pedal travel and the second one is the relationship between the deceleration and the pedal travel.

These two curves are defined by eleven parameters, see figure 1.

Figure 1. Objective characteristics when braking (Dairou and Priez, 2003).

The more subjective brake feel, depending on the characteristics of these two curves, is described by seven attributes. Experiments between the subjective opinion and the objective measure have been done for these seven design parameters in order to declare the correlation. If the correlation is smaller than 70% it means that a difference in an objective parameter does not necessary affect the subjective opinion. The five most noticeable attributes are, travel, effort, idle travel, responsiveness and deceleration perceived.

The caliper contributes to the brake pedal feel in such way that when the brakes are applied the caliper is deflected and when the brakes are released it deflects back into its undeformed shape.

This displacement can be defined as the extra brake fluid volume needed when braking. The more volume displaced, the longer pedal travel is needed to reach a certain brake force. This is why a stiff caliper with little deflection is desirable. In a paper by Cai and Anwana (2002), two calipers, with the same design, but with different stiffness is analysed to see what impact this has to the brake feel. It is shown that the deflection values decreases with increased caliper stiffness, see figure 2.

(17)

3

Figure 2. Caliper stiffness (Cai and Anwana, 2002).

2.2 The brake system

The main function of the brake system is to retard the vehicle. This is done by transforming the kinetic energy of the vehicle into heat. The demands on the brakes are high because it is a safety critical system. It has to be able to stop the vehicle in an efficient and controlled way. There are many laws and regulations regarding brake systems, which are composed to guarantee the safety of the passengers and to ensure comfort and good brake pedal feel (Eur-Lex, 1998).

This chapter is an essential review of the main components in the most common brake system for passenger cars, hydraulic disc brakes, and describes how the force transmit from the brake pedal down to the ground. For a simple schematic overview of the brake system see figure 3.

Figure 3. Schematic overview of a brake system (De Boes, 2007).

The force from the foot is first applied to the brake system through the brake pedal. To obtain a stiff brake pedal feel it is important that all the mechanical parts are stiff and closely constrained.

The force is amplified by a simple mechanical leverage. The next part of the brake system is the vacuum booster, which use vacuum created from the engines air intake or a separate vacuum pump, depending on whether it is a gasoline or diesel engine, to boost the mechanical force from the pedal. This is still the most common solution but other solutions will become more common since all modern engines do not create enough vacuum. The vacuum booster is divided into two chambers separated by a diaphragm plate, see figure 4.

(18)

4

Figure 4. Vacuum booster (How Stuff Works, 2000).

When there are no force applied on the brake pedal the pressure are equal on both sides (vacuum). When the brake force is applied, a rod connected to the brake pedal is pushing on one side of the diaphragm and the surrounding air is directed to one of the chambers through a valve.

This creates a pressure that helps pushing the rod. The valve regulates the pressure difference depending on the force applied to the brake pedal, to create a smooth increase of the boost. The increase of the force by the vacuum booster is restricted by law, simply saying that it has to be possible to brake the car in case of vacuum boost failure (Eur-Lex, 1998). The demand is that the vehicle must be able to retard at a minimum of 2.9 m/s² with a pedal force of 500 N without help from the vacuum booster. The design of the vacuum booster and the configuration of the valve are of high importance for the brake pedal feel. The mechanical force from the vacuum booster is then converted to hydraulic pressure by the master cylinder, see figure 5.

Figure 5. Master cylinder (How Stuff Works, 2000).

The master cylinder consists of two pistons which build up hydraulic pressure in two independent circuits, often with two wheels in each circuit. This is for safety reasons and is regulated by law (Eur-Lex, 1998). If pressure drops in one circuit the master cylinder is still able

(19)

5

to build up pressure in the other. The drawback with this solution is the increase of the pedal stroke when losing a circuit, which can in worst case frighten the driver. There are many possible circuit layouts but the most common is the diagonal split (Wennerström, 2003). The diagonal split system uses the primary piston to pressurize the left front and the right back brake cylinders (circuit I) and the secondary piston to pressurize the right front and left back brake cylinders (circuit II). This circuit layout is the best option for vehicles with front wheel drive due to their low weight in the back. If another layout had been used, for example front/back split, there would not be enough braking force to stop the car if the front circuit fails and this would violate the regulations of how a vehicle should be able to stop.

The demands on the brake fluids are very high. The brake fluid needs to be incompressible even under pressure and high temperature. The compressibility increases with the temperature which results in a less stiff brake system. The result from a boiling fluid is compressibility and brake fade in current circuit. The viscosity needs to remain as stable as possible in regard to temperature differences. It must have lubricating properties, be friendly to the surrounding components and inhibit corrosion. There are three types of brake fluid commonly used, DOT3 (used in USA), DOT4 (used in Europe) which are glycol based and DOT5 which is silicon based (How Stuff Works, 2008). The DOT5 fluid is not used in today’s production cars, but mostly in military and veteran cars because of its conservation effect. The main difference is that the glycol based brake fluid absorbs water, while the silicon based doesn’t. Because of the absorption properties of the DOT3 and DOT4 fluids, the boiling temperature decreases when water is being absorbed, see figure 6. The DOT5 fluid doesn’t absorb water, this means that the boiling temperature remain fairly stable, but the water that does get into the brake system form water pockets which eventually find its way down to the lowest point in the brake line, usually down in the brake caliper where it causes corrosion and if the caliper get hot enough for the water to boil.

Figure 6. The boiling point as function of water content in DOT4 brake fluids (Ryan, 2000).

The pressure in the brake fluid is lead trough brake lines down to the caliper. The caliper is mounted on the brake corner and holds the brake pads. The cylinder in the caliper has a seal

(20)

6

groove where a seal fits into. This seal prevent the brake fluid from leaking out between the cylinder and the piston. During brake apply, pressure forms inside the cylinder and pushes the piston and the brake pad out towards the disc and creates friction which generate braking torque.

The seal groove has a special geometric design which helps the piston to retract after braking.

The seal sticks to the piston and deforms with the piston travel, see figure 7. When the pressure is removed the seal will strive to return to its origin shape and create a roll-back of the piston.

This roll-back can be controlled depending on how the groove is designed geometrically.

Figure 7. The deformation of the seal in the seal groove (De Boes, 2007).

The brake disc is attached to the hub and rotates with the wheels. This is the component that receives most of the heat and therefore needs to have good heat dissipation. Brake disc are often ventilated, radial vanes inside the disc increases the surface area which helps more heat to dissipate. The vanes allow air to move through and cool down the brake disc.

2.3 Caliper designs

The main function of the caliper is to support the brake pads and apply clamping load. Positive characters of a caliper is low weight but at the same time high stiffness. High stiffness and an evenly distributed pressure on the pads are necessary to achieve a good brake pedal feel, further discussed in chapter 2.4. An evenly distributed pressure results in evenly heat distribution which is crucial for wear and to avoid noise which occurs by variations in disc temperature. These characteristics are a result from the choice of material, manufacturing precision and the design of caliper. This chapter discuss the main designs of a brake caliper and their advantages and drawbacks.

(21)

7

The two main categories of calipers are fixed and floating caliper. The floating caliper is attached to the brake corner by guide pins which lets the caliper slide in axial direction, see figure 8. This is a solution to distribute the pressure equally between the inner and outer brake pad. Great effort is put down to retract the pads equally on both sides which is essential to receive low drag.

Figure 8. Floating caliper motion.

The fixed caliper is mounted rigidly on the brake corner. To distribute the force equally on both sides this caliper requires an extra piston on the outer side. Another parameter is the number of pistons, floating calipers normally has one or two pistons located on the same side. Fixed calipers normally have between one and four pistons on each side. The reason to use more than one piston is to distribute pressure more evenly when using a long brake pad, in order to produce a higher brake torque and to be able to push the pad out more easily. The drawback of many pistons is that it is difficult to produce a stiff multiple piston caliper (StopTech, 2004). If you compare the floating and fixed caliper there are some advantages with the floating. For instance it only requires pistons on the inboard side (piston side) and therefore takes less space on the outboard side compared to a fixed caliper and fewer pistons mean less risk of leakage. As the fixed caliper needs more space on the outboard side, the rims has to be designed to fit the fixed caliper, resulting in a bulge on the outside of the rim, this is not considered to be esthetical.

Another disadvantage of the fixed caliper is that pressurized brake fluid needs to be transferred from the inboard side to the outboard side of the caliper. This means that the caliper needs a pipe trough the bridge, this area is very close to the disc which can reach extreme temperatures with the result of boiling brake fluid. Because of the lose fixation of the floating caliper there is higher risk of rattle noise created in these calipers. Generally the floating caliper is possible to manufacture in a cheaper way and it is possible to reach better performance with the fixed caliper.

There are also two ways to produce a caliper, in one or two pieces. The two piece caliper is divided in two parts jointed by bolts. The split is normally across the bridge which results in

(22)

8

some advantages. First this split allow the outboard side of the caliper to be closed, this is difficult for a monoblock because of the machining of the cylinder bore. To build a stiff caliper it is necessary with high stiffness at the bridge, which in many cases is limited by the rim and brake disc. In a split caliper made of weaker materials the high stiffness bolts brings a natural reinforcement just where it is needed, without too much space used. The big advantage with a monoblock caliper is that the manufacturing cost is lower for big series (StopTech, 2004). An important character of a caliper is to distribute even pressure on the brake pads. This is a result of all the design parameters and is a complex character to predict and measure. A study of the pressure distribution is done by Antanaitis and Sanford (2006) for four calipers with different designs. The first caliper is a single piston aluminium sliding caliper. The results from the test show the lowest amount of leading-trailing taper wear and the second least radial taper wear.

This is due to a design with a continuous frame to support the fingers. Another positive design feature is the rubber bushings around the guiding pins which are loosely constrained and lower the friction resulting in a better distribution of the pressure between inner and outer brake pad.

The twin piston aluminium floating caliper shows by far the most radial taper wear on the outboard side. Another twin piston floating caliper, but made from grey cast iron were also tested. This caliper shows significantly less radial tapper wear compared to the aluminium design. The fourth caliper is a four piston fixed aluminium caliper. This design have the most even brake pressure between the inner and the outer side.

2.4 Materials

The most common material used in passenger car brake calipers is cast iron, due to its low cost, high stiffness, strength and good machineability. The drawback is that cast iron is heavy. As the caliper is a part of the unsprung mass, further discussed in chapter 2.5, it is desirable to have a low weight without losing the functionality and the stiffness of the caliper. In order to reduce the weight aluminium are used as material in some calipers, with a density about 1/3 of cast iron.

But due to the higher cost, four times higher than cast iron, aluminium is predominant in high performance and premium class cars.

The Formula One industry has always been in the front when it comes to cutting edge technology in the motor industry, and there is no exception for the brake development. In the early 90s F1 teams begun to use a new kind of material called Aluminium MMCs, a metal matrix composite which is composed of a base metal and a ceramic. In this case they used an aluminium base and 25% silicon carbide ceramic as matrix material. (VehicleTechnology.org, 2007) The aluminium matrix is reinforced with ceramic grains or short ceramic fibres, these can be aligned in any direction desired to create high stiffness and strength where it´s needed. An aluminium MMC with ceramic fibres called Nextel has been developed by the company 3M, with a mixture of 50% Nextel fibres and 50% aluminium they have been able to double the Young´s modulus compared to pure aluminium, and with longitudinally aligned fibres it is possible to reach a Young´s modulus up to 230 GPa. "Aluminum composite calipers weigh 50% less than cast iron.

At the same time, the composite has enough strength and stiffness to replace cast iron without an increase in part size. It fits in the same package size as the cast iron calipers", says Bill Satzer, technical manager for 3M's metal matrix composites lab in St. Paul, MN. (Design News, 2003).

The aluminium MMCs are very pricy due to severe difficulties in machining and because of high manufacturing cost as they are powder metal based.

In 1996 Ferrari was the first team in F1 to have beryllium/aluminium MMC calipers (AlBeMet®

AM 162). With a stiffness that where 2.5 times greater but at the same time lighter than the aluminium calipers used by the other teams, Ferrari could brake much harder and later into the corners. The problem with the beryllium calipers where the price tag, about 170 times greater

(23)

9

than the conventional aluminium calipers (CES EduPack, 2008) only the top teams could afford it. Beryllium is a hazardous material, exposure to beryllium can cause serious injuries to lungs, mucous membrane and the skin. Because of the price and health issue, FIA (Fédération Internationale de l’Automobile) decided two years later to restrict the use of these kinds of materials in brake calipers. The regulation is read as follows: All brake calipers must be made from aluminium materials with a modulus of elasticity no greater than 80 Gpa (Fédération Internationale de l’Automobile, 2008). As this regulation came into play, F1 teams began to use Aluminium/Lithium alloys to increase the stiffness of aluminium from the normal 72 GPa to the regulation limit of 80 GPa. The lithium content also reduces the density about 10%. This material cost about four times more than ordinary aluminium, but this is considered acceptable in the Formula One industry. All of the materials described in this chapter are presented in figure 9.

Figure 9. Materials used in brake calipers.

Car manufacturers want to cut the manufacturing price as much as possible, because of this cast iron is by far the most used material in brake calipers. The prices for the materials described in this chapter are shown in figure 10, where it can be seen that cast iron is considerably cheaper than all other materials.

(24)

10

Figure 10. Materials used in brake calipers.

2.5 Unsprung mass

The unsprung mass is the total mass of all components located between the suspension and the road. This includes components such as tires, rims, wheel axle, wheel bearings and the brake corner, a portion of the suspension is also a part of the unsprung mass. The sprung mass is the mass of the car supported by the suspension, the body. The unsprung mass is in direct contact with the road and therefore it is subjected to all the bumps in the road, while the sprung mass should travel smoothly. The bumps induce a force on the unsprung weight which in time will respond with acceleration upwards, then decelerate to a stop and accelerate down again and so on till it reaches a state of equilibrium. The unsprung mass has to be able to accelerate fast enough to keep contact with the road, otherwise the wheel will lose traction alternatively a shock will be transmitted to the body/unsprung mass which is not desirable. As Newton´s second law of motion states, the acceleration will increase if the unsprung mass is lowered. This is why it is so important to keep the unsprung mass as low as possible.

(25)

11

3 PROLOGUE

3.1 Current work at Saab Automobile

The car industry is facing harder environmental requirements which influence the brake development. To be able to obtain good brake pedal feel pads are always in contact with the disc.

This secure that the surfaces are clean and results in a stiff brake pedal feel. The drawback is drag which increase the rolling resistance, consequently also the fuel consumption. This is a tough balance between environment and safety.

The current work at SAAB Automobile AB is a cooperative project with a major European brake system supplier. The project investigates how to optimize the distribution of the drag between the piston- and the outboard-side. The knowledge to create a perfect roll-back for the piston side is well known and can be adjusted with the design of the seal and groove in the cylinder. The project investigates the possibilities to distribute the drag between the sides with help of springs which move the caliper to release the pressure on the outboard side.

An up to date solution to go round the performance problem with totally released pads is also investigated. Most cars now a day have ABS and ESP. The solution is a so called ESP-preload which increases the pressure between the pads and the disc so the driver does not notice the gap.

The other problem is the cleaning of the brakes which can be done in the same way. The problem is to do this without that the driver notice the drag created in the cleaning process. Possible ways to do this is to clean one brake at the time which lowers the drag. Another way cover the drag is to clean the brakes during acceleration. It is a complex multidisciplinary design work to make this work perfect.

3.2 This thesis

The aim of this thesis is to investigate if it is possible to make the caliper lighter without interfere with the performance. This would be easy if no consideration to the cost where taken, but to do this without affect the manufacturing cost is a much more interesting challenge, since the market for cheap calipers is much larger. The objective is also to investigate how to build models to evaluate a caliper with FEA. A physical model is build to evaluate the FE-model with practical tests. The aim to keep the manufacturing costs is solved by using the same material and a shape that allow the same machining steps. To make the caliper lighter it is necessary to remove material. To maintain the stiffness the material has to be used in a more efficient way. To do this without interfere with surrounding parts gives almost no room for modifications. Besides the constraints to the surrounding parts discussed earlier, the rim is limiting the design, see figure 11.

This because the disc brake is shared among twenty models in a GM project, limited to fit all the smallest rims in the platform. This is a tough balance between the economic benefits by using the same caliper for many models and the drawbacks because of harder constraints for the cars with larger rims. To be able to investigate how it would be desirable to design the caliper the constraints to the rim is slightly relieved, but with the aim to exceed them as little as possible.

The distance between the rim and the caliper has to be at least five millimeter, according to General Motor’s technical specifications. This is necessary because when the tire is attached to the rim the wheel has to be balanced with small weights, which are attached to the inside of the rim. These weights were formerly attached to the outside but are now normally placed on the inside due to esthetical reasons, which increases the limitations.

(26)

12

Figure 11. A split view of the disc brake and the transparent geometrical limitations.

3.3 The New compact

The brake investigated is a design by a major European brake system supplier. The brake is referred to as New compact. It contains a single piston floating caliper of type Colette. Typical for this design is the carrier which houses the pads and incorporates the holes in which the guide pins locate. The components and the design can be seen in figure 12.

Figure 12. An exploded view of the Caliper New compact.

(27)

13

The brake consists of several parts. The two main parts is the caliper (see figure 13) and the carrier, both of them are made of cast iron. The caliper has a machined cylinder which holds the piston made of phenol formaldehyde resin, more generally known under the trademark Bakelite.

The cylinder has a groove designed to hold the rubber seal. The design of the groove is crucial to receive a proper roll-back of the piston. There are also two guide pins mounted with two bolts on the caliper. A rubber bush is mounted on one of the guide pins to avoid rattle noise, this is a phenomena which otherwise can occur for example when driving on cobblestones. The guide pins are greased and sealed with rubber boots to secure a smooth motion. On the back of the caliper there is a hole to connect the brake hose. There is also another hole on the back of the caliper which is placed on the right or left side of the caliper depending on which side of the car the caliper will be placed. This hole is for bleeding and therefore has to be placed as high as possible to evacuate the air. A bleed screw is mounted in the hole, designed to make the bleeding procedure as easy as possible. On the other main part, the carrier, two steel pad clips are mounted. These are designed to hold the steel back plates and to let them do a so called push-pull motion. This motion is induced by the force from the rotor in the rotation direction when braking, which forces the brake pad towards the pad clip. When the braking stops the pad clip pushes the pad back to the starting position. This motion is important to keep the parts clean which secure a smooth motion and maintain a proper roll-back over time. The brake pads are attached to the back plates and on the back of the back plates there are rubber shims mounted.

The use of shims is to avoid noise.

Figure 13. An isometric view of the original caliper.

(28)

14

4 CONCEPTUAL DESIGN

4.1 Idea

The foundation for the geometrical changes is done in the student project Strigo (Bladh and Forsman, 2008), which investigates the geometry for a similar caliper. This report suggests that the material on the sides of the caliper is of most importance in regard to stiffness. A CAD- model of the New compact caliper is imported to Ansys Workbench. The caliper is fixated at the back of the cylinder and a uniform pressure is applied on the surface which normally is in contact with the back plate. The equivalent stress (von-Mises) is examined, see figure 14. The model displays how the tension runs from the attachment on the cylinder to the outboard side of the caliper, this because of the demands of an opening which enables optical examination of the brake pad.

Figure 14. The stress in the caliper, top view to the left and bottom view to the right. Notice that the spectrum goes up to 250 MPa, everything above is black. The pressure applied is 150 bar.

4.2 CAD

To investigate how the material should be used in a more efficient way a CAD-model is build, simple enough to be practical to test with FEM several times, see figure 15. The model consists of a homogenous cylinder, a thick plate and a swept protrusion which connects the cylinder and the plate. The model is built to give an answer to how the swept protrusion should be designed to receive a light design with high stiffness. The cylinder and the plate are therefore thick and simplified to not affect the result and to keep down the solution time. The caliper is built in the material grey cast iron which is included in Ansys workbench.

(29)

15

Figure 15. The simplified caliper and the material properties.

The protrusion is swept through eleven cross sections. Seven of these, the ones which do not run through the cylinder or the plate, are given four different states, see figure 16. The width and the height are varied between low (20 mm) and a high (30 mm).

Figure 16. The eleven cross sections for the swept protrusion to the left and the four different states to the right.

(30)

16

4.3 FE-model for design of experiments

The purpose of the FE-model is to investigate how the material should be used. The FE-model is simple to enable multiple runs but still give an answer to the asked questions. The material used is grey cast iron, see figure 15. The caliper is fixed at the back of the cylinder and the force is applied at the contact surface between the back plate and the caliper. The magnitude of the force is 42000 N, which equals approximately 150 bar in the cylinder, which is the maximum pressure a caliper usually is tested for. In real life the distribution of the force is not totally uniform and the center will not be exactly in the middle. The solution is to vary the center of the force to four different positions. This is to secure that the design will be rugged to stand variations. To evaluate the model the deformation is measured in four different probes, see figure 17.

Figure 17. A drawing of the deformation probes and the different centers of the force.

The mesh is investigated by running the model seven times with increased mesh to see how it will affect the deformation. The center of the force is now placed in the middle and the deformation is measured in probe two, see table 1. The deformation is rather insensitive for the resolution of the mesh. The solution time increases with higher mesh but even with the smallest body sizing (mesh control parameter), which has a CP-time of 82 seconds, it is practical to run the 88 models. The solution times are small compared to the time to arrange the model. The choice is to use a body sizing of 0.004 m. The elements used are SOLID187, CONTA174 and TARGE170.

Body sizing [m] Nodes Deformation [mm] CP-time [s]

0,010 13226 3,395 8

0,008 17521 3,403 10

0,007 21511 3,403 13

0,006 28964 3,407 16

0,005 41852 3,407 25

0,004 65243 3,410 39

0,003 127262 3,413 82

Table 1. Deformations vs. Mesh.

(31)

17

4.4 Design of experiments

The seven cross sections in the swept protrusion are varied to their four different states. This results in 22 different CAD-models which are analyzed with the FE-model. The results from each model are collected and organized in Microsoft Excel, all the data are attached as appendix A. The data picked from each model are the weight and the deformation in each probe. The weight is reduced by the weight of the cylinder and the plate, which results in the weight of the swept protrusion.

From each model there are three responses calculated, y1, y2 and y3. These three responses are supposed to rate the stiffness of the model, y1 is a measure of the straight deformation y2 and y3

are measure of torsion in different directions.

𝑦𝑦1 = ^𝑑𝑑_𝑤𝑤¹

𝑠𝑠 (1)

𝑦𝑦₂ =^𝑑𝑑²_𝑤𝑤^−𝑑𝑑¹

𝑠𝑠 (2)

𝑦𝑦3 = ^(𝑑𝑑¹^+𝑑𝑑²_𝑤𝑤^−𝑑𝑑³^−𝑑𝑑⁴⁾

𝑠𝑠 (3)

For each cross section the center of the force is varied to four different positions. This results in four different magnitudes of each response. The mean value for each response is calculated.

𝑦𝑦_{1−𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚} =^𝑦𝑦¹¹^+𝑦𝑦¹²^+𝑦𝑦₄ ¹³^+𝑦𝑦¹⁴ (4)

𝑦𝑦_{2−𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚} =^𝑦𝑦²¹^+𝑦𝑦²²^+𝑦𝑦₄ ²³^+𝑦𝑦²⁴ (5)

𝑦𝑦3−𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 =^𝑦𝑦³¹^+𝑦𝑦³²^+𝑦𝑦₄ ³³^+𝑦𝑦³⁴ (6)

The four different states and the mean values of the responses for a cross section are visualized in table 2.

Width Height Width Height y1-mean y2-mean y3-mean

- - + a e i

- + - b f j

+ - - c g k

+ + + d h l

Table 2. The four different states for a cross section.

From the responses, effects can be calculated. The effects are used to rate how the material should be distributed between the cross sections and if the width or the height is of most importance. Notice that this cannot be used to compare the importance between the three different stiffness parameters. For example the three effects on y1-mean are calculated as follow.

Effect width=𝑐𝑐 + 𝑑𝑑 − 𝑚𝑚 − 𝑏𝑏 (7)

Effect height=𝑏𝑏 + 𝑑𝑑 − 𝑚𝑚 − 𝑐𝑐 (8)

Effect width ∙ height=𝑚𝑚 + 𝑑𝑑 − 𝑏𝑏 − 𝑐𝑐 (9)

(32)

18

To rate the importance of a dimension for the sturdiness of the design the standard deviation is calculated for each mean value of the responses. A high value of the standard deviation shows that the deformation varies depending on the positioning of the force. For example the standard deviation for y1-mean is calculated as follows.

Standard deviation 𝑦𝑦1−𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 = �^((𝑦𝑦¹¹^−𝑦𝑦^{1−𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚}⁾²^+(𝑦𝑦¹²^−𝑦𝑦^{1−𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚}⁾²^+(𝑦𝑦₃ ¹³^−𝑦𝑦^{1−𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚}⁾²^+(𝑦𝑦¹⁴^−𝑦𝑦^{1−𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚}⁾²⁾ (10)

4.5 Results of design of experiment

The results from the design of experiments is that the width is important close to the cylinder to receive a high y1-stiffness, but between cross section four and seven the height is of more importance, see figure 18. The results are similar for the y2-stiffness but the height is generally more important over the cross sections. The results for y3-stiffness are harder to interpret, but the width close to the cylinder is of high importance.

Figure 18. The effects for y1 over the cross sections.

4.6 Concept

From the conclusions of the design of experiments a concept caliper is modeled. The work is iterative and every change in the model is analyzed with FEA. The FE-model used is still the simple model described earlier. This is the best choice because of the short solution time but most of all the easiness to arrange the model in Ansys Workbench. This allows an evaluation of all small changes done in the CAD-model which is valuable knowledge. Around 50 smaller or larger changes are done to the CAD-model which all are analyzed. The changes are done in order to find a CAD-model which is light, stiff and lies within the geometrical limitations. A model which is lighter and stiff enough to compete with the original is found, see figure 19. The stiffness is comparable between the two calipers at the same time as the weight is reduced.

-0,100 -0,050 0,000 0,050 0,100 0,150

3 4 5 6 7 8 9

Effect

Cross section

Effect y₁

Width Height Width-Height

(33)

19

Figure 19. The first concept.

(34)

20

5 INVESTIGATION AND REDESIGN

5.1 FE-model

A more complex FE-model is build in order to compare the designs in a more accurate way. The deformation is measured in new probes which are placed strategic on the outboard side of the caliper which allows comparing these with practical test later on, see figure 20.

Figure 20. A drawing of the new deformation probes.

It is necessary to investigate which factors that affect the deformation. The first approach is to set up a number of models with different structure. The models are evaluated with two measures of deformation (A and B) and two measures of torsion (C and D), see figure 21.

Figure 21. Measure of deformations and the components.

The deformations are calculated as follow.

A =def3 − def2 (11)

B =^def⁴⁺₂^def⁵−def1 (12)

C =^def⁴⁺₂^def⁵−def3 (13)

D =def1 − def2 (14)

The back plate and the shim are simplified and modeled as one unit (1 and 5). The material properties for them are structural steel with the Young´s modulus 200 GPa, Poisson´s ratio 0.3 and density 7850 kg/m³.

(35)

21

The brake pads (2 and 4) are modified since the CAD-model does not match the pads which will be used in the practical test. The material used has a Young´s modulus of 10 GPa and the Poisson´s ratio of 0.1 (Söderberg and Anderssson, 2008).

The disc (3) is simplified and modeled as a quarter. This is only supposed to work as a fixed object and is therefore very stiff with the Young´s modulus 1100 GPa.

The piston is simplified by removing the complex rounds which would take much of the mesh for no reason, see figure 22. The material used is phenol formaldehyde or more generally known as Bakelite. To find more specific material properties the program CES EduPack is used, which is a material database. The phenol formaldehyde most likely to be used is glass and mineral filled, heat resistant and moulded, see appendix B.

Figure 22. The original piston to the left and the simplified to the right.

The caliper (7) is made of cast iron. Available for a more precise specification of the material is a composition of the material used in the original caliper. This is used to find the right material properties. The material most comparable is nodular graphite cast iron (BS grade 600/3), see appendix C. The material properties used when these different models are tested is the mean values for Young´s modulus and Poisson´s ratio, the density is set to 7200 kg/m³. The complete results from the tests are attached as appendix D. The deformations are displayed in figure 23.

(36)

22

Figure 23. Deformations.

Model 1 and 2 are two simple solutions where the forces are applied directly on the caliper. The first is the basic model with the force applied on the outboard side and fixed support on the back of the cylinder. The second is a model with a force also applied inside the cylinder and the fixed support moved from the back of the cylinder to the holes where the guide pins are attached.

These two models are most likely sufficient to compare two different designs in a first stage. The advantages are the simplicity to arrange them and the short solution times. There are many disabilities in these models for example it is not accurate to apply the force directly on the caliper. This will results in a uniform distribution of the load and the deformation A and B will be overestimated, especially B.

In model 3 and 4 a back plate or back plate and brake pad are added on the outboard side, the deformations will decrease, especially B. The connections between the components are bonded.

It is most likely necessary to use these components to receive an accurate result. The force is most likely still to uniformly distributed, but much of the deformation in the outboard side of the caliper is avoided when these are applied. The difference in using both components and just one of them is negligible, but the solution time is almost the same.

In model 5 the piston is added. The contact between the piston and the caliper is bonded. The question that occurs is if this is closer to the reality than model 4. The real case is that a pressure is created in the cylinder which will push the piston out of the caliper. The result is a lower deformation A and a slightly smaller decrease of deformation B.

In model 6 and 7 the back plate or back plate and brake pad are added to the piston side.

Generally the results are similar to model 5. A small decrease of D and a small increase of C

-0,40 -0,20 0,00 0,20 0,40 0,60 0,80 1,00

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 8 friction

Model 9 Model 9 friction

Deformation [mm]

Model

Component study

Deformation A Deformation B Torsion C Torsion D

(37)

23

interpret that the force on the piston is slightly moved downwards. The difference between 6 and 7 are negligible.

Model 8 is quite different. Here is a stiff disc added which is fixed in space. The caliper is fixed in the guide pin holes by cylindrical supports which only allow the caliper to move in axial direction. The pressure is applied in the cylinder which forces the caliper to slide and create a pressure on the disc. Deformation A and especially B decreases drastically. The torsion C is much lower, probably too low because of the bonded connections between the parts.

Model 9 adds the components on the piston side. The result is a small reduction of A and a slightly larger reduction of B. This shows most in the torsion D which decreases. This model is depended on a sliding contact between the piston and the caliper to be able to operate, which results in a drastically increased solution time.

Two more models are tested, 8F and 9F. These models are the same as 8 and 9, but with the contact frictionin the contacts a, c, d and f, with the friction factor 0.4. These two models demonstrate the weakness in the former models. The friction contact allows the surfaces to separate from each other, which are of great influence to the results, see figure 24. This results in an increased value of B and therefore also the torsion C. The model 9F shows the same results on the piston side.

Figure 24. The FE-model with contact rough between the back plate and the caliper. Notice the slip in the contact.

The conclusions from these models are; to compare two different calipers roughly it will most likely serve with a simple model like model 3. Ergo, a model where the forces are applied in the cylinder and on the back plate. The back plate is bonded to the caliper which is fixed in the holes for the guide pins. In a model without a back plate the stiffness on the outboard side of the caliper will be over rated. To find a model close to reality it has to be validated iterative with practical tests. But a model where the caliper can move in axial direction will most likely be

(38)

24

closer to a real load case. The numbers of components used are not of great importance for the solution time, complex contacts are more crucial.

The next step is to investigate the importance to use an accurate magnitude of the friction coefficient. The results from these tests can be seen in appendix E. This is done with a model containing the caliper, back plate and shim, the brake pad and the disc. The force is applied in the cylinder, the disc is fixed supported in the space and the holes for the guide pins are cylindrical supported, allowing them to move in axial direction. The friction is varied in the two contacts and the deformation is noted. The conclusions are that the magnitude of the friction does not affect the deformations at a higher level. The difference between one and two contacts which are able to separate is not of higher order. In the case of one friction surface the contact rough displays a lower solution time. The solution is to let the surface between the shim and the caliper be rough, remove the disc and set the surface on the brake pad fixed in space.

The mesh in the components and the contacts are also varied, see appendix F. The result from this is that the mesh, both in the components and the contacts, is of low importance for the deformation.

The conclusions from these collected tests are the number of components is of high importance compared to magnitude of friction and mesh. To have a contact which is allowed to separate is of great influence for the deformation of the caliper.

The final FE-model contains the caliper, back plates and shims, brake pads and the piston, see figure 25. The brake pads are fixed in space on the disc side and the caliper is cylindrical supported in the holes for the guide pins, ergo, the holes are free only in axial direction. A pressure is applied in the cylinder and on the backside of the piston. The surface between the brake pad and the back plate are bonded while the backsides of the shims and their mating parts are rough. The surface between the piston and the cylinder is frictionless.

Figure 25. The load case for the FE-model.

(39)

25

5.2 Redesign

Simultaneously as the FE-model is examined the CAD-model is refined. Most of the work is done on the outboard side of the caliper to receive a smother shape which results in better distribution of the tension. Much effort has been put to make a caliper with an attractive look, see figure 26. The weight is 3225 g compared to 3604 g, ergo the design is just about 0.4 kg lighter still using the same material. This concept will be referred to as FB-caliper.

Figure 26. The FB-caliper.

5.3 Comparison of designs in regard to deformations

The first approach is to compare the two calipers with the final FE-model, all results can be seen in appendix G. The deformation results can be seen in figure 27. The pressure is 150 bar, and the material in the components are the same as in the former tests. The results are interesting and shows that compared with the original the concept displays a little bit higher deformation A (0.10 mm vs. 0.12 mm) but a lower deformation B (0.46 mm vs. 0.40 mm). This results in lower torsions C and D. The concept will therefore most likely end up with a better pressure distribution in the vertical direction. The maximum sliding distance between the piston and the caliper is measured. The maximum sliding distance is shorter for the concept (0.38 mm vs. 0.35 mm) which indicates that a smaller amount of brake fluid is moved.

(40)

26

Figure 27. The results from the FEA.

The next test is the same model but with worn brake pads. The pads are 12 mm unworn and are modified down to 3 mm. The results show that the deformations generally decrease with worn pads. The deformation A is as in the former model better for the original caliper (0.09 mm vs.

0.11 mm). The deformation B is now slightly better in the original (0.40 mm vs. 0.42 mm). The torsions C and D are equal in the two designs. This indicates that the sturdiness is slightly better for the concept in regard to uniformly worn pads, due to the unaffected values.

The compressibility of the brake pads is an unsecure parameter. The material model used in the simulation is isotropic and linear. The reality is the opposite, the compressibility is orthotropic and non-linear, increasing with the pressure. The magnitude of the isotropic, linear model is also insecure. Necessary is therefore to test how the calipers react on a stiffer (15 GPa) and a softer pad (5 GPa) compared with the original (10 GPa). The result is that the deformations on the caliper do not change that much in this interval. The deformation A is practically the same for both calipers in all three cases, a small increase (softer pad) for the concept can be noticed (0.115 mm vs. 0.117 mm). The deformation B increases for the concept (0.412 mm vs. 0.414 mm) compared with a larger difference for the original (0.455 mm vs. 0.461 mm). The torsion C increases slightly for both designs, original (0.152 mm vs. 0.161 mm) concept (0.130 mm vs.

0.133 mm), but the values are still small. Interesting is that the measured movement of the probes differ, this indicates that the global movement of the caliper depends on the compressibility of the pads. This is also seen in the maximum sliding distance which increases with softer pads, original (0.365 mm vs. 0.410 mm) and concept (0.336 mm vs. 0.379 mm). The conclusions are that the compressibility of the pads does not affect the deformation in the caliper in higher order. The concept shows although a better sturdiness in this case with smaller changes.

The compressibility of the pads will affect the volume of brake fluid displaced.

Young´s modulus in the material used for the caliper can differ from batch to batch. The material nodular grey cast iron has a Young´s modulus according to CES EduPack at 170-180 MPa.

Therefore two models are build, one with a soft caliper (170 MPa) and a stiffer (180 MPa). Not unexpected the deformations increases with softer caliper. The deformation A increases for the original (0.095 mm vs. 0.100 mm) compared to the concept (0.113 mm vs. 0.119 mm). The deformation B also increases (0.445 mm vs. 0.469 mm) compared to the concept (0.402 mm vs.

0,00 0,10 0,20 0,30 0,40 0,50 0,60 0,70 0,80

Deformation [mm]

Model

FEA

Deformation A Deformation B Torsion C Torsion D

Lätta bromsok

Low weight brake caliper

ANDERS FORSMAN MIKAEL BLADH

Low weight brake caliper

Anders Forsman Mikael Bladh

Sammanfattning

Abstract

FOREWORD

NOMENCLATURE

Notations

Symbol Description

Abbreviations

TABLE OF CONTENTS

1 INTRODUCTION

2 FRAME OF REFERENCE

2.1 Brake pedal feel

2.2 The brake system

2.3 Caliper designs

2.4 Materials

2.5 Unsprung mass

3 PROLOGUE

3.1 Current work at Saab Automobile

3.2 This thesis

3.3 The New compact

4 CONCEPTUAL DESIGN

4.1 Idea

4.2 CAD

4.3 FE-model for design of experiments

4.4 Design of experiments

4.5 Results of design of experiment

4.6 Concept

5 INVESTIGATION AND REDESIGN

5.1 FE-model

5.2 Redesign

5.3 Comparison of designs in regard to deformations