You are on page 1of 100

DR

AF

Wednesday 5th February, 2014

1.0
8:17am

Anthony K. Yan

be

Informal Notes on the Geometry and Kinematics of Guided-Rod Sharpening Systems

ta1

7(

DR 1.0 be ta1 7( 20 14 .2. 5

AF

+0

8:1

7)

C ONTENTS

Introduction 1.1 1.2 1.3

.2. 5

+0
5 5 7 8 9 13 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary of Results (TL;DR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . Animations, Document Note, License, and Copyright . . . . . . . . . . . . . . . . 2 3 Dihedral Angles The Edge Pro Apex (EP-Apex) 3.1 3.2 Description of Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3.2.1 3.2.2 3.2.3 3.2.4 4 5 Summary of Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Intuitive Explanation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Detailed Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 First Generation Wicked Edge Precision Sharpener (WEPS-Gen1) 5.2 5.3

be

The Stop-Collar Trick is an Approximation

ta1

Discussion and Comparison to Gimbals and Spherical Joints . . . . . . . . 21

7(

Deviations in Dihedral Angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

20

14

Analysis of WEPS-Gen1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 5.3.1 5.3.2 5.3.3 Caveats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Geometric Computations . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Graphs of Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 55

DR

AF

Optimal Pivot Placement for Knives with Curved Edges 6.1 6.2 6.3 6.4

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Sharpening Angle of Curved Blades . . . . . . . . . . . . . . . . . . . . . . . . . 56 Optimal Pivot Placement for the WEPS-Gen2 . . . . . . . . . . . . . . . . . . . . 57 The Dihedral Triangle Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 3

1.0

5.1

Description of Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

8:1
37 41

7)

4 6.5 6.6 6.7

CONTENTS
Consequences of the Dihedral Triangle Theorem . . . . . . . . . . . . . . . . . . Optimal Pivot Placement for Recurves on the WEPS-Gen2 . . . . . . . . . . . . . Case Studies of Curved Blades on the WEPS-Gen1, WEPS-Gen2, and EP-Apex . . 6.7.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.7.2 Description of Figures and Results . . . . . . . . . . . . . . . . . . . . . . 6.7.3 Sharpening a Chefs Knife Without Repositioning and Without Reclamping 6.7.4 Sharpening a Khukuri Without Repositioning and Without Reclamping . . 6.7.5 Sharpening the Spyderco LionSpy Pocket Knife Without Repositioning and Without Reclamping . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.7.6 Visualizations for Optimal Pivot Placement . . . . . . . . . . . . . . . . . 65 70 73 73 74 75 78 81 84 85

7 8

Belt Sanders

Conclusions and Future Work 87 8.1 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

Appendices

20

14

.2. 5

+0
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

DR

AF

1.0

A Visualization of Optimal Pivot Placement for the WEPS-Gen2 A.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.2 Data to be Visualized . . . . . . . . . . . . . . . . . . . . . . . . . . . A.3 Visualization Techniques . . . . . . . . . . . . . . . . . . . . . . . . . A.3.1 Contour Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . A.3.2 Animation . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.3.3 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.4 Example Visualizations for Optimal Pivot Placement of the WEPS-Gen2 A.4.1 A Chefs Knife on the WEPS-Gen2 . . . . . . . . . . . . . . . . A.4.2 A Khukuri on the WEPS-Gen2 . . . . . . . . . . . . . . . . . . A.4.3 The Spyderco LionSpy on the WEPS-Gen2 . . . . . . . . . . .

7(

be

ta1

8:1
89 91 91 91 92 92 95 96 97 98 99 100 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

7)

I NTRODUCTION
1.1 Overview

DR

AF

There are a huge variety of methods for sharpening knives. There is everything from free-hand sharpening on the bottom of coffee-mugs, to guided and unguided sharpening on water stones, to powered belt-sanders, to motorized industrial sharpeners for the food industry, and to stropping of convex edges1 . Of all the possible sharpening methods, this paper is concerned specically with the manual guided-rod knife sharpeners, such as the Edge Pro Apex (EP-Apex) and the Wicked Edge Precision Sharpener (WEPS). This is not saying that any method is superior to any other; it is simply that the author is interested in the geometry of mechanical devices. As someone who loves mechanisms, one can wonder if there are any ways to improve the EP or WEPS. One immediate idea is to improve the precision of the EP and WEPS mechanisms with more accurate parts machined to a ner tolerance, as well as increasing the rigidity of the parts. However, after careful thinking, one may realize that even if these mechanisms were perfectly precise and innitely rigid, that they would not always grind a perfect dihedral angle (informally known as a V-edge). That is, if we used a perfect EP or WEPS to sharpen a tanto knife, then the knife edge would not have a perfectly uniform dihedral angle. There will be a tiny variation in the included angle of the knife bevel. For users who prefer convex edges this as a feature, rather than a aw. However, the EP and WEPS appear to be designed for creating V-edges with as little convexing as possible2 , so, informally, we call these deviations a aw. The main purpose of this document is to study these tiny aws in the design of the EP and WEPS. We will consider four sharpening mechanisms: (1) the Edge Pro Apex, (2) an Edge Pro Apex modied with a spherical joint, (3) the original Wicked Edge Precision Sharpener (WEPS-Gen1), and (4) the newly updated3 Wicked Edge Precision Sharpener (WEPS-Gen2) with spherical joints. As we shall see, the EP and WEPS-Gen1 have slight variations in their sharpening angle. However when these mechanisms are modied to use spherical joints, their designs become perfect with
The knife community uses the term convex informally to mean a knife cross section with curved sides similar to the cross section of a convex lens. This usage is different from the mathematical denition of convex ( https: //en.wikipedia.org/wiki/Convex_set ). 2 Of course, the EP and WEPS are designed to create convex edges by replacing the sharpening stone with a leather strop. However, their most common usage appears to be creating V-edges with sharpening stones, and that is what will be studied here. 3 In early 2013.
1

1.0

be

ta1

7(

20
5

14

.2. 5

+0

8:1

7)

C HAPTER 1

CHAPTER 1. INTRODUCTION

These critiques are valid. Consequently, the results presented here are irrelevant to virtually all users. Instead, our intended audience is very narrow: namely the designers and makers of knife sharpeners, and also knife sharpeners who have a technical interest in mechanisms (for example, a knife user who happens to be a mechanical engineer). We encourage those who are not interested, to simply skip this document. In the following sections of this document, we will rst briey state the overall results followed by a review of dihedral angles, and then analyze the Edge Pro Apex, the so-called stop-collar trick, and the WEPS-Gen1. In passing, we will consider the WEPS-Gen2 as well as the Edge Pro Apex with modications to use a spherical joint, as well as the Professional version of the Edge Pro. Then, we will consider the WEPS-Gen2 and the optimal knife/pivot location for making the sharpening angle as uniform as possible along the entire length of a curved blade. Finally, we briey discus the sharpening of curved knives on belt sanders.

DR

AF

1.0

be

ta1

Consequently, there are four natural critiques of this document. The rst is that most knives have curved cutting edges, so we are only looking at a tiny subset of knives. The second critique is that no mechanism in the real-world is, in fact, made of perfectly accurate parts that are innitely rigid. A third critique is that even if the mechanisms were perfect, the deviation in sharpening angle is very small with typical values on the order of or less than a tenth of a degree ( 0:1 ), and with the largest realistic case being at most half a degree ( 0:5 ). Such tiny deviations in sharpening angle, are simply too small to be noticed. Furthermore, in any practical situation, the slight inaccuracy of parts means that the play or slack in the mechanisms is likely to cause larger changes in angle. Therefore, it is pointless to remove a small theoretical deviation in angle, when, much larger deviations are caused by the imperfections that exist in any practical mechanism. A fourth and nal critique is that we are not considering knives with so-called convex edges; that is, knives which have curved bevels that are not planar.

7(

20

14

.2. 5

To keep the analysis as simple as possible, we explicitly restrict our study to idealized mechanisms which are perfectly precise and innitely rigid. And, for the most part, we restrict ourselves to the sharpening of tanto knives with straight cutting edges. (We consider curved knives in in Chapter 6, where we discuss how the WEPS-Gen2 varies the sharpening angle along a curve. We also briey consider curved knives in Chapter 7 which is about sharpening with a belt sander.)

+0

8:1

no variation in sharpening angle (when used on knives with straight cutting edges). For example, when sharpening a tanto knife, the WEPS-Gen2 is capable of sharpening perfect dihedral edges, and so is the Edge Pro after it has been modied to use a spherical joint.

7)

1.2. SUMMARY OF RESULTS (TL;DR)

For those of you who are not technically minded, here is a summary of the main results, but without explanation or geometrical proof.

Third, this document mainly analyzes knife edges which are straight. That is, the cutting edges of the knife are straight lines, such as in a tanto knife.

DR

AF

Seventh, I would like to note that both the EP-Apex and WEPS-Gen1 designs can be xed by either of two modications: One can replace their pair(s) of pivoting hinges with a single spherical joint. Because the WEPS-Gen2, which uses spherical joints, it grind perfect V-edges. Alternatively, one can take the pair of pivoting hinges and move them as close together as possible, so that their axes intersect. That is, consider two lines, where each line goes through the axis of a hinge. If these two lines intersect at a point, then the sharpener will grind a perfect V-edge. Either of these xes will result in a design, that, under ideal circumstances, would grind a perfect V-edge. Eighth, we analyze the WEPS-Gen2 when sharpening a knife with a curved edge. Ninth, we end with a brief discussion about sharpening V-edges with belt sanders.

My hope is that these geometric considerations are of interest to anyone who design knifesharpeners and/or precision mechanisms.

1.0

be

Sixth, we consider the so-called stop-collar trick which is used to compensate when switching between sharpening stones of different thickness. Mathematically, the stop-collar trick is only approximately correct. However, it is a fairly accurate approximation.

ta1

Fifth, there are cases where the EP-Apex and WEPS-Gen1 can grind perfect V-edges. For the EP-Apex, this is the special case when the plane of the grinding-stone is perfectly perpendicular to the vertical rod in the EP-Apex. That is to say, when one is sharpening a knife edge which is 30 per side (because the EP-Apex platform is inclined at 30 degrees). In a tanto knife, the WEPS-Gen1 will grind a perfect V-edge along the long edge of the knife, but will slightly change the angle along the tantos tip. As we shall see, the design of the WEPS-Gen1 is slightly clever and subtle.

7(

20

14

Fourth, the deviations of the sharpening angles are very small! Typically, the sharpening angle changes by less than a tenth of a degree ( 0.1 ). In the worst case (that is realistic), the angle variation is about half a degree ( 0.5 ).

.2. 5

Second, we assume that the mechanisms are perfect. That is, all parts are innitely rigid, and all mechanical joints (pivoting or sliding) are perfectly precise.

+0

First of all, we are mainly considering the following two mechanisms: the Edge Pro Apex and the rst generation of the WEPS, which well call WEPS-Gen1.

8:1

7)

1.2 Summary of Results (TL;DR)

CHAPTER 1. INTRODUCTION

DR

AF

1.0

be

This document is released under a Creative Commons Attribution Non-Commercial ShareAlike 3.0 license. For details on this license, please see https://creativecommons. org/licenses/by-nc-sa/3.0/. Other licenses may be available upon request; for inquiries, please contact the author at snfe.agaupt@gmail.com.

ta1

7(

A quick note about the level of presentation: the authors background contains university research in physics and computer science. However, because of a very broad audience, this document avoids using terminology/jargon from technical elds, such as robotics and kinematics. This document is presented at a conceptual or semi-technical level, and can be thought of as an informal tutorial, and it is not a paper at the level of technical research. The material here would be elementary for undergraduate students of mechanical engineering (for example, the detailed mechanics of a universal joint in a car transmission). Those who are interested in a more technical discussion should feel free to contact the author at snfe.agaupt@gmail.com and/or consult textbooks on mechanical design and robotics.

20

14

.2. 5

In this PDF document, there are embedded animation videos. To view these animations, we recommend using Adobe Reader version 9.0 or later. Other PDF viewers may not work. In addition, one might need to give Adobe Reader permission to play videos. For those who are unable to view videos embedded within PDFs, they may watch the animations as separate .MP4 video les. These les are included in a folder named Movies within the original .ZIP le that also contains this PDF document. Any software that understands MPEG-4 and/or H.264 should be able to play the included .MP4 videos. For example, VideoLAN and Apples QuickTime Player should work.

+0

8:1

7)

1.3 Animations, Document Note, License, and Copyright

D IHEDRAL A NGLES

A dihedral angle is created when two planes intersect at a line. Another way to say this is, a dihedral angle is the angle of a 3d wedge or ramp. Here is an example of a dihedral angle.

DR

AF

Figure 2.1: Dihedral Angle. A dihedral angle is formed by the intersection of two planes.

To measure a dihedral angle, we can slice the dihedral by any plane perpendicular to the edge of the dihedral. (The edge of the dihedral is the line that is the intersection of the two planes.) This perpendicular slice creates a cross section which is a 2d angle that we measure in the usual way. Of course, there are other cross-sections of a dihedral, but only a perpendicular cross-section is 9

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

C HAPTER 2

10 used to measure the dihedral angle.

CHAPTER 2. DIHEDRAL ANGLES

of intersection of the two planes. The red plane cuts a cross-section of the dihedral angle. This perpendicular cross-section is used to measure the size of the angle of the dihedral.

DR

Before continuing, we need a couple of simple facts about dihedral angles. Imagine a dihedral angle as if it were a door hinge. The two sides of the hinge are the two planes, and the pin they rotate around is their line of intersection. Next, imagine that we x one side of the dihedral, but we allow the other side to ap around. Then, we imagine that this dihedral hinge is spring-loaded, so that it wants to completely open, except it cannot because we are pressing on it with our nger. Our nger is the red arrow in in Figure 2.3. If we move our nger around, we can change this dihedral angle. However, if we move within the plane, then the angle of the dihedral does not change. It is only when our nger moves with a component perpendicular to the plane that the dihedral angle changes. That is, we must lift our nger off of the original plane, in order to change the dihedral angle. To help visualize this more clearly, please consider Figure 2.3. The geometry

AF

1.0

be

Figure 2.2: Measuring a Dihedral Angle. The red plane is perpendicular to the line

ta1

7(

20

14

.2. 5

+0

8:1

7)

11 in Figure 2.3 is the basic foundation for understanding the analysis of the EP-Apex and is closely related to the analysis of the WEPS-Gen1. Arguably, this is the most important gure of this entire document, so it is worthwhile to spend time understanding it.

DR

AF

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

12

CHAPTER 2. DIHEDRAL ANGLES

DR

AF

Figure 2.3: Changing a Dihedral Angle. We imagine a dihedral angle to be hinge, where one side of the hinge
is xed, and the other side must rotate to stay in contact with the tip of the red arrow. The top shows a perspective view, and the bottom shows a side view. In this animation, we show three motions of the red arrow. In the rst motion, the red arrow tip moves within the original plane of the dihedral angle. Notice that this motion does not change the dihedral angle. In the second motion, the red arrows moves forward and backward, which makes the arrow tip move out of its original plane. Notice this changes the dihedral angle. Finally, in the third motion, the red arrow moves in a circular path that is horizontal. This circular path can be thought of as a rotation around a vertical line (drawn in blue). Because the top plane of the dihedral is not horizontal, this circular motion moves the arrow tip out of its original plane, and the dihedral angle changes.

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

T HE E DGE P RO A PEX (EP-A PEX )


3.1 Description of Mechanism

If you are reading this document, then you probably are familiar with the Edge Pro. Here, we will specically analyze the Edge Pro Apex. The analysis for the Professional version of the Edge Pro is very similar, however it appears that the angle of the platform is different.

DR

AF

The Edge Pro Apex (EP-Apex) consists of a main platform which is inclined at 30 , a vertical mast which is attached to a pivot which we call the pin. The other side of the pin is attached to a sliding arm which holds the sharpening stone. (See Figure 3.2.) There are three axes of rotation in this mechanism: The vertical mast, the pin, and the sliding arm. The sliding arm is an axis of 13

Figure 3.1: Photograph of an Edge Pro Apex. This photo has been ipped horizontally

to be approximately consistent with later gures. Photo used without permission from http://www.edgeproinc.com.

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

C HAPTER 3

14

CHAPTER 3. THE EDGE PRO APEX (EP-APEX)

10

8 6 4

be

ta1

ai

7(

nP

lat fo rm

Pin

20

14
z y x

Ta nto K
Vertical Mast

nif e

.2. 5
id Sl g in

+0
m Ar
ng ni e p e ar on Sh St
10 8 2

DR

AF

3.2 Deviations in Dihedral Angle


In this section we consider Edge Pro Apex and its deviations from grinding a perfect dihedral angle.

Figure 3.2: Diagram of an Edge Pro Apex. This diagram shows the various parts an

their names for an Edge Pro Apex.

1.0

8:1

rotation because the sharpening stone can rotate around it. If we consider each axis of rotation to be a line, then all three lines are skew. That is, there is no intersection between any two axes of rotation. Figure 3.3 shows the rotation around the vertical mast, and Figure 3.4 shows rotation of the pivot. The rotation of the sharpening stone around the sliding rod is not illustrated.

7)

3.2. DEVIATIONS IN DIHEDRAL ANGLE

15

Figure 3.3: Edge Pro Vertical Mast Rotation. In this animation we only rotate the

DR

AF

Figure 3.4: Edge Pro Pivot Rotation. In this animation we only rotate the pivot that joins the vertical mast to the sliding rod. The three long blue cylinders represent the three rotating joints of the Edge Pro.

1.0

be

ta1

7(

vertical mast. The three long blue cylinders represent the three rotating joints of the Edge Pro.

20

14

.2. 5

+0

8:1

7)

16

CHAPTER 3. THE EDGE PRO APEX (EP-APEX)

50 45 40 Dihedral Angle [degrees per side] 35 30 25 20 15 10 5 0 6

Edge Pro Apex

45 40 35 30 25 20 15 10 5

ta1

7(

20
0 X [inches] 2 4 6

Figure 3.5: Edge Pro Example Dihedral Angles. The dihedral angle is set at position

DR

AF

3.2.2 Intuitive Explanation


Perhaps the most important thing is to have an intuitive understanding of why the Edge Pro Apex deviates (slightly) from a perfect dihedral angle. In order to explain this intuition, we simplify the geometry of the mechanism as much as possible. In addition to simplifying the geometry of the Edge Pro, we will deliberately exaggerate its geometry for clarity. Then, in the following section, we will return to the full and standard geometry of the Edge Pro, and present results for it.

1.0

X D 0 and then varies as the sharpening stone moves to different positions along the knife edge. Notice that as the dihedral angle approaches 30 , the dihedral angle becomes constant. Also notice that the variations in dihedral angle are generally very small, especially when the sharpening stone is close to the center of the platform (which is when -1 X +1 inches).

be

14

.2. 5

+0

8:1

7)

3.2.1 Summary of Results

3.2. DEVIATIONS IN DIHEDRAL ANGLE

17

DR

AF

Now that we see the sharpening angle (ie: dihedral angle) changes in our simplied Edge Pro, we can ask,Why does this happen? In our Simplied Edge Pro, the sliding arm must go through the point where it intersects the pin of the joint (that connects it to the vertical mast). From our side view, we can see that one side of the dihedral angle must go through this point. This intersection point is very much like our nger in the example of Chapter 2, Figure 2.3.

So if we move this point (ie: nger), then we may or may not change the dihedral angle. If we move this point within the plane of the original dihedral angle, then there is no change in the dihedral angle. However, if we move this point out of the original plane, then the dihedral angle will change. So our question becomes: where does this intersection point move, when we rotate the Edge

1.0

In our example, we can see that the dihedral angle changes as the sharpening stone sweeps over the edge of the knife. To see how much the sharpening angle has changed, we compare the rst and last frame of the animation in Figure 3.9. We are looking at a perfectly aligned side-view, where the line of the knife edge is exactly perpendicular to the page. Therefore, our side view is looking at a perpendicular cross section of the knife edges dihedral angle. This means we can exactly measure the sharpening angle in this view. In Figure 3.10 we compare the initial and nal dihedral angle. The nal angle is represented by the inclined blue line, which is the angle of the Edge Pros sliding arm and also the angle of the sharpening stone. The initial angle is represented by the red line. (If you check, the red line is in the same position as the Edge Pros arm in the rst frame of Figure 3.9.) Clearly the dihedral angle has changed as the sharpening stone has moved along the knife edge.

be

ta1

7(

20

Now that we have fully exaggerated and simplied the geometry of the Edge Pro, we are ready to illustrate that it does not grind a perfect dihedral angle. We now sweep our sharpening stone along the edge of the knife. If there are any changes in the dihedral angle, then they will be visible in the side view. The side view looks directly into the line of the knife edge, so it will see the perpendicular cross section of the dihedral angle. Any visible changes in this perpendicular cross section will represent changes in the measured angle of the dihedral. In addition to a perspective and side view, we also illustrate a top view. Hopefully this makes the geometry completely clear (and intuitive).

14

.2. 5

+0

Here, we explain, step-by-step, the exaggeration and simplications used to demonstrate an intuitive understanding of the EP-Apex. Please see Figure 3.6. We begin with the original EPApex. Next, we exaggerate the geometry by dramatically lengthening the pin. We then simplify by making the sharpening stone and the three rods innitely thin (the vertical mast, the pin, and the sliding arm). Finally, we move the pin slightly forwards and down so that it intersects the vertical mast and also intersects the sliding arm. In Figure 3.6 the nal panel (f) uses red stars to mark the intersection of the pin with the vertical mast and the sliding arm.

8:1

7)

18

CHAPTER 3. THE EDGE PRO APEX (EP-APEX)

12 10 8 6 4 2 0 2 4 10 8 6 4 2 0 zy x

12 10 8 6 4 2 0 2 4 10 8 zy x

(a)

6 12 10 8 4

.2. 5
2 0

+0
(b)
12 10 8

12 10 8 6 4 2 0 2 4 10 8 6 4 2 0 zy x

12 10 8 6 4 2 0

14
8 6 4 2 0 8 6 4 2 0

zy x

ta1

(c)

7(

20
2 4 10 12

10

(d)

12 10 8 6 4 2 0 2 4 10 8 6

12 10 8 6 4 2 0 2 4 10 zy x

1.0

be

zy x

(e)

12 10 8

(f )

DR

AF

Figure 3.6: Simplied Edge Pro Apex. For the sake of intuition, we simplify the geometry of the Edge Pro Apex.
(a) Original Edge Pro Apex. (b) Edge Pro Apex with exaggerated geometry. (c) Sharpening stone is made innitely thin. (d) The three rods are made innitely thin (vertical mast, pin, and sliding arm). (e) We slightly reposition the pin so that it intersects both the vertical mast as well as intersects the sliding arm. (f) Same as (e) except the points of intersection have been marked by red stars.

8:1
12 10 8 12 10 8

7)

3.2. DEVIATIONS IN DIHEDRAL ANGLE

19

Figure 3.7: Sharpening with Simplied Edge Pro (Perspective View). In this animation, we sweep the sharpening stone along the main edge of our tanto knife.

DR

AF

Figure 3.8: Sharpening with Simplied Edge Pro (Top View). We sweep the sharpening stone along the main edge of our tanto knife.

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

20

CHAPTER 3. THE EDGE PRO APEX (EP-APEX)

DR

AF

Figure 3.10: Sharpening with Simplied Edge Pro (Side View). As we sweep the sharpening stone along

the edge of the knife, we can see that the dihedral angle changes. The initial dihedral angle is represented by the red line, and the nal dihedral angle is represented by the diagonal blue line. We are using a perfectly aligned side view, so the two-dimensional angles in this gure represent the dihedral angles.

1.0

be

ta1

7(

20

Figure 3.9: Sharpening with Simplied Edge Pro (Side View). We sweep the sharpening stone along the main edge of our tanto knife. In this side view, please notice that the dihedral angle (sharpening angle) changes as the sharpening stone moves along the length of the knife edge.

14

.2. 5

+0

8:1

7)

3.2. DEVIATIONS IN DIHEDRAL ANGLE

21

3.2.3 Discussion and Comparison to Gimbals and Spherical Joints


We are now in a position to make several key observations. Our rst observation is that there is a special case where the simplied Edge Pro will sharpen a perfect dihedral angle. This happens when a horizontal displacement does not move our nger point out of the top plane of the dihedral. This can only occur, when the top plane of the dihedral is horizontal, because then any horizontal movement of our nger point is still within the same horizontal plane. The Edge Pro Apex has a main platform that is inclined at 30 from horizontal. Therefore, if we sharpen a knife on the Edge Pro Apex at 30 per side, we will have a perfect dihedral angle for our tanto knife. A related observation is that as we sharpen our knife to angle closer-and-closer to 30 per side, the Edge Pro Apex will sharpen an edge which is closer-and-closer to being a perfect dihedral angle. I believe this is why the design of the Edge Pro Apex is inclined at 30 : because most household knives are sharpened at angles which are not too different from 30 per side. The Professional model of the Edge Pro has a different angle of inclination for its main platform. Without access to a Professional Edge Pro, it is unclear what the angle is for the main platform, but from photographs, it appears to be approximately 15 . Our speculation is that the Professional Edge Pro has this design because most professional chefs like to have their knives sharpened at angles close to 15 per side. Of course, the general conclusion is that sharpening at an angle (per side) which is

DR

AF

1.0

be

ta1

7(

20

After some thought, one can see that an un-simplied version of the Edge Pro has a similar effect. Not exactly the same, of course, because our simplied Edge Pro is greatly modied compared to a real Edge Pro. However, the basic idea is the same: a rotation around the vertical mast causes a horizontal displacement of a point on the sliding-arm. This horizontal-displacement causes the dihedral angle to change. (In a full analysis of the real Edge Pro, we would have to remove every single simplifying step taken to create the simplied Edge Pro. At that point, even things, such as the thickness of the sharpening stone, will affect the results.)

14

.2. 5

Pro around its vertical mast? To answer this, we look at the pin which is the joint that connects the vertical mast to the sliding arm. We have grossly exaggerated the length of this pin. Because this pin rotates in a plane perpendicular to the vertical mast, it moves our nger point in a circle that lies in a horizontal plane. In general, this horizontal plane is not parallel to the top plane of our dihedral angle! Therefore, as we rotate the Edge Pro around its vertical mast, our nger point moves out of the plane of the original dihedral angle. (It moves out of the top plane of the original dihedral angle.) For a direct analogy between the EP-Apex and a changing dihedral angle, please see Figure 3.11. Therefore, the dihedral angle changes and the Edge Pro cannot sharpen a perfect dihedral edge (ie: V-edge).

+0

8:1

7)

22

CHAPTER 3. THE EDGE PRO APEX (EP-APEX)

DR

Figure 3.11: Simplied Edge Pro compared to Dihedral Angle. Left column show the perspective, side, and top views of the Simplied Edge

Pro model. The red arrow points at the intersection of the sliding arm and the pin. This intersection is analogous to the tip of the red arrow on the right column. Right column shows perspective, side, and top of a dihedral angle. One side of the dihedral angle is constrained to be touching the tip of the red arrow which moves in a horizontal circle (see Figure 2.3). Notice the change of dihedral angle in both the left and right columns.

AF

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

3.2. DEVIATIONS IN DIHEDRAL ANGLE

23

DR

AF

We note in passing, that several clones and modications of the Edge Pro use these ideas. The rst idea of getting the three axes to intersect is used in clones that have a universal joint, just like in a trucks rear-wheel drive. In a universal joint, there are four axes or joints: The input shaft, the output shaft, and then two joints in the gimbal that connects the shafts. If you look carefully, you will see that these four axes all intersect simultaneously at a single point in the center of the gimbal. (See Figure 3.12.) Without going into any detail, we note that having all four axes intersect is a necessary condition for the universal joint to function properly (otherwise the shafts or other parts would be forced to bend and ex during operation).
1

For those less familiar with scientic measurements, it may be worth reviewing the technical meaning of the terms repeatability, resolution, accuracy, and precision. Some discussion of these terms can be found on Wikipedia: https://en.wikipedia.org/wiki/Repeatability , https://en.wikipedia.org/ wiki/Accuracy_and_precision.

1.0

be

We can also see that shortening the pin also reduces the amount of horizontal displacement, and therefore reduces the deviation from an ideal dihedral angle. (The pin is the joint that connects the vertical mast to the sliding arm.) In fact, if we could somehow make the pin have zero length, then rotation around the vertical mast would have no change in the dihedral angle. In this case, the three axes of the vertical mast, the pin, and the sliding arm would simultaneously intersect at a single point. This three-way intersection would act as if it were a single spherical joint. With this arrangement, our nger point would actually be on the axis of the vertical mast. Then, as the vertical mast rotates, it cannot move any points which lie on its axis, so our nger point would not move at all! The resulting dihedral angle would remain constant.

ta1

7(

20

dihedral angle varies, however it always varies by the same amount for the same rotation of the vertical mast. Each sharpening strategy, in theory, has different tradeoffs. In practice, the deviation in dihedral angle is very small (less than 0.1 typically) and is not noticeable.

14

Also related, is an observation that if we only perform a smaller rotation around the vertical mast, then the Edge Pro will have a smaller deviation in the dihedral angle. This is because smaller rotations will rotate the horizontal pin less, and cause a smaller horizontal displacement of the nger point. This leads to two possible sharpening strategies. In the rst strategy, we only allow minimal rotations of the vertical mast, which means the sharpening stone stays close to the centerline of the platform. Consequently, we have to repeatedly move the knife across the platform to sharpen the entire edge. The second strategy is to not move the knife at all, but to allow the vertical mast to rotate as we sharpen the entire edge. In this second strategy, we allow the dihedral angle to vary, but we rely on the fact that the Edge Pro mechanism is highly repeatable1 . That is, the

.2. 5

+0

8:1

equal to the tilt of the main platform will result in perfect dihedral knife edges (regardless of which Edge Pro model you are using). And, if you sharpen close to this specic angle, then you will get close to an ideal dihedral angle.

7)

24

CHAPTER 3. THE EDGE PRO APEX (EP-APEX)

Figure 3.12: Universal Joint. Used in many mechanisms, including the rear-wheel drive

DR

AF

Figure 3.13: Gimbal. A gimbal similar to this one could be used to guide the sliding

rod of an Edge Pro sharpener. Notice that like a universal joint, every joint and rotating axel has an axis that goes through a single point at at the center of the red ring. One can think of this arrangement of joints as a way to emulate a spherical bearing. Image from http://www.dvinfo.net/forum/stabilizers-steadicam-etc/ 119756-glidecam-4000pro-balance-keeps-shifting-2.html

1.0

be

ta1

7(

20

Closely related to a universal joint, is a gimbal. (Actually, a universal joint contains a gimbal.) A gimbal similar to Figure 3.13 could be used to guide the sliding rod in an EP-Apex. Notice that in a gimbal, all rotating axes intersect at a single point.

14

.2. 5

of trucks. Notice that all four rotating axels/joints have axes that intersect at a single point in the center of the green gimbal. Image from Wikipedia. https://en.wikipedia. org/wiki/Universal_joint

+0

8:1

7)

3.2. DEVIATIONS IN DIHEDRAL ANGLE

25

Figure 3.14: Animated Gimbal. Used in many mechanisms, including gyroscopes, universal joints, and telescope mounts. Notice that all rotating axels/joints have axes that intersect at a single point in the center of the blue ring. Furthermore, the center of the blue ring is a xed point even through all the joints are rotating. Image from Wikipedia. https://en.wikipedia.org/wiki/Gimbal

DR

AF

Figure 3.15: Example Gimbal in a Sharpener. Here is an example of a gimbal used to to guide the sliding rod in a sharpener. Notice that like a universal joint, every joint and rotating axel has an axis that goes through a single point at at the center of the white block. One can think of this arrangement of joints as a way to emulate a spherical bearing. Image by forum user Sticky at http://www.bladeforums.com/forums/showthread.php/ 975542-Lets-see-your-home-made-knife-sharpening-devices/ page2

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

26

CHAPTER 3. THE EDGE PRO APEX (EP-APEX)

Figure 3.16: Edge Pro Modication with a Spherical Rod End. Some users have

DR

We can now see that a gimbal and/or universal joint are effectively the same as a spherical joint. While not exactly the same, they both can constrain a guide-rod in the same way. First, let us consider a gimbal (or universal joint). Let p be the point which is the intersection of all the rotating axes of a gimbal and/or universal joint. Under normal operation, this point p is xed and never moves despite any combination of rotations from the other joints. This is simply due to the mathematical fact that a rotation around an axis cannot move any points which lie on the axis. Because p is the intersection of all rotating axes, it is on each individual axis. Therefore, the position of p cannot be changed by rotating any of the joints individually. It then follows that p cannot be moved by any combination of joint rotations in the gimbal (or universal joint). Therefore, the point p is xed in a gimbal. Next, let us consider a spherical bearing which is mounted to a xed frame. For examples, see Figures 3.16, 3.17, 3.18, 5.18, and 5.19. Let q be the center of the ball in the spherical joint. Notice that the center of the ball, q , never moves even when the spherical joint pivots. If a guide-rod is mechanically constrained to always passes through the center of the spherical joint, then the center

AF

1.0

be

ta1

modied their Edge Pro sharpeners to use a spherical rod-end. Notice that the axis of the sliding arm always intersects the point at the center of the spherical joint. (Image via forum user MadRookie at www.knifeforums.com)

7(

20

14

.2. 5

+0

8:1

Instead of using a universal joint, it is also possible to use a spherical bearing, or a spherical rod-end. This same idea has been around for awhile, and is used in some custom modications of the Edge Pro sharpener, and is also part of the basic design of the KME Knife Sharpening System (see Figure 3.18).

7)

3.2. DEVIATIONS IN DIHEDRAL ANGLE

27

rod end. The spherical rod end is manufactured by Igus Inc. Image from Nosmo on www. bladeforums.com

DR

AF

Figure 3.18: KME Knife Sharpening System. The KME sharpener uses a polymer spherical joint. Image via http://www.kmesharp.com/

1.0

be

ta1

7(

20

Figure 3.17: Clone of Edge Pro with a Spherial Rod End. Notice the polymer spherical

14

.2. 5

+0

8:1

7)

28

CHAPTER 3. THE EDGE PRO APEX (EP-APEX)

3.2.4 Detailed Results

DR

In addition, there are other effects which will vary from knife-to-knife, and user-to-user. For example, the following analysis assumes that the knife edge is exactly at the edge of the Edge Pros platform. This ignores two real-world effects: First, this is impossible because all knives have signicant thickness; therefore the knife edge will be above the platform surface by half the knife thickness (for a knife with symmetric bevels). Second, most users have the knife edge protrude out and beyond the main platform. This means a real knife would have its cutting edge in a different vertical and a different horizontal position than the edge of the platform. These small variations will affect the dihedral angle and how it varies as the Edge Pro sharpens. For the sake of brevity, we have not performed any analyses which account for all possible effects. Instead, we present a single example of results.

AF

1.0

For the record, we did not have physical access to an Edge Pro Apex. Instead, we had to rely on friends who own Edge Pros to take measurements and photographs. So while the geometric model of the Edge Pro Apex may not be perfectly accurate, it should be an excellent approximation. Therefore, the reader should not take the following graphs and numbers too literally, but instead use them as approximations.

be

ta1

Caveats

7(

In the previous section, we used a simplied and exaggerated version of the Edge Pro Apex. In this section, we return to the real world and consider the full geometry of the Edge Pro Apex. We would like to know how much the Edge Pro Apex will deviate from an idea dihedral angle.

20

14

.2. 5

axis of the guide rod always passes through the xed point q . Similarly, for a gimbal, if a guide-rod is mechanically constrained to always pass through the center of the gimbal, then the center axis of the guide rod always passes through the xed point p . Therefore, a gimbal and a spherical joint are similar in that they both constrain the axis of the guide rod to go through a xed point. It is precisely in this way that they are equivalent. In theory, one could replace a spherical joint with a gimbal and vice-versa, without any change in how the sharpener functions. (Of course, they are not the same in other respects, which is why the gimbals in universal joints are used to transmit power from the engine to the rear axel in a truck, but spherical joints are not.) Further discussion and analysis of spherical joints in guided-rod sharpeners can be found in Chapter 6 and also in Appendix A.

+0

8:1

7)

3.2. DEVIATIONS IN DIHEDRAL ANGLE


Geometric Computations

29

To represent the Edge Pro Apex, a kinematic chain was built in software with four components and four joints (three revolute joints and one prismatic joint). For the sharpening stone to touch a specic point on the knife edge, we must solve a kinematic loop-closure problem. At the sharpening stone, we have two constraints: First, there is a positional constraint that a specied point on the sharpening stone must coincide with a specied point on the knife edge. Second, there is an orientational constraint, because the face of the sharpening stone must be a tangent plane of the knife edge. Initially, we relax the orientational constraint, and solve the closure problem using a combination of the following: Closed form solutions were calculated as roots of a single-variable polynomial of degree four. However, in general, these roots are not accurate because the locations of the roots can be highly sensitive to small perturbations of the polynomial coefcients. (In numerical-analysis and scientic-computing, this a known and well-studied problem.) The closed form solutions are then polished by local optimization with a special type of cyclic coordinate descent that is designed to overcome slow convergence caused by the fact that we are not using any methods related to conjugate gradient descent. (A full explanation of these methods is beyond the scope of this document.)

Once the kinematic chain satises positional and orientational constraints, it is straightforward to compute the dihedral angle using elementary linear algebra and trigonometry. Final solutions were checked for positional and orientational constraint satisfaction and typically satisfy these constraints to relative accuracies of better than 10 12 and all solutions satisfy the constraints to a relative accuracy of better than 10

1.0

DR

AF

Graphs of Example Results In this section we present several examples using the Edge Pro Apex. As mentioned earlier in the Section 3.2.4 Caveats, our specic examples will not be the same as any specic case, depending on a number of factors including knife shape, position, and other geometrical effects. Below we present several graphs showing how the dihedral angle varies for the Edge Pro Apex

be

Because we have relaxed the orientation constraint, we get an innite set of solutions which is parameterized by one degree of freedom. To satisfy our orientation constraint, we optimize this one degree of freedom using Brents Method.

ta1

(single precision).

7(

20

14

.2. 5

+0

8:1

Without going into the technical details, here is a brief description of how the results were calculated. This brief description is primarily for those who are familiar with the technical aspects of forward and inverse kinematics, which is the study of mechanisms (typically studied by people in the elds of robotics, computer graphics, and mechanical engineers).

7)

30

CHAPTER 3. THE EDGE PRO APEX (EP-APEX)

5.5 Dihedral Angle [degrees per side]

AF

1.0

4.5

3.5

be

3 6

ta1

Edge Pro Apex. Starting Dihedral Angle = 5 Degrees per Side

7(

20
0 X [inches] 2

In the following graphs, we consider dihedral angles ranging from 5 per side through 45 per side in steps of 5 . On the x -axis of each graph is a position along the knife edge, specied in inches. Initially, the sharpening stone is positioned so that it initially forms the desired dihedral angle when the Edge Pros horizontal arm is inside a plane perpendicular to the knife edge. This contact point is labeled as X D 0. Increasingly negative values of X represent points on the knife edge which are towards the knife handle, and increasingly positive values of X points on the knife edge which are towards the knife tip. To cover a broad range of knife sizes and positions, the sharpening stone is moved over a range of 12 inches (from X D 6 to X D C6 inches). However, the primary area of interest is when 1 X C1 inches because in standard operation the sharpening stone is kept near the centerline of the main platform.

14
4

DR

Figure 3.19: Edge Pro Example Dihedral Angles for 5 per side. Initial dihedral angle
set at 5 per side at X D 0. X represents a position along the knife edge in inches.

.2. 5
6

+0

8:1

under idealized circumstances, namely we assume that the Edge Pro is mechanically perfect and innitely rigid. In addition, the knife edge is assumed to coincide with the edge of the Edge Pros main platform. Lastly, please note that the Professional version of the Edge Pro appears to have a different platform angle than the Edge Pro Apex.

7)

3.2. DEVIATIONS IN DIHEDRAL ANGLE


Edge Pro Apex. Starting Dihedral Angle = 10 Degrees per Side

31

11

10.5 Dihedral Angle [degrees per side]

9.5

8.5 6

0 X [inches]

14
2 4 2 4

Figure 3.20: Edge Pro Example Dihedral Angles for 10 per side. Initial dihedral
angle set at 10 per side at X D 0. X represents a position along the knife edge in inches.

15.8 15.6 15.4 Dihedral Angle [degrees per side] 15.2 15

AF

1.0

14.8 14.6 14.4 14.2 14

13.8 6

be

ta1

Edge Pro Apex. Starting Dihedral Angle = 15 Degrees per Side

7(

20
0 X [inches]

DR

Figure 3.21: Edge Pro Example Dihedral Angles for 15 per side. Initial dihedral
angle set at 15 per side at X D 0. X represents a position along the knife edge in inches.

.2. 5
6 6

+0

10

8:1

7)

32

CHAPTER 3. THE EDGE PRO APEX (EP-APEX)


Edge Pro Apex. Starting Dihedral Angle = 20 Degrees per Side

20.6

20.4

Dihedral Angle [degrees per side]

20.2

20

19.8

19.6

19.4

0 X [inches]

14
2 4 2 4

Figure 3.22: Edge Pro Example Dihedral Angles for 20 per side. Initial dihedral
angle set at 20 per side at X D 0. X represents a position along the knife edge in inches.

25.3

25.2

Dihedral Angle [degrees per side]

25.1

AF

1.0

24.9

24.8

24.7

24.6 6

be

25

ta1

Edge Pro Apex. Starting Dihedral Angle = 25 Degrees per Side

7(

20
0 X [inches]

DR

Figure 3.23: Edge Pro Example Dihedral Angles for 25 per side. Initial dihedral
angle set at 25 per side at X D 0. X represents a position along the knife edge in inches.

.2. 5
6 6

+0

8:1

7)

3.2. DEVIATIONS IN DIHEDRAL ANGLE


Edge Pro Apex. Starting Dihedral Angle = 30 Degrees per Side

33

30

Dihedral Angle [degrees per side]

30

30

30

30

30

30

30

30 6

0 X [inches]

14
2 4 2 4

Figure 3.24: Edge Pro Example Dihedral Angles for 30 per side. Initial dihedral
angle set at 30 per side. X represents a position along the knife edge in inches.

35.5

35.4 Dihedral Angle [degrees per Side]

35.3

AF

1.0

35.2

35.1

34.9

34.8 6

be

35

ta1

Edge Pro Apex. Starting Dihedral Angle = 35 Degrees per Side

7(

20
0 X [inches]

DR

Figure 3.25: Edge Pro Example Dihedral Angles for 35 per side. Initial dihedral
angle set at 35 per side at X D 0. X represents a position along the knife edge in inches.

.2. 5
6 6

+0

8:1

30

7)

34

CHAPTER 3. THE EDGE PRO APEX (EP-APEX)


Edge Pro Apex. Starting Dihedral Angle = 40 Degrees per Side

41

40.8

Dihedral Angle [degrees per side]

40.6

40.4

40.2

40

39.8

39.6 6

0 X [inches]

14
2 4 2 4

Figure 3.26: Edge Pro Example Dihedral Angles for 40 per side. Initial dihedral
angle set at 40 per side at X D 0. X represents a position along the knife edge in inches.

46.2

46

Dihedral Angle [degrees per side]

45.6

AF

1.0

45.4

45.2

44.8

44.6

be

45.8

45

ta1

Edge Pro Apex. Starting Dihedral Angle = 45 Degrees per Side

7(

20
0 X [inches]

DR

Figure 3.27: Edge Pro Example Dihedral Angles for 45 per side. Initial dihedral
angle set at 45 per side at X D 0. X represents a position along the knife edge in inches.

.2. 5
6 6

+0

8:1

7)

3.2. DEVIATIONS IN DIHEDRAL ANGLE

35

50 45 40 Dihedral Angle [degrees per side] 35 30 25 20 15 10 5 0 6

Edge Pro Apex

40 35 30 25 20 15 10 5

ta1

7(

20
0 X [inches] 2 4 6

Figure 3.28: Edge Pro Example Dihedral Angles. The dihedral angle is set at position
X D 0 and then varies as the sharpening stone moves to different positions along the knife edge. Notice that as the dihedral angle approaches 30 , the dihedral angle becomes constant. Also notice that the variations in dihedral angle are generally very small, especially when the sharpening stone is near or above the main plaform (that is, when 1 X C1 inches).

DR

AF

The data in the above graph(s) represent the primary results for this chapter. We hope this is useful and of interest to knife sharpeners as well as hobbyist designers of precision mechanisms.

1.0

be

14

.2. 5

+0

45

8:1

7)

In this next gure, we over-lay all the graphs. It is worth noting that as the sharpening angle gets closer and closer to 30 per side, the dihedral angle becomes more and more constant.

36

CHAPTER 3. THE EDGE PRO APEX (EP-APEX)

DR

AF

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

T HE S TOP -C OLLAR T RICK IS AN A PPROXIMATION

where

is the sharpening angle (per side) for this side of the knife bevel.

DR

In the stop-collar trick, we are simply making the approximation that a D b. This will be perfectly accurate when a and b are measured in the same direction, namely when b is also perpendicular to the two parallel lines. This happens when the two parallel lines are perpendicular to the vertical mast. So, for the Edge Pro Apex, when we sharpen at an angle of 30 per side, the stop collar trick is perfectly accurate (because the Edge Pro Apex has its main platform inclined at 30 ). However, as we sharpen at angles more and more different from 30 , the stop-collar trick will be more and more inaccurate. As a specic example, let us consider the case where we are using the Edge Pro Apex to sharpen 37

AF

1.0

In Figure 4.1(D), we see that we have increased the stone thickness by a distance represented by a, and to compensate, we needed to move the sliding arm upwards (along the vertical mast) by a distance b. Both distances a and b are trapped between two parallel lines (the lines are parallel because we have maintained the same sharpening angle). However, a and b are not equal in general because they are not measuring their distances in the same direction. While a is the perpendicular distance between the parallel lines, b is not. Instead b is strictly the vertical distance between the two parallel lines. Because a is perpendicular to the parallel lines, it is possible to create a right triangle whose sides are a, b, and a segment from one of the parallel lines. Then applying trigonometry, one can see that, a D b cos.30 / (4.1)

be

ta1

7(

20

We begin by considering a side view of the Edge Pro Apex in Figure 4.1(A). We next make a geometry diagram for this in Figure 4.1(B). The long thin blue rectangle represents our sharpening stone. We can now make this stone much thicker while maintaining the same sharpening angle, as in Figure 4.1(C) By over-laying Figure 4.1(B) and Figure 4.1(C) we can see exactly how the geometry is affected by changing the stone thickness. See Figure 4.1(D) for the overlay.

14

.2. 5

The stop collar trick is a technique used to compensate for a change in sharpening stone thickness, when switching from one stone to another. In this section, we describe exactly how much compensation is needed for stone thickness, and then compare that with the stop-collar trick. As we shall see, the stop-collar trick is an approximation that assumes a trigonometric cosine factor is approximately unity.

+0

8:1

7)

C HAPTER 4

38

CHAPTER 4. THE STOP-COLLAR TRICK IS AN APPROXIMATION

aDb

a D b cos.30 D .1 1

/ / //b a // cos.30 / a

b cos.30 cos.30 cos.30

+0
/

D .1 Vertical Error D b aD

.2. 5

cos.30 cos.30 /

DR

AF

For comparison, the typical thickness of paper is around 0.1 mm (for 24 lbs weight copier paper). So a error of around 0.224 mm is about the thickness of two sheets of ofce paper.

1.0

be

only off by 0.0588 degrees when the stone thickness increases by 0.25 inches and we are using the stop collar trick. So although the stop-collar trick is an approximation, it is quite accurate.

ta1

In general, we see that as the sharpening stone gets closer and closer to perpendicular to the vertical mast, then the stop-collar trick will become more and more accurate. And for our specic example with the Edge Pro Apex, we see that sharpening at 15 per side, leads sharpening angle is

7(

20

Plugging in values for our example, we get that the stop collar trick underestimates the compensation needed by about 0.0088 inches (which is 0.224 mm). This is a rather small error. 1 If horizontal distance between the vertical mast and the knife edge is 8 inches, then this small error of 0.0088 inches leads to a dihedral error of about 0.0588 degrees (calculation details omitted). For virtually all practical purposes, this will not be noticeable.

14

8:1
(4.2) (4.4) (4.3) (4.5) (4.6)

7)

a knife at 15 per side, and our stone thickness increases by 0.25 inches. Then using Equation (4.1), we have

39

(A)

14 7( 20
b

.2. 5 ta1

+0
(B)
a

(C)

Figure 4.1: Stop Collar Trick. (A) Side view of Edge Pro Apex. (B) Diagram of Edge
Pro geometry. (C) Same as (B) with a much thicker sharpening stone. (D) Stop collar trick. Notice that the distance a is the increase in stone thickness. The distance b is the required distance we need to compensate for the increased stone thickness. The two distances, a and b, are between two parallel lines. However, they are not measured in the same direction, so they must not be equal, namely a b. Using trigonometry, one can see that a D b cos.30 / where we are sharpening at an angle of degrees per side. (The 30 is from the 30 inclination of the Edge Pro Apexs main platform.)

DR

AF

1.0

be

8:1
(D)

7)

40

CHAPTER 4. THE STOP-COLLAR TRICK IS AN APPROXIMATION

DR

AF

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

The next pivot controls the side-to-side direction of the main guide rod. (See Figure 5.3.)

DR

AF

Like our analysis of the Edge Pro, we assume that the WEPS-Gen1 has a mechanism that is perfectly accurate and that the parts are innitely rigid. Similarly, we will restrict our analysis to the sharpening of a tanto knife. Our main concern will be whether the WEPS-Gen1 can sharpen a perfect dihedral angle for the long edge of a tanto knife. In fact, this is the case. It turns out that once elevation is set for sharpening, that there is no motion of the lowest pivot in the WEPS-Gen1. Instead, only the second higher pivot will rotate. To see why this is, consider the diagram of a dihedral angle in Figure 5.4. Let us consider the lower pivot in the WEPS-Gen1. Suppose we x its angle of rotation. Then, if we rotate the upper pivot back and forth, the guide rod will sweep out a circle. Notice that this circle is in a xed plane; the inclination of this xed plane was determined by the lower pivot. Because this xed plane is actually parallel to the bevel of the knife edge, the WEPS will grind a perfect dihedral angle (V-edge). 41

1.0

be

5.2 Analysis of WEPS-Gen1

ta1

The pivot closest to the base controls the elevation of the main guide rod. (See Figure 5.2.)

7(

In the WEPS-Gen1, the knife is clamped with the knife edge cutting directly upwards and the line of the knife edge is horizontal. The guide rods are on the two sides of the knife, and are each constrained by a pair of pivoting joints (hinges) near the base. (See Figure 5.1.)

20

5.1 Description of Mechanism

14

.2. 5

In this section, we will examine the rst-generation of the Wicked Edge Precision Sharpener (WEPS-Gen1). As we shall see, the WEPS-Gen1 can grind perfect dihedral angles for the main edge of a tanto knife. However, if the tip of the tanto knife is slanted, then the WEPS-Gen1 will not grind it at a constant dihedral angle.

+0

F IRST G ENERATION W ICKED E DGE P RECISION S HARPENER (WEPS-G EN 1)

8:1

7)

C HAPTER 5

42CHAPTER 5. FIRST GENERATION WICKED EDGE PRECISION SHARPENER (WEPS-GEN1)

DR

AF

Figure 5.2: Wicked Edge Precision Sharpener, First Generation (WEPS-Gen1). The
pivot closest to the base controls the elevation of the guide rod.

1.0

be

ta1

7(

simplify this diagram, we have omitted the knife-clamp and have drawn only one guide-rod with its two pivots. In addition, the length of each pivot has been greatly exaggerated in length to make them more visible. This exaggeration does not change the mechanics of the WEPS-Gen1.

20

Figure 5.1: Wicked Edge Precision Sharpener, First Generation (WEPS-Gen1). To

14

.2. 5

+0

8:1

7)

5.2. ANALYSIS OF WEPS-GEN1

43

second pivot controls the side to side direction of the guide rod.

DR

The WEPS-Gen1 design is slightly subtle. If we were to swap the location of the two pivots, it would cause a problem. That is, suppose the lowest pivot in the WEPS-Gen1 controled the backand-forth direction of the guide arm. And suppose the higher pivot controlled the elevation of the guide arm. In this alternative design, the WEPS-Gen1 would not be able to grind a perfect dihedral angle. In fact, the geometric analysis would be somewhat similar to the Edge Pro, where the lower-pivot represents the vertical mast of the Edge Pro, except that the entire Edge Pro would have been turned on its back so that the main platform is now in a vertical plane. As in the case

AF

At this point, it is worth comparing the WEPS-Gen1 to the Edge Pro Apex. Both mechanisms use a pair of joints to constrain the position and orientation of their guide rods. And in both cases, the pair of joints have axes of rotation that are separated by a short distance (ie: the pair of axes do not intersect). However, in the Edge Pro, both joints must continuously rotate to cover the entire knife edge (except for the special case when sharpening at 30 per side). In the Edge Pro, one of the joints is a rotation around the vertical mast which causes a horizontal displacement that in turn changes the dihedral angle (see Section 3.2.2 for details). On the other hand, the WEPS-Gen1 has only one joint continuously rotating as the sharpening stone moves along the tantos edge. As a result, the lower pivot in the WEPS-Gen1 does not continuously rotate, and cannot cause any type of displacement. Then end result is that the WEPS-Gen1 can grind perfect dihedrals on the long edge of a tanto knife.

1.0

be

ta1

7(

20

Figure 5.3: Wicked Edge Precision Sharpener, First Generation (WEPS-Gen1). The

14

.2. 5

+0

8:1

7)

44CHAPTER 5. FIRST GENERATION WICKED EDGE PRECISION SHARPENER (WEPS-GEN1)

Figure 5.4: WEPS-Gen1 Dihedral Angle. In the WEPS-Gen1, the elevation or sharpening angle is controlled by the pivot closest to the base. (See Figure 5.1) The second (higher) pivot in the WEPS is represented by the green line in this diagram. Because the second pivot is perpendicular to the plane of the dihedral, it can rotate to cover the entire knife edge. The red fan of line segments represent possible positions of the guide rod. Notice that the guide rod can cover the entire knife edge even though the location of the second pivot (green line) is xed.

DR

AF

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

5.3. RESULTS

45

5.3 Results
5.3.1 Caveats

5.3.2 Geometric Computations

DR

AF

5.3.3 Graphs of Results


We set up a tanto knife in the WEPS-Gen1 where the knife protrudes 0.5 inches from the top of the knife clamp. And we adjust the WEPS-Gen1 so that it will grind a 15 per side bevel on the main edge of the tanto. This set-up is illustrated in Figure 5.5. The corner of the tanto edge is 3 inches infront of the WEPS-Gen1 center, and the very tip of the tanto is one inch further than the corner.

The same computer program, and the same methods were used as in Section 3.2.4, which resulted in the same levels of relative accuracy. For additional details, see Section 3.2.4.

1.0

be

The caveats are similar to the ones for the Edge Pro. We did not have access to a WEPS-Gen1, so we had to rely on measurements and photographs taken by friends who did own a WEPS-Gen1. As before, we are only considering a perfectly accurate WEPS-Gen1 with innitely rigid parts. And we are only considering the sharpening of the straight tip of a tanto knife. (Of course the non-tip edge is boring, as the WEPS-Gen1 will create a perfect dihedral angle for the long-edge of the tanto.)

ta1

7(

20

14

Considering the above analysis, we can ask,Does this mean that the WEPS-Gen1 geometry cannot be improved? The answer is no, there is an improvement however it is not a major one. Without going into detail, it turns out that the slanted angle of the tantos tip will not be ground as a perfect dihedral angle in the WEPS-Gen1. This can be xed by replacing the pair of joints in the WEPS-Gen1 with either a spherical joint (as in the design of the current WEPS-Gen2), or moving the two pivots close together so that their axes intersect (See the latter part of Section 3.2.2 for dtails). In the following section, we analyze the deviations from a perfect dihedral angle when the WEPS-Gen1 sharpens the tip of a tanto knife.

.2. 5

+0

I believe the WEPS-Gen1 was carefully designed so that the specic order of the pivots would allow it to grind perfect dihedral angles.

8:1

of the Edge Pro, this alternate WEPS-Gen1 would require a continuous rotation about the main mast to cover the length of the knife edge, and this continuous rotation would cause a (vertical) displacement that would change the dihedral angle. (This is very similar to the Edge Pro analysis where a horizontal displacement causes a change in the dihedral angle.)

7)

46CHAPTER 5. FIRST GENERATION WICKED EDGE PRECISION SHARPENER (WEPS-GEN1)

Figure 5.5: Example Set-up of WEPS-Gen1.

DR

In Figure 5.7, each red line is seven inches long. One inch protrudes into the tantos tip, and the remaining six inches go above the tantos main edge. If we were to sharpen an edge along this red line, the dihedral angle would vary. In the following plots, we show how the dihedral angle varies along each of these red lines. Please keep in mind that when we set up the WEPS-Gen1, we adjusted it so that it grinds a 15 per side angle on the main edge of the tanto. This means that the tantos tip is generally not going to be ground at 15 per side, because the tips edge is in a different position and orientation. Also note that when the tips edge has no inclination, that D 0 and the tips edge is the same as the main edge of the tanto. Therefore, in the case that D 0, the WEPS-Gen1 will grind a perfect dihedral angle. But as we incline the direction of the tips edge, the dihedral angle will change as well as vary along the tips edge. In the gures, the corner of the tantos edge is at X D 0 and positive values of X move along the red lines towards the tip of the knife, while negative values move along the read line towards the handle of the knife.

AF

1.0

be

Next, we vary the direction of the edge of the tantos tip. We will use the variable the direction of the tantos tip, as shown in the gures below.

ta1

7(

20
to denote

14

.2. 5

+0

8:1

7)

You can also refer to Figure 5.7 to see the set-up from the side. The coordinate axes are labeled in inches.

5.3. RESULTS

47

Figure 5.6: WEPS-Gen1 Sharpening a tantos tip (Perspective View). We use


denote the edge direction of the tantos tip.

20

14
to

DR

AF

Figure 5.7: WEPS-Gen1 Sharpening a tantos tip (Side View). We use edge direction of the tantos tip.

1.0

be

ta1

7(

.2. 5
to denote the

+0

8:1

7)

48CHAPTER 5. FIRST GENERATION WICKED EDGE PRECISION SHARPENER (WEPS-GEN1)


Wicked Edge Precision Sharpener (Gen1). (15 Degrees per Side). Edge Direction =0

15.1 15.08 15.06 Dihedral Angle [degrees per side] 15.04 15.02 15 14.98 14.96 14.94 14.92 14.9 6

X [inches]

Gen1.( D 0 .) X represents a position along a red line of Figure 5.6 with positive being towards the tip of the knife, and X D 0 representing the corner of the tanto knife.

14.2

14.15 Dihedral Angle [degrees per side]

AF

1.0

be

14.1

14.05

14

13.95 6

ta1

Wicked Edge Precision Sharpener (Gen1). (15 Degrees per Side). Edge Direction =5

7(

20
3 X [inches] 2 1

Figure 5.8: Example Dihedral Angles at the Tip of a Tanto Knife using a WEPS-

14
0

DR

Figure 5.9: Example Dihedral Angles at the Tip of a Tanto Knife using a WEPS-

Gen1. ( D 5 .) X represents a position along a red line of Figure 5.6 with positive being towards the tip of the knife, and X D 0 representing the corner of the tanto knife.

.2. 5
0 1 1

+0

8:1

7)

5.3. RESULTS
13.6 Wicked Edge Precision Sharpener (Gen1). (15 Degrees per Side). Edge Direction =10

49

Dihedral Angle [degrees per side]

13.5

13.45

13.4

13.35

13.25

13.2

X [inches]

Gen1. ( D 10 .) X represents a position along a red line of Figure 5.6 with positive being towards the tip of the knife, and X D 0 representing the corner of the tanto knife.

13.1

13 Dihedral Angle [degrees per side]

AF

1.0

be

12.9

12.8

12.7

12.6

12.5 6

ta1

Wicked Edge Precision Sharpener (Gen1). (15 Degrees per Side). Edge Direction =15

7(

20
3 X [inches] 2 1

Figure 5.10: Example Dihedral Angles at the Tip of a Tanto Knife using a WEPS-

14
0

13.15 6

DR

Figure 5.11: Example Dihedral Angles at the Tip of a Tanto Knife using a WEPS-

Gen1. ( D 15 .) X represents a position along a red line of Figure 5.6 with positive being towards the tip of the knife, and X D 0 representing the corner of the tanto knife.

.2. 5
0 1 1

13.3

+0

8:1

13.55

7)

50CHAPTER 5. FIRST GENERATION WICKED EDGE PRECISION SHARPENER (WEPS-GEN1)


Wicked Edge Precision Sharpener (Gen1). (15 Degrees per Side). Edge Direction =20

12.7

12.6

Dihedral Angle [degrees per side]

12.5

12.3

12.2

12.1

X [inches]

Gen1. ( D 20 .) X represents a position along a red line of Figure 5.6 with positive being towards the tip of the knife, and X D 0 representing the corner of the tanto knife.

12.3

12.2

Dihedral Angle [degrees per side]

AF

1.0

be

12.1

12

11.9

11.8

11.7

ta1

Wicked Edge Precision Sharpener (Gen1). (15 Degrees per Side). Edge Direction =25

7(

20
3 X [inches] 2 1

Figure 5.12: Example Dihedral Angles at the Tip of a Tanto Knife using a WEPS-

14
0

12 6

DR

Figure 5.13: Example Dihedral Angles at the Tip of a Tanto Knife using a WEPS-

Gen1. ( D 25 .) X represents a position along a red line of Figure 5.6 with positive being towards the tip of the knife, and X D 0 representing the corner of the tanto knife.

.2. 5
0 1 1

+0

12.4

8:1

7)

5.3. RESULTS
12.2 Wicked Edge Precision Sharpener (Gen1). (15 Degrees per Side). Edge Direction =30

51

Dihedral Angle [degrees per side]

12

11.9

11.8

11.7

11.5

11.4

X [inches]

Gen1. ( D 30 .) X represents a position along a red line of Figure 5.6 with positive being towards the tip of the knife, and X D 0 representing the corner of the tanto knife.

12.1 12 11.9 Dihedral Angle [degrees per side] 11.8 11.7 11.6 11.5 11.4 11.3 11.2

AF

1.0

be

ta1

Wicked Edge Precision Sharpener (Gen1). (15 Degrees per Side). Edge Direction =35

7(

20
3 X [inches] 2 1

Figure 5.14: Example Dihedral Angles at the Tip of a Tanto Knife using a WEPS-

14
0

11.3 6

DR

Figure 5.15: Example Dihedral Angles at the Tip of a Tanto Knife using a WEPS-

Gen1. ( D 35 .) X represents a position along a red line of Figure 5.6 with positive being towards the tip of the knife, and X D 0 representing the corner of the tanto knife.

.2. 5
0 1 1

11.6

+0

8:1

12.1

7)

52CHAPTER 5. FIRST GENERATION WICKED EDGE PRECISION SHARPENER (WEPS-GEN1)


Wicked Edge Precision Sharpener (Gen1). (15 Degrees per Side). Edge Direction =40

12 11.9 11.8 Dihedral Angle [degrees per side] 11.7 11.6 11.5 11.4 11.3 11.2 11.1 11 6

X [inches]

Gen1. ( D 40 .) X represents a position along a red line of Figure 5.6 with positive being towards the tip of the knife, and X D 0 representing the corner of the tanto knife.

DR

AF

1.0

be

ta1

7(

20

Figure 5.16: Example Dihedral Angles at the Tip of a Tanto Knife using a WEPS-

14

.2. 5
0 1

+0

8:1

7)

5.3. RESULTS
Wicked Edge Precision Sharpener (Gen1). (15 Degrees per Side). 15 =0

53

14.5 Dihedral Angle [degrees per side] =5

14

13 =15 12.5

12

11.5

7(

20
X [inches] 2 1 0 1

DR

AF

The data in this plot are the main results for our analysis of the WEPS-Gen1. Given that a tantos tip is only an inch or two long around X D 0, we see that the deviation in dihedral angle is quite small, typically smaller than 0.5 per side (for 1 X C1 inches). We conclude this chapter by noting that the current WEPS-Gen2 uses a spherical joint (spherical rod end) in its design, and can grind perfect dihedral angles both along the main edge and the tip edge of a tanto knife. For a discussion of spherical joints and possible ways to x the WEPS-Gen1, please see the latter part of Section 3.2.2.

1.0

be

Figure 5.17: Example Dihedral Angles at the Tip of a Tanto Knife using a WEPSGen1. X represents a position along a red line of Figure 5.6 with positive being towards the tip of the knife, and X D 0 representing the corner of the tanto knife. The variable represents the inclination of the tantos tip as shown in Figure 5.6.

ta1

14
=20 =25 =30 =35 =40

.2. 5

13.5

=10

+0

8:1

7)

54CHAPTER 5. FIRST GENERATION WICKED EDGE PRECISION SHARPENER (WEPS-GEN1)

Pro Pack II upgrades. Note the spherical joints for the guide rods. Image from http: //www.wickededgeusa.com

DR

AF

User Modied WEPS-Gen1. Photo of a WEPS-Gen1 that was modied by a user to use large spherical bearings. Image via beltman on http://www.bladeforums.com. (http://www.bladeforums.com/ forums/showthread.php/809918-New-and-improved-Wicked-Edge )

Figure 5.19:

1.0

be

ta1

7(

20

Figure 5.18: WEPS-Gen2. Photo of the new Wicked Edge Precision Sharpener with

14

.2. 5

+0

8:1

7)

6.1 Introduction

DR

user technique can come into play. If a user is careful to consistently position the guide rod so that it always touches the same side of the slot, then the Gatco and Lansky will have repeatable sharpening angles. That is, the sharpening angle may deviate from a perfect dihedral, but the deviations will be the same for the same position along the knife edge. Nevertheless, an analysis of the Gatco and Lansky is potentially complicated by user technique, and so will not be covered here.
1

AF

Users who prefer convex edges may consider this to be a feature of the Lansky and Gatco sharpeners.

Some fully-clamped systems include the Lansky sharpener as well as the similarly designed Gatco sharpener (see Figure 6.1). One interesting aspect of the Gatco (and Lansky) systems, is the long horizontal slots in the sharpening guides. The primary reason for the length of these slots (they are wider than they are tall) is to allow the guide-rod to swing towards the tip and the heel of the knife, so that the sharpening stone may cover the entire knife edge. These wide slots do not affect sharpening of the long edge of a tanto knife (so long as the width of the slot is parallel to the tantos long edge). However, for the tip of a tanto knife, and also for curved edges, the exact positioning of the rod within the wide slot can have an effect the sharpening angle 1 . Here again,

1.0

be

ta1

Of course, it is possible to sharpen curved edges with the Edge Pro sharpeners, however the analysis will be greatly affected by the users technique. This is because the knife could be frequently repositioned during sharpening with an Edge Pro. Other sharpeners, such as the WEPS, typically clamps the blade so that its position is completely xed throughout the sharpening process. This difference is mostly a matter of personal preference. However, because of varying user technique, the Edge Pro is much more difcult to analyze. As a result, this section will focus only on sharpening systems where the knife is completely xed by a clamp.

7(

20
55

14

So far in our discussions, we have only considered blades with straight cutting edges, such as a tanto knife. In the more general case, knives have curved edges, such as any of the spear-point, drop-point, trailing-edge, recurve, etc. knife shapes. In this section, we consider sharpening knives with curved edges using the Wicked Edge Precision Sharpener generation 2 (WEPS-Gen2).

.2. 5

+0

O PTIMAL P IVOT P LACEMENT FOR K NIVES WITH C URVED E DGES

8:1

7)

C HAPTER 6

56

CHAPTER 6. OPTIMAL PIVOT PLACEMENT FOR KNIVES WITH CURVED EDGES

Figure 6.1: Gatco Knife Sharpener. Long slots in the knife clamp are used to guide the metal rods attached to the sharpening stones. Image used without permission from http://www.gatcosharpeners.com/ )

DR

AF

6.2 Sharpening Angle of Curved Blades


Before we can begin our analysis, we need an understanding of what a sharpening angle is for a curved knife. This section will give a simple explanation, that avoids going into the full calculus and differential geometry required for a mathematical denition.

1.0

be

ta1

7(

20

14

.2. 5

Therefore, we turn our attention to a case which is more easily studied, namely a guided-rod sharpener with fully-clamped knife and a guide rod that is completely constrained at one end. Examples of this would include the WEPS (Gen1 and Gen2), the KME Sharpening System, and others. Most of these sharpeners have a mechanism similar to either the WEPS-Gen1, with a pair of pivots close together, or a mechanism similar to the WEPS-Gen2, where a spherical joint is used to guide each rod. For example, the KME Knife Sharpening System uses a spherical joint (see Figure 3.18) and so does the WEPS-Gen2 (see Figure 5.18). To make the analysis simple and concrete, we will consider the WEPS-Gen2. (A similar analysis exists for the WEPS-Gen1, but it made more complicated by the variations in dihedral angle that are caused by the effects discussed in Chapter 5.)

+0

8:1

7)

6.3. OPTIMAL PIVOT PLACEMENT FOR THE WEPS-GEN2

57

6.3 Optimal Pivot Placement for the WEPS-Gen2


For the WEPS-Gen2, the knife is fully clamped and typically is not repositioned during the sharpening process. As a result, one may ask, what is an optimal position for clamping the knife? Or, equivalently, we can ask, given a specic knife shape, what is the best place for the spherical bearing in the WEPS-Gen2? That is, different relative positions between the knife and spherical joint will cause different variations in sharpening angle. Which position minimizes these deviations? We call this the optimal pivot placement.
2

DR

AF

Mathematically, we say the knife edge is a differentiable curve.

1.0

So now we know how to dene the dihedral angle of a curved knife: We rst choose a specic point p on the knife edge. Next, we zoom into that point until, in the limit, the knife edge becomes straight therefore has a straight dihedral angle. This limiting dihedral angle is our denition of the dihedral angle at point p on the knife edge. Furthermore, this limiting dihedral angle has a edge which is a line tangent to the original knife edge.

be

ta1

zooming in will make it look more and more like a straight edge. 2

7(

Third, there are cases where zooming in innitely does not make the knife edge more and more like a line. Just consider a tanto knife with a corner between its main edge and the edge of the tip. If one magnies the corner, then it still remains a corner. For example, under a microscope, the 90 corner of a square is still a 90 corner. However, if the knife edge is a smooth curve then

20

The second point is to consider the line that results from innitely zooming in. Suppose we draw a true line along this straight edge and then zoom out all the way back to our circular knife. What do we get? That line we drew is simply a line tangent to the circle, and contacts the circle at one point. If we zoom back into the contact point, then the circle gets straighter and straighter as well as closer and closer to the line, until, in the limit, they become indistinguishable.

14

.2. 5

There are three important observations to make. First, we zoomed into a specic point on the knife edge. So this dihedral angle is only for this specic point. To get the dihedral angle for another point on the knife edge, we would have to zoom into that other point.

+0

We already have a denition for the dihedral angle for a knife with a straight edge (see Chapter 2). When we have a curved knife edge, we can imagine zooming into it with a microscope. For example, let us consider a circular knife. We pick a point on the circular edge, and then zoom in as if with a microscope. (See Figure 6.2.) As we zoom in, the edge becomes straighter and straighter. The curved edge looks more and more like a straight edge. In the limit of innite zoom, the circular edge becomes a line. Now that the edge is a line, we can use our denition of dihedral angle for knives with straight edges.

8:1

7)

58

CHAPTER 6. OPTIMAL PIVOT PLACEMENT FOR KNIVES WITH CURVED EDGES

1x

2x

8x

7(

20
16x 32x

1.0

64x

be

ta1

128x

14
256x

DR

AF

Figure 6.2: Zooming Into a Curved Knife. We consider a circular knife. As we zoom

in, the knife edge appears straighter and straighter. In the limit of innite magnication, the knife edge becomes a line.

.2. 5

+0
4x

8:1

7)

6.3. OPTIMAL PIVOT PLACEMENT FOR THE WEPS-GEN2

59

Figure 6.3: Simplied WEPS-Gen2 Dihedral Angle. The intersection of the green and
red lines is a point that represents the spherical joint of the WEPS-Gen2 (see Figure 5.18). The fan of red lines represents possible positions of the guide-arm. As this diagram illustrates, the WEPS-Gen2 can grind a perfect dihedral angle on a knife with a straight cutting edge.

DR

AF

To study the WEPS-Gen2, we will imagine the following set-up. The plane of the knife is perfectly vertical (ie: perpendicular to the ground plane). We then place the spherical joint at some xed position. Then we ask, what knife silhouettes can the WEPS-Gen2 sharpen at a constant dihedral angle? Of course from earlier chapters, we know that the WEPS-Gen2 can sharpen straight edges at a pefect dihedral angle (see Section 3.2.3 and Chapter 5). This is illustrated in Figure 6.3

1.0

be

ta1

7(

20

14

.2. 5

Naturally, there is not a single optimal pivot placement because there are different ways to measure the size of all the deviations. For example, are we interested in minimizing the average deviation, or are we interested in minimizing the maximum deviation? Or, for those familiar with curve-tting, are we interested in minimizing the root mean squared (RMS) deviation? Rather than pick one specic measure of overall deviation, we instead show the types of knife shapes that can be sharpened with perfectly uniform dihedral angles along the entire curved edge. As we shall see, the WEPS-Gen2 can perfectly sharpen knives that have a silhouette that is a straight main edge which smoothly transitions to a tip that is a circular arc. We leave it up to the user, as to how best approximate any given knife by these shapes.

+0

8:1

7)

60

CHAPTER 6. OPTIMAL PIVOT PLACEMENT FOR KNIVES WITH CURVED EDGES

Figure 6.4: WEPS-Gen2 Dihedral Angle. Here we demonstrate that a WEPS-Gen2 (full geometry with no
simplications) can sharpen a perfect dihedral angle. The sequence of steps in this animation are sufciently well-dened, that they could be used as the basis for a mathematical proof. However, instead of showing the proof, we animate the conceptual steps of the proof. We construct several positions of the WEPS-Gen2, and then show that these constructed positions will place the sharpening stone on the edge of a dihedral knife edge. We start with a plane that contains the spherical joint of the WEPS-Gen2 (this is represented by the red dot). Next, we show ve positions of the guide rod (thick blue lines). We place the axis of these guide rods within the plane. Next, we align the sharpening stones so that their top faces (away from the knife) are aligned parallel to our plane (in the animation, these faces are actually on the plane). Because of this parallel alignment, the grinding face of the sharpening stones are all within a new plane. This new plane is an offsetplane which is parallel to our initial plane. Since all the sharpening stones are grinding this offset-plane, the WEPS-Gen2 is able to sharpen a dihedral angle where the offset-plane is one side of the dihedral angle. For clarity, the animation ends with only the dihedral angle and the grinding faces of the sharpening stones.

DR

AF

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

6.3. OPTIMAL PIVOT PLACEMENT FOR THE WEPS-GEN2

61

Figure 6.5: Circular Knife Edge (Perspective View). The perfectly vertical plane represents the plane of the knife. The blue cone represents the bevel of a circular knife edge. The apex of the cone represents the spherical joint of the WEPS-Gen2. Finally, the inclined plane represents the dihedral angle for the part of the knife with a straight edge. Notice that the sharpening angle of the blue cone is the same as the angle of the black dihedral. Also notice that the cone and the dihedral angle touch along one line segment that runs from the apex of the cone to the red dot on the edge of the dihedral angle.

DR

AF

One critical point to notice, is that the spherical joint is located at the apex of the cone. One may wonder if other positions of the spherical joint are possible. Also, one may wonder if other

1.0

be

ta1

7(

20

14

.2. 5

What about curved edges? After a little reection, it is natural to consider circular blades. In a circular blade, one side of the knife bevel is part of a cone. From solid geometry, the shape of the circular bevel is a truncated cone. If a cone is circular and also not oblique, then the rim of the cone has a constant and uniform angle. See Figure 6.5 and Figure 6.6.

+0

and also in Figure 6.4. In Figure 6.3, we consider a simplied version of the WEPS-Gen2, where the thickness of the guide-rod is zero, and the thickness of the sharpening stone is also zero. This greatly simplies the geometry, and except for Figure 6.4, we will only consider this simplied WEPS-Gen2 to keep the discussion clear. For the sake of completeness, Figure 6.4 demonstrates that the full geometry of an un-simplied WEPS-Gen2 can sharpen a perfect dihedral angle for a straight edge.

8:1

7)

62

CHAPTER 6. OPTIMAL PIVOT PLACEMENT FOR KNIVES WITH CURVED EDGES

DR

AF

6.4 The Dihedral Triangle Theorem


We now prove our key geometric theorem using a combination of elementary solid geometry and vectors. The proof presented is only an informal sketch. This is for brevity and clarity; otherwise it would be necessary to clearly state the various axioms and theorems that are used from elementary geometry and linear algebra. (Our name for this theorem is informal; as far as we know, it is sufciently elementary that it does not have an ofcial name.)

1.0

shapes (other than lines and circles) are possible as perfect dihedral angles on the WEPS-Gen2. It turns out, no other shapes are possible; any knife silhouette that can be perfectly sharpened at a uniform angle on the WEPS-Gen2 must be comprised of line segments and circular arcs where the circular arcs are all part of the same circle and the line segments are tangent to their adjacent circular arcs. This may not be obvious, so briey, we rst prove a useful theorem. We will then use this theorem as a tool to show our nal result. Those who are not interested in the details of a proof may wish to skip the following section.

be

ta1

7(

Figure 6.6: Circular Knife Edge (Side View). Side view of Figure 6.5.

20

14

.2. 5

+0

8:1

7)

6.4. THE DIHEDRAL TRIANGLE THEOREM


Theorem 6.4.1 (Dihedral Triangle Theorem) Please see Figure 6.7 Given: (1) A plane S (2) A point p not on S (3) A plane T that contains p and intersects S at a line l (4) The dihedral angle formed by S and T has a measure strictly less than 90 .

63

DR

AF

We now can start at r and travel to q and to p by using vectors k and n. Because vectors k and n are perpendicular to l , it follows that the triangle 4pqr is in a plane perpendicular to line l . (Here we are using some elementary solid geometry.) We can now make the following conclusions:

(a) Because vector n is perpendicular to every line in plane S , it must be perpendicular to the

Similarly, let k be the vector which points from r to q . Because r is the perpendicular projection of q onto line l , we know that k is perpendicular to l . (Here we are using the denition of perpendicular projection of a point onto a line.)

1.0

Let n be the vector which points from q to p . Because q is the perpendicular projection of p onto plane S , we know that vector n is perpendicular to every line in plane S . (Here we are using the denition of perpendicularity between a line and a plane, plus the denition of perpendicularity between a vector and a plane.) In particular, n is perpendicular to line l which is in plane S .

be

If we can show that triangle 4pqr is in a plane perpendicular to line l , then it will directly follow that qrp has the same measure as our dihedral angle. Conclusions (a), (b), and (c) will then follow.

ta1

S KETCH OF P ROOF. When two distinct planes S and T form a dihedral angle, they intersect at a line l . Furthermore, by denition, the measure of the dihedral angle is taken in a plane perpendicular to line l (See Chapter 2).

7(

20

Then: Let q be the perpendicular projection of p into plane S . Let r be the perpendicular projection of q onto line l . We conclude: (a) The triangle 4pqr is a right triangle with right angle at vertex q . (b) The measure of angle qrp is the same as the measure of the dihedral angle formed by the planes S and T . (c) The distance between q and r must be equal to d D knk= tan. /, where n is a vector that points from p to q .

14

.2. 5

+0

8:1

7)

64

CHAPTER 6. OPTIMAL PIVOT PLACEMENT FOR KNIVES WITH CURVED EDGES

(b) Point r is on line l and point p is not on line l (because p is not on the plane S by assumption (2)). From the above, it follows that the measure of prq is the same as the measure of the dihedral angle.

where

is the measure of the dihedral angle. Then d D kkk D knk= tan. /. 2

be

ta1

7(

20
n p T

DR

AF

Figure 6.7: Dihedral Triangle Theorem. We project point p onto the plane S to obtain

point q . We then project point q onto line l to obtain point r . The plane T goes through point p and forms a dihedral angle with plane S . The line l is the intersection of planes S and T . See Theorem 6.4.1 for a full description.

1.0

14

.2. 5

(c) Let d be the distance from q to r . Using some trigonometry, we see that tan. / D knk=kkk

+0

8:1

7)

line segment qr . Therefore pqr is a right angle.

6.5. CONSEQUENCES OF THE DIHEDRAL TRIANGLE THEOREM

65

ta1

7(

20
h p T

DR

AF

So now we can ask an interesting question. Suppose our WEPS-Gen2 has its spherical joint

Figure 6.8: Dihedral Triangle in the WEPS-Gen2. This gure is similar to Figure 6.7.

We project point p onto plane S to obtain point q . We use h to denote the length of line segment pq . We project q onto line l to get point r . We use d to denote the length of line segment qr . As before in Figure 6.7, plane T contains p and forms a dihedral angle with S where l is the line of intersection between S and T .

1.0

be

14

.2. 5
u

To apply Theorem 6.4.1 to knife sharpening, we consider the following. Suppose we want to sharpen our knife at degrees per side. Then let S be a vertical plane which represents the plane of our knife as clamped in the WEPS-Gen2. Let us pick some arbitrary point u on the knife edge. Then there is a line l that is tangent to the knife edge at u. (See Figure 6.8.) For a curved edge, we want the sharpening angle at point u to be degrees. To be precise, we want the sharpening stone to be in a plane T such that it intersects plane S at line l where l is tangent to the knife edge at u and the dihedral angle formed is degrees.

+0

8:1

7)

6.5 Consequences of the Dihedral Triangle Theorem

66

CHAPTER 6. OPTIMAL PIVOT PLACEMENT FOR KNIVES WITH CURVED EDGES

DR

AF

In addition to the visualization above, it may be useful to consider an alternate way of conceptualizing the geometry. In Figure 6.8, suppose the following are xed: the spherical joint at p and its projection q and also the point u. However we will allow the line l to rotate around the xed point u. Keep in mind that l represents a tangent line of the knife edge. So, changing the direction of line l is equivalent to changing the direction of the knife edge by modifying its belly or silhouette. Alternatively, this can be thought of as rotating the entire knife within the plane of the knife. We can now ask, how does the dihedral angle change as we rotate l around the point u? We can answer the question by applying the Dihedral Triangle Theorem. Using Figure 6.8, the theorem tells us that d D h= tan. /. Solving for we get D arctan.h=d /. Our variable h is

1.0

Finally, we reach our primary conclusion: That the WEPS-Gen2 can only sharpen a constant dihedral angle on knives with curved edges, if the curve(s) are arcs of the same circle. The most common shape would be a knife with a main edge that is straight, followed by a circular arc at the knife tip. Notice that the spherical pivot must be located at a point that projects into the center of the circle. See Figure 6.11. For knives which are not exactly this shape, one can use the best approximation of the knife shape by a circular arc and a line that is tangent to the arc.

be

ta1

7(

In Figure 6.9, we have draw a head-on view of plane S and an example of some lines l 2 L. And in Figure 6.10 we animate a side view of both the set L (blue lines) plus we also draw the cone of possible guide-rod positions (black lines). The apex of the code is the location of the spherical bearing in the WEPS-Gen2.

20

So what is the set of possible lines L? Theorem 6.4.1 gives the answer. If our spherical joint is at p , then let h be the perpendicular distance from p to plane S . (See Figure 6.8.) Every line that can be sharpened at degrees is a distance d D h= tan. / from point q . Here we are using the fact that the perpendicular projection of a point q onto a line l gives us the point r which is the point on l which is closest to q . So we have discovered a necessary condition for all lines l 2 L. Without going into details, one can show that this condition is also sufcient. (If the reader understood the proof of Theorem 6.4.1, then the proof of sufciency is straightforward.) With necessary and sufcient conditions, we conclude: L is the set of all lines in plane S that are a distance d from point q .

14

.2. 5

+0

centered at point p . What is the set L that contains all lines l that can be sharpened at degrees? If our set L were to contain every possible line in plane S , then we could sharpen any curved knife edge at degrees per side. However, if our set L does not contain all lines in plane S , then there are some curves which we cannot sharpen at degrees. To be precise, if a curve has a point u with a tangent line l L, then we cannot sharpen the curve at degrees at that point u. In other words, every curve that we can sharpen at degrees must have only tangent lines in L.

8:1

7)

6.5. CONSEQUENCES OF THE DIHEDRAL TRIANGLE THEOREM

67

DR

AF

Figure 6.9: The Set L. This is a head-on view where the plane S is in the plane of the page. The blue lines represent examples of lines l 2 L and are used to illustrate the set L.

1.0

be

ta1

7(

20
q
0 1 2 3 4

14
d

.2. 5

+0

8:1

7)

68

CHAPTER 6. OPTIMAL PIVOT PLACEMENT FOR KNIVES WITH CURVED EDGES

DR

AF

Figure 6.10: Animation of Set L (Side View). The black lines running down the side of the cone represent possible positions of the guide-rod in the WEPS-Gen2. The red point at the apex of the cone is p the spherical joint in the WEPS-Gen2. The second red point at the base of the cone is q which is the projection of p into the plane S . The blue lines which are tangent to the rim of the cone are examples of lines l 2 L. Any curve which can be sharpened at a given degrees must only have tangent lines l 2 L.

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

6.5. CONSEQUENCES OF THE DIHEDRAL TRIANGLE THEOREM

69

7(

DR

AF

Figure 6.11: Sharpening at a Constant Angle with the WEPS-Gen2 (Side View) This is a head-on view where the plane S is in the plane of the page. Here we show an example of a knife shape that can be sharpened at a constant angle with the WEPS-Gen2. Notice that the knife edge contains a straight part which is tangent to the circle. The tip of the knife is an arc of the circle. The position of the spherical joint must be placed at a point which would project into the center of the circle q .

1.0

be

ta1

20
q

14

.2. 5

a given point u, the smallest dihedral angle is obtained when the line l is far away as possible, namely when d is equal to the distance between u and q . This occurs when line l is perpendicular to the line segment uq . So for a given point u, we have lower and upper bounds on the possible dihedral angles. The lower bound is D arctan.h=dmax / where dmax is the length of line segment uq . The upper bound is 90 .

+0

Furthermore, has an inverse relationship with d ; namely as d gets larger and larger, becomes smaller and smaller (and vice versa). Notice that if the tangent line l goes through the point q , then d D 0. In that case we have D limd !0 arctan.h=d / D arctan.C1/ D 90 . Also, for

8:1

xed, because we are assuming that the spherical joint p is xed. So that means the dihedral angle is only varies with d . We now have our answer: If everything else is xed, then given a tangent line l , the dihedral angle only depends on the distance between the tangent line and the point q .

7)

70

CHAPTER 6. OPTIMAL PIVOT PLACEMENT FOR KNIVES WITH CURVED EDGES

DR

For the WEPS-Gen2 this is impossible, so we instead turn the knife upside-down so that its edge is cutting down into the ground. Now the concave knife edge can be approximated by a circle whose center is below the knife. We can then locate the spherical joint accordingly. See Figure 6.13. In this case, we are able to sharpen a perfect dihedral angle on a karambit styled blade. However, the sharpening stones must cross-over from one side of the knife to sharpen the bevel on the opposite side. That is, the spherical joint on the right side of the knife is used to sharpen the bevel on the left side (and vice versa). While this is correct in theory, it may be cumbersome in practice.
3

AF

http://www.bladeforums.com/forums/showthread.php/1131476-Geometry-and-Kinematics-of-Guid p=12923640#post12923640

1.0

Optimal pivot-placement for a recurve or karambit appears to be unusual. We can still apply the Dihedral Triangle Theorem (see Theorem 6.4.1), however the consequences are unconventional. The following solution was suggested by on www.bladeforums.com by user brplatz 3 . Using the Dihedral Triangle Theorem, we approximate the concave curve of the blade with a circle. The spherical joint of the WEPS-Gen2 must be aligned with the circles center as before: a projection of the spherical joint into the plane of the knife must land on the circle center. If the knife edge is facing upwards, this would mean that the spherical joint must be above the knife clamp!

be

ta1

7(

A less common shape of knives includes karambit and recurve blades. These shapes are unusual, and typically not as easily handled by most sharpening systems. Because their silhouettes contain concave curves, they cannot be sharpened by wide at sharpening stones. However, they may be sharpened using narrow stones with rounded edges, abrasive rods, and/or sharpening stones with cylindrical surfaces. For example, in Figure 6.12, one can see WEPS sharpening stones with curved surfaces, and are specically designed for recurve and karambit shaped knives. It is worth noting that the curvature appears to be approximately cylindrical where the cylinder axis is at the center of the guide rod. This prevents any twist rotation from changing the sharpening angle. Contrast this with a large wide and at sharpening stone: a twist rotation along the guide rod will tip up the stone on one of the stones edges. This tipping motion will move the guide-rod out-of-plane from the original dihedral angle, and thus change the sharpening angle. Unlike a wide at stone, an appropriate cylindrically shaped stone will maintain the same angle despite any twist. This may not be an issue for thin rectangular stones, because their two rounded edges will be simultaneously in contact with the knife edge, which will constrain the twist. However, if the two point contacts are not well rounded, then they may tend to gouge the knife edge.

20

14

.2. 5

+0

8:1

7)

6.6 Optimal Pivot Placement for Recurves on the WEPS-Gen2

6.6. OPTIMAL PIVOT PLACEMENT FOR RECURVES ON THE WEPS-GEN2

71

Figure 6.12: WEPS Sharpening Stones for Curved Blades. These stones from Wicked

DR

AF

1.0

Edge are designed for sharpening knives that have a curved and concave silhouettes. Notice that the stone surface is cylindrical, with the axis of cylinders located along the central axis of the guide rod. This prevents any advertent twisting of the stone from changing the sharpening angle.

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

72

CHAPTER 6. OPTIMAL PIVOT PLACEMENT FOR KNIVES WITH CURVED EDGES

ta1

DR

AF

Figure 6.13: Optimal Pivot Placement for a WEPS-Gen2 on a Karambit Knife. In the top
gure, we have the typical shape of a knife that can be sharpened to a perfect dihedral angle as described in Section 6.5. In the bottom gure, we have a karambit knife with the edge cutting downwards into the ground. We then follow the consequences of the Dihedral Triangle Theorem. Assuming that the karambits edge is circular, we can sharpen it to a perfect dihedral angle so long as we place the spherical joint in alignment with the circle center. That is, a projection of the spherical joint center into the vertical plane of the knife, will land on the circles center. Because the cutting edge is facing downward, the guide-rods will have to cross over to the other side of the knife. The spherical joint on the right side of the knife will be used when sharpening the bevel on the left side, and vice-versa.

1.0

be

7(

20
q

14

.2. 5

+0

8:1

7)

6.7. CASE STUDIES OF CURVED BLADES ON THE WEPS-GEN1, WEPS-GEN2, AND EP-APEX 73

6.7.1 Methodology

DR

AF

If the knife has a signicant circular silhouette, then the knife is clamped so that the spherical joint of the WEPS-Gen2 would project into the circle center as described in Section 6.5. For the WEPS-Gen1, the midpoint between the two pivots is placed at the circle center. Unfortunately for the Edge Pro Apex, it is typically not possible to place the vertical mast at the circle center. This is because the distance from the vertical mast to the edge of the main platform is xed. The horizontal distance from the vertical mast to the edge of the main platform is approximately 8 inches, so it is not possible to bring the vertical mast into alignment for circles of radii smaller than 8 inches. Consequently, for the Edge Pro Apex, the relative position between the knife and the sharpener is chosen somewhat arbitrarily. Typically, we have chosen the vertical mast to line up with a position approximately half way down the length of the blade. We illustrate the position of the vertical mast in our results, so the reader may judge if they are reasonable. After the knife is positioned either in the WEPS clamp or the AP-Apex main platform, we next adjust the sharpener to achieve a target dihedral angle at a single calibration point on the knife edge. We use the same calibration point and target angle for all three sharpeners. Therefore, at the

1.0

be

ta1

Here we describe how these case studies were performed. First a photo is found that has a direct side-view of the knife. Beyond visual inspection, no consideration was given to perpendicularity of the camera view or other technical issues such as perspective transform or lens distortion. However, we do present the original photographs used so the reader may judge their reasonableness. From the photo, the cutting edge of the knife was traced as a polygonal chain and interpolated with cubic splines.

7(

20

14

In this section, we consider three knives with curved edges: a chefs knife, a khukuri, and the Spyderco LionSpy pocket knife. The khukuri and LionSpy were selected for their unusual and extreme shapes which illustrate the geometry discussed above as well as in earlier chapters. The results in this section should not be taken too literally because these studies assume that the knives are never repositioned or re-clamped throughout the entire sharpening process. That is, even on the Edge Pro Apex, we pretend that these knives are clamped to the main platform and are never moved. This assumption is arguably unfair. However, it allows us to perform an apples to apples and oranges to oranges comparison. In practice, one would either reposition (EP-Apex) or re-clamp (WEPS) large knives like the khukuri several times during sharpening. Therefore, the results here should be considered with these caveats clearly in mind.

.2. 5

+0

8:1

Gen2, and EP-Apex

7)

6.7 Case Studies of Curved Blades on the WEPS-Gen1, WEPS-

74

CHAPTER 6. OPTIMAL PIVOT PLACEMENT FOR KNIVES WITH CURVED EDGES

6.7.2 Description of Figures and Results

However, the EP-Apex has a main platform that is inclined 30 from horizontal. In a top-down view, one would not see a perfect silhouette of the knife, because it has been rotated by 30 to
Although not discussed in earlier chapters, this effect is much easier to analyze for tanto knives. For a tanto the contact between the sharpening stone and knife is a line, not a point. This means that when the stone contacts the knife at a point off of the center-line, the stone center-line is in contact with a nearby point on the knife edge. Therefore in a plot of dihedral angle versus position on a tanto knife, one can simply look a short distance to the left and right on the graph to see the maximum and minimum dihedral angles. For example, if the sharpening stone is in contact with the tanto edge and is also perpendicular to the line of the edge, then one can look at points 0.5 inches away on the graph (assuming the sharpening stone is 1 wide). However, the stone could be in contact with the tanto edge, but tilted at an angle so that the stone is not perpendicular to the line of the edge. In that case, one would have to look beyond 0.5 inches away because slicing a rectangle at a diagonal angle has a longer length (the length is increased by a trigonometric factor). From Figures 3.28 and 5.17, we can see that this effect is typically extremely small. 5 The author would like to thank forum member bubo for suggesting the layout of gures in this section. See http://www.cliffstamp.com/knives/forum/read.php?5,21981 .

DR

AF

1.0

For each knife, results are shown by two aligned pictures 5 . For an example, see Figure 6.15. The rst picture is a graph of dihedral angle versus horizontal position. The second picture is a diagram of the knife silhouette which is accurately aligned to the graph. In addition, we show the calibration point as well as the position of the WEPS-Gen2 spherical joint (for the WEPS-Gen1, we show the midpoint of the two joints), and the position of the EP-Apexs vertical mast. It is important to note that for the WEPS-Gen1 and WEPS-Gen2, this diagram is fully accurate.

be

ta1

7(

For the dihedral sharpening angle, we assume that the sharpening stone contacts the knife edge on the center-line of the sharpening stone, where the center-line is parallel to the guide-rod that holds the stone. Of course this is not true in general. However, the change in angle between contacting on the center-line of the stone versus contacting off of the center-line is extremely small (largely in part because the sharpening stones are only about an inch wide). Technically, each point in the graphs should show an error bar which shows the smallest and largest dihedral angle possible for different points of contact on the sharpening stone 4 . To keep the discussion straightforward, we will not consider this effect.

20

14

.2. 5

As mentioned, cubic splines are used to represent the knife edge. These splines were sampled at approximately every 1/4 of edge-length (arc length) or closer. These sample points divide the knife edge into short segments. Each segment was approximated as a line segment, and a midpoint of the line segment is found. This is done because the sharpening angle depends on both the position of a point on the knife edge, and also depends on the direction of the tangent line of the knife at that point. The midpoints of the line segments were used for positions, and the direction of the line segments were used as tangent lines. Then the same software as used in Chapters 3 and 5 were used to calculate dihedral sharpening angles for every midpoint.

+0

8:1

7)

calibration point, all three sharpeners will have the target dihedral angle.

6.7. CASE STUDIES OF CURVED BLADES ON THE WEPS-GEN1, WEPS-GEN2, AND EP-APEX 75
lie on the main platform and would be foreshortened. Similarly, any circle that approximates the knife edge would appear as an ellipse. Therefore, for the Edge Pro Apex, the silhouette, the circle, and the circle center are drawn incorrectly. However, the relative position between the calibration point and the vertical mast is accurately drawn for the Edge Pro Apex as seen from an orthographic top-down view.

6.7.3 Sharpening a Chefs Knife Without Repositioning and Without ReclampFor an example of a Chefs Knife, we have chosen the Zwilling J.A. Henckels Damascus Chefs Knife, 8. The rst third of the knife near the tip is well approximated by a circle. We choose a target angle of 15 per side.

DR

AF

Figure 6.14: Chefs Knife. Zwilling J.A. Henckels Damascus Chefs Knife, 8. Photo used without permission from http://www.williams-sonoma.com/products/2919595/ .

1.0

be

ta1

7(

20

14

.2. 5

ing

+0

8:1

7)

76

CHAPTER 6. OPTIMAL PIVOT PLACEMENT FOR KNIVES WITH CURVED EDGES

DR

On the left side of the graph, the WEPS-Gen2 shows a decreased dihedral angle. This is understandable given the consequences of the Dihedral Triangle Theorem. Consider a point near X D 1 inches in Figure 6.15. This point is approximately where the knife edge deviates from being a perfect circular arc. From Section 6.5, we know that the WEPS-Gen2 can sharpen a uniform dihedral angle for a knife edge that is a circular arc followed by a straight section that is tangent to the circular arc. In the Chefs Knife, a tangent line at X D 1 inches would continue to the left and above the straighter part of the knife edge. Therefore, the dihedral angle cannot be constant on the Chefs Knife. In addition, a tangent line on the straighter part of the Chefs knife will have a distance to the circle center that is larger than the circles radius. Therefore, the sharpening angle must decrease.

AF

1.0

be

ta1

7(

20

14

.2. 5

The results show in Figure 6.15 show some interesting features. First, as expected, the WEPSGen2 sharpens the circular part of the knife at nearly a constant dihedral angle. This is because we located the spherical joint of the WEPS-Gen2 at the center of the circle. The WEPS-Gen1 shows a downward trend on the right of the graph. This is downward trend is the same effect that was analyzed earlier in Figure 5.17. (See Chapter 5 for details.) As for the EP-Apex, it was not possible to align the sharpener with the center of the circle due to the xed length of the main platform. Consequently, the sharpening angle varies more than for the WEPS-Gen1 and WEPSGen2. (Keep in mind that in practice, we would be allowed to reposition the knife multiple times on the EP-Apex, and this would allow the EP-Apex to sharpen a much more uniform angle.)

+0

8:1

7)

6.7. CASE STUDIES OF CURVED BLADES ON THE WEPS-GEN1, WEPS-GEN2, AND EP-APEX 77
Sharpening Angles for Chef's Knife 8" No repositioning and no re-clampinng of knife.

16.5

EP Apex

16

Dihedral Angle [degrees per side]

15.5

15

14.5

14

13.5

13 5

14
1 0

.2. 5
1 2 3 4 Position of X Coordinate [inches]

be

ta1

7(

20 1.0 T
8 6 4 2

Calibration Point

Circle Center WEPS-Gen1 (midpoint of pivots) WEPS-Gen2 (spherical joint)

EP Apex Vertical Mast (approx)

AF

DR

+0
2

Figure 6.15: Sharpening a chefs knife without repositioning or re-clamping. Positional coordinates are in inches. Angular measurements are in degrees. See Sections 6.7.1 and 6.7.2 for details. The target angle was chosen to be 15 per side at the calibration point.

8:1
WEPS-Gen2 WEPS-Gen1

7)
4

78

CHAPTER 6. OPTIMAL PIVOT PLACEMENT FOR KNIVES WITH CURVED EDGES

ing

The author was asked to analyze a khukuri knife. As suggested by forum members, the target angle was chosen to be 10 per side6 .

DR

We also see that the WEPS-Gen2 is the most uniform over the circular part of the blade. The WEPS-Gen1 shows a slight decrease in sharpening angle on the right of the graph. This is downward trend is the same effect that was analyzed earlier in Figure 5.17. (See Chapter 5 for details.) As for the EP-Apex, it was not possible to align the sharpener with the center of the circle due to the xed length of the main platform. Consequently, the sharpening angle varies more than for the WEPS-Gen1 and WEPS-Gen2. (Keep in mind that in practice, we would be allowed to reposition the knife multiple times on the EP-Apex, and this would allow the EP-Apex to sharpen a much
6

AF

http://www.cliffstamp.com/knives/forum/read.php?5,21981

1.0

Below in Figure 6.17 we see a huge increase in sharpening angle on the left side of the graph. To understand this increase, we can apply the Dihedral Triangle Theorem. From the heel of the cutting edge, we have drawn a tangent line (see Figure 6.17). This tangent line is a distance d from the projected point of the spherical joint in the WEPS-Gen2. Because d is much smaller than the radius of the circle, we know that the sharpening angle is much larger there.

be

ta1

Figure 6.16: Khukuri. Condor Tool & Knife CTK255-10HC Heavy Duty Kukri Knife 10. Photo used without permission from http://www2.knifecenter.com/item/CN25510HC/ Condor-Tool-and-Knife-CTK255-10HC-Heavy-Duty-Kukri-Knife-10-inch-Carbon-Steel-Blade-

7(

20

14

.2. 5

+0

8:1

7)

6.7.4 Sharpening a Khukuri Without Repositioning and Without Reclamp-

6.7. CASE STUDIES OF CURVED BLADES ON THE WEPS-GEN1, WEPS-GEN2, AND EP-APEX 79

DR

AF

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

more uniform angle.)

80

CHAPTER 6. OPTIMAL PIVOT PLACEMENT FOR KNIVES WITH CURVED EDGES

35

30

Dihedral Angle [degrees per side]

25

20

15

.2. 5
5 4 3

10

5 6

14
2 1 0 Pos ition o fXcoo rd ina te [inches ]

20 7(
Ta n ge nt

Calibration Point

ta1

Lin

fro

He

el

d
Circle Center WEPS-Gen1 (Midpoint of Pivots) WEPS-Gen2 (Spherical Joint)

1.0

be

EP Apex Vertical Mast (approx)

AF

12

10

+0
EP-Apex WEPS-Gen2 WEPS-Gen1
1 2 3 4

DR

Figure 6.17: Sharpening a Khukuri without repositioning or re-clamping. Positional coordinates are in inches. Angular measurements are in degrees. See Sections 6.7.1 and 6.7.2 for details. The target angle was chosen to be 10 per side at the calibration point.

8:1
4 6

Sharpening Angles for Khukuri 10" No repositioning and no re-clamping of knife.

7)

6.7. CASE STUDIES OF CURVED BLADES ON THE WEPS-GEN1, WEPS-GEN2, AND EP-APEX 81

ing and Without Reclamping

1.6

80 (approx)

Figure 6.18: Spyderco LionSpy C157GTI. Photo used without permission from http://www. spyderco.com/catalog/details.php?product=720 .

DR

AF

When we look at the results in Figure 6.19 we see an increase in sharpening angle near the tip of the knife. As before, this can be understood in terms of the Dihedral Triangle Theorem and its consequences. Consider a point on the knife edge near X D 0 inches. This is approximately where the knife edge transitions from a straight line into a circular arc. However, the WEPS-Gen2 does not have its spherical joint located at the center of the circle. Instead, it is relatively far away at approximately 4 inches below the knife edge. So if we drew a large circle around the spherical

1.0

be

ta1

3.62

rox) p p 5" (a

7(

rox) p p a 3" (

20

14

.2. 5

For a pocket knife, we choose the Spyderco LionSpy for its unusual shape. See Figure 6.18. The tip of the knife is almost 80 from the main edge. This is an unusually large angle. Furthermore, the belly of the knife has a very tight radius of curvature that is approximately 1.63 inches. It is not reasonable to expect the WEPS-Gen1 or WEPS-Gen2 to be able to place their joints close to the center of this circle. Consequently, the position of the knife and sharpeners were chosen somewhat arbitrarily. For sharpening angle, we choose a target value of 15 per side.

+0

8:1

7)

6.7.5 Sharpening the Spyderco LionSpy Pocket Knife Without Reposition-

82

CHAPTER 6. OPTIMAL PIVOT PLACEMENT FOR KNIVES WITH CURVED EDGES

DR

Interestingly, the WEPS-Gen1 sharpens a more uniform dihedral angle than the WEPS-Gen2. This is due to the downwards effect on sharpening angle of the WEPS-Gen1 as illustrated in Figure 5.17. (See Chapter 5 for details.) In this specic case, the WEPS-Gen1 mechanism has a deviation in sharpening angle that partially cancels out the effect of the Dihedral Triangle Theorem for the WEPS-Gen2. Namely, the WEPS-Gen1 has an artifact in its mechanism that causes its sharpening angle to be smaller than that of the WEPS-Gen2 near the tip of the LionSpy. (We can see a similar effect in Figures 6.15 and 6.17.) In this case, the decrease in angle is actually benecial for creating a more uniform sharpening angle along the length of the blade.

AF

1.0

be

ta1

7(

20

14

.2. 5

+0

joint that reached out to the point at X D 0 inches, it would curve gently above the knife. The LionSpy with its very tight radius of curvature bends sharply into the interior of the large circle. As a result, tangent lines near the knife tip are closer to the spherical joint than the radius of the large circle. Then, from the Dihedral Triangle Theorem, we know that the dihedral angles near the knife tip must be larger. This effect is even greater for the EP-Apex because its vertical mast is even farther away from the LionSpys center of curvature. (Keep in mind that in practice, we would be allowed to reposition the knife multiple times on the EP-Apex, and this would allow the EP-Apex to sharpen a much more uniform angle.)

8:1

7)

6.7. CASE STUDIES OF CURVED BLADES ON THE WEPS-GEN1, WEPS-GEN2, AND EP-APEX 83
30

28

26 Dihedral Angle [degrees per side]

24

22

20

+0
1.5

18

16

14 2

.2. 5
1 0.5 0 0.5 Pos ition o fXcoo rd ina te [inches ] 1

ta1

7(

20
WEPS-Gen1 (Midpoint of Pivots) WEPS-Gen2 (Spherical Joint)

1.0

be

14
1

Calibration Point

DR

AF

EP Apex Vertical Mast (Approx)

Figure 6.19: Sharpening a Spyderco LionSpy without repositioning or re-clamping. Positional


coordinates are in inches. Angular measurements are in degrees. See Sections 6.7.1 and 6.7.2 for details. The target angle was chosen to be 15 per side at the calibration point.

8:1
WEPS-Gen2 WEPS-Gen1
1.5 2

7)
EP Apex

Sharpening Angles for Spyderco Lionspy No repositioning and no re-clamping of knife.

84

CHAPTER 6. OPTIMAL PIVOT PLACEMENT FOR KNIVES WITH CURVED EDGES

DR

AF

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

For readers who wish for more in-depth analysis of sharpening curved blades, please see Appendix A. Appendix A uses animated contour plots to provide a global visualization for optimal pivot placement of the WEPS-Gen2.

7)

6.7.6 Visualizations for Optimal Pivot Placement

B ELT S ANDERS

1.0

be

ta1

7(

20

14

In this chapter, we very briey discuss the use of belt-sanders to sharpen a knife at a constant dihedral angle. It turns out, this is very simple. See Figure 7.1. The vertical plane S is colored blue, and represents the sanding belt. The inclined plane T represents a guide platform where the knife rests. The planes S and T form a dihedral angle where line l is the intersection between the planes. The plane of the knife is the same as plane T . Then when the curved knife edge contacts the sanding belt, it will have line l as a tangent. Consequently, the belt sander is capable of sharpening a perfect dihedral angle, even for knives with curved edges. (Here we are assuming the sanding belt remains planar because it is supported by a at platen from behind.)

.2. 5

+0

DR

AF

Figure 7.1: Belt Sander. The blue plane is vertical and represents the sanding belt. The inclined plane represents a at platform to guide the knife. We assume the plane of the knife is parallel to the platform. Notice that this arrangement can sharpen a uniform dihedral angle on knives with straight and curved edges.

85

8:1
T

7)

C HAPTER 7

86

DR 1.0 be ta1 7( 20 14 .2. 5

AF

CHAPTER 7. BELT SANDERS

+0

8:1

7)

C ONCLUSIONS AND F UTURE W ORK

DR

For future work, it may be interesting to consider guided-rod sharpening mechanisms in the real world, where there is slight exing of parts, as well as looseness and/or imprecision in parts and how they t together. An initial study could be carried out by applying perturbation theory to the idealized models discussed in this document. For example, in addition to analytic perturbation theory, the kinematic software created for our current analysis could be extended to numerically
1

AF

Although, it is unknown to the author if the ball in a spherical rod end is manufactured to the same tolerance as a precision ball bearing.

and at most 0.5 under the most extreme conditions that are still realistic. Such small deviations in angle are unlikely to be noticed in any practical situation encountered by typical users. Generally speaking, they are more than accurate enough.

1.0

That being said, both the Edge Pro and the WEPS-Gen1 are very accurate and repeatable mechanisms. And the deviations from a perfect dihedral edge are quite small; typically less than 0.1 ,

be

Depending on how they are used, spherical joints could have a disadvantage in that they could have more limited ranges of motion that designs based on pivoting joints (revolute joints). If revolute joints are to be used, then we recommend arranging their geometry so that the axes of the joints intersect at a point (which forms a gimbal). In terms of theoretical design, a gimbal is virtually equivalent to a spherical joint. However, using a gimbal is not always necessary, as is shown by the WEPS-Gen1 which can grind perfect straight dihedral angles.

ta1

7(

Spherical bearings can be highly precise (ball bearings typically have extremely accurate sphericity of 25 micro-inches1 (0.000025 inches)), are mechanically simple, and very common (spherical bearings and rod ends are standard mechanical parts). For both the Edge Pro and the WEPS-Gen1, a spherical bearing could replace two revolute joints and also remove any problems with grinding a perfect straight dihedral edge.

20
87

14

We have examined two guided-rod sharpening systems, namely the Edge Pro Apex as well as the Wicked Edge Precision Sharpener (Generation 1). In both cases, we analyzed the behavior of the mechanisms when used to sharpen a tanto knife. Each sharpener has a variety of cases, in some of which they can sharpen a perfect dihedral angle, and in other cases where they cannot. A precise geometric analysis of these two mechanisms has lead us to better understandings how to design high precision sharpening mechanisms. For sharpeners that use a pivoting guide-rod our general recommendation is to use a spherical bearing or a gimbal.

.2. 5

+0

8:1

7)

C HAPTER 8

88

CHAPTER 8. CONCLUSIONS AND FUTURE WORK

DR

AF

First of all, I would like to thank Clay Allison (creator of the WEPS) and Ben Dale (creator of the EP) for making superb sharpening devices. Discussions with Tom Blodgett, Ken Schwartz and others of http://www.knifeforums.com have been great. Feedback and discussions from Mark de Weger have been very helpful. In addition I would like to thank members of #sharpthings for their friendship and support. In particular, Josh and Chuck were very helpful in pinning down the exact geometry of the WEPS-Gen1 and Edge Pro Apex.

1.0

be

8.1 Acknowledgements

ta1

It is sincerely hoped that this document has been of interest to knife sharpeners as well as to anyone who is mechanically minded and/or interested in precision mechanisms.

7(

study small perturbations in geometry and rigidity. Additional effects might include the slight change in sharpening angle for the EP and WEPSGen2, due to the removal of metal from the knife bevel. Each stroke of the sharpening stone slightly changes the geometry of the knife, which could possibly affect the guide rod mechanisms. If the accuracy and repeatability of the mechanism is ner than the changes due to metal removal, then one has reached the point of diminishing returns, and further increases in accuracy would not be meaningful. It would be interesting to determine how much metal is removed per stroke, particularly at very ne grits. (Extremely ne grits can have average particle sizes on the order of microns or smaller.) An even broader consideration may wish to account for costs of manufacturing, wear and reliability of parts, as well as complexity of design. For example, as mentioned in Section 3.2.3, it is possible to replace the pivots in the EP or WEPS-Gen1 with gimbals and/or universal joints (see Figures 3.13 and 3.12). However, these designs are much more complicated (and possibly more expensive) and have many more moving pieces, each of which could be an additional source of play or slack due to the inaccuracy of physical parts. It would be necessary to manufacture and test real mechanisms to see which designs would actually lead to improved accuracy in practice. Many of these additional effects are very interesting, but would require a much more in depth discussion of mechanical engineering and manufacturing.

20

14

.2. 5

+0

8:1

7)

DR 1.0 be ta1
Appendices

AF

T 7( 20 14

89

.2. 5

+0

8:1

7)

DR 1.0 be ta1 7( 20 14 .2. 5

AF

+0

8:1

7)

A.1

Introduction

Suppose we have a chefs knife that we want to sharpen on a WEPS-Gen2. Where should we clamp the knife to minimize the variation in sharpening angle? What we could do, is try lots of different clamping arrangements and see how each one varies the sharpening angle, and then somehow graph or plot all the results. So what does this data look like? To nd out, let us go through an example in full detail. So, say we want to sharpen the chefs knife at 15 per side. We pick a point on the knife edge that we want to be exactly 15 per side; this point is our calibration point on the knife edge. Next, we need to try many different positions for the spherical joint in the WEPS-Gen2. We can specify the position by .x; y/ coordinates where these coordinates are in the plane of the knife. The z -coordinate is perpendicular to the plane of the knife, and it is adjusted until we get 15 per side exactly at our calibration point. Now our knife and WEPS-Gen2 are fully set up. Finally, we get a sharpening angle for each point along the knife edge. Given the above, we have the following:

DR

AF

1.0

be

ta1

A.2

Data to be Visualized

7(

In Chapter 6 we considered the use of guided-rod sharpeners for sharpening knives with curved edges. In particular, in Section 6.7 we examined detailed case studies where the WEPS-Gen1, WEPS-Gen2, and EP-Apex were used to sharpen a chefs knife, a khukuri, and the Spyderco LionSpy. However, these case studies only examined one pivot position for each possible pairing of sharpener and knife. Ideally, one would like to show results for many different pivot positions so that the most desirable one can be chosen. In this Appendix, we demonstrate a method for visualizing the resulting sharpening angles for all (reasonable) pivot positions of the WEPS-Gen2.

20
91

14

.2. 5

+0

V ISUALIZATION OF O PTIMAL P IVOT P LACEMENT FOR THE WEPS-G EN 2

8:1

7)

A PPENDIX A

92APPENDIX A. VISUALIZATION OF OPTIMAL PIVOT PLACEMENT FOR THE WEPS-GEN2 Let x = x -coordinate of the spherical joint. Let y = y -coordinate of the spherical joint. Let xk = x -coordinate of a point on the knife edge. Let f = sharpening angle (degrees per side) at some specic point. So our data looks like this:

A.3

Visualization Techniques

DR

AF

In a contour plot (topographic map), each contour line represents a specic height. See Figure A.1. It is like having an enormous layer cake where each layer is evenly spaced. We then carve away the cake to form our mountains, valleys, and landscape. Each contour line is just a layer of icing. We then view everything from the top. Where the lines are closely spaced, the landscape is very steep (we cross many cake layers in a short distance). Where the lines are very widely spaced, the landscape is at (we have to travel a long way before we get to the next layer). So, if we x the x -coordinate, we get that the sharpening angle is f .y; xk /, which we can plot as a contour map. Figure A.2 is an example for our chefs knife. Dont worry; well explain what this picture means.
1

https://en.wikipedia.org/wiki/Topographic_map

1.0

Suppose we x the x -coordinate of the spherical joint. Then we now have a function f .y; xk /. This would require 3 dimensions to plot. However, we can use just 2 dimensions if we use a contour plot. You may be familiar with contour plots from topographical maps 1 .

be

A.3.1

Contour Plots

ta1

We will try to solve this with two techniques: (1) Contour plots. (2) Animated video so that we can use time as an extra dimension.

7(

20

Now we have a problem: How to visualize f .x; y; xk /? To fully plot this, we need three inputs and one output, which would be four dimensions. Sadly, we only live in 3 spatial dimensions, so we cant do that. In fact, we only have a computer-screen which is 2 dimensions. So how to go from 4 dimensions down to 2?

14

.2. 5

f .x; y; xk / = sharpening angle on the knife edge at point xk , when the spherical pivot is at .x; y/, and z is adjusted to sharpen at 15 per side at our calibration point.

+0

8:1

7)

A.3. VISUALIZATION TECHNIQUES

93

Figure A.1: Topographic Map. In a topographic map, each contour line represents altitude above sea level. Image from www.wikipedia.org.

DR

AF

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

94APPENDIX A. VISUALIZATION OF OPTIMAL PIVOT PLACEMENT FOR THE WEPS-GEN2

Figure A.2: Contour Plot of Sharpening Angles. The red point on the knife edge
represents our calibration point (set to 15 in this example). The vertical black line at x D 2:4 represents the x -coordinate of the spherical joint of the WEPS-Gen2. Each horizontal gray line represents a choice of y -coordinate for the spherical pivot. In addition, each point on a gray line corresponds to a point on the knife edge (namely, the point on the knife directly above on the page). When a gray line crosses a contour line, the sharpening angle has changed by 0.1 per side.

DR

AF

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

A.3. VISUALIZATION TECHNIQUES

95

Please ignore the colors in the contour plot. Im thinking about what a good color scheme should be and learning how to set the colors in Matlab. But for now, Im just using Matlabs default colors, which do not mean anything in this plot. I kept the colors because they are still useful for seeing the direction of contours when they get very dense.

DR

AF

A.3.2

So far, we have a landscape and the horizontal gray lines are our foot paths. And we can walk along the foot-paths and see how many contours we cross to see how the sharpening angle varies. But this landscape is only for a specic value of x , our choice of x -coordinate for the spherical joint! We want to try many different x -coordinates for the spherical joint. This is where we use animation as a visualization technique. We can make many landscapes:

Animation

1.0

be

The landscape we plotted has sea level set to at 15 degrees per side. So the contour labeled 0 means no deviation from our target of 15 per side. The contours labeled 0.5 means we have increased the sharpening angle by 0:5 per side, so we would be at 15 C 0:5 D 15:5 per side. Similarly for the -0.5 contour, and so on.

ta1

Additional Notes

7(

So in our example above, we see lots of widely spaced contours near the heel of the knife. So with our pivot at .x D 2:4; y D 1/, the sharpening angle near the heel is almost constant. However near the tip of the knife, the contours get very close together! So the sharpening angle changes a lot here. So how much does the sharpening angle vary? We can nd out by counting how many contours we cross as we walk along the gray line. Each time we cross a contour line, our sharpening angle (ie: altitude) has changed by 0:1 per side.

20

14

along this gray line, our altitude represents the sharpening angle for the point on the knife with the same x -coordinate (on the page, draw a vertical line until it touches the knife edge).

.2. 5

But how do we read off the sharpening angle? This is where the contour map comes in. Each of the horizontal gray lines represents a foot-path through our landscape. From the point .x D 2:4; y D 1/ in the gure, we can travel horizontally (left or right) along one of these gray lines. Each time we cross a contour, our sharpening angle has changed by 0:1 per side. As we walk

+0

Let us explain all the different parts of this picture. First of all, you can see the silhouette of the chefs knife. The red point on the knife edge is our calibration point: the sharpening angle at this point will always be exactly 15 degrees per side. Suppose we want to try placing our spherical joint at coordinates x D 2:4 and y D 1. So, we rst x x D 2:4 which is represented by the black vertical line in the middle. Next we move along this vertical line until we get to y D 1. This is how we set the .x; y/ position of the spherical joint of the WEPS-Gen2.

8:1

7)

96APPENDIX A. VISUALIZATION OF OPTIMAL PIVOT PLACEMENT FOR THE WEPS-GEN2

A.3.3

Example

DR

AF

Next, the red point is on a horizontal gray line. We can walk left-and-right along the gray line.

Figure A.3: Example Contour Plot of Sharpening Angles. To place the spherical joint

at the red dot below the knife, we rst go to the animated frame where the moving vertical like goes through the red dot. To nd the variation in sharpening angle, we travel left and right on the gray line that goes through the red dot. Each contour line represents a change in sharpening angle of 0.1 per side.

1.0

be

ta1

7(

20

14

.2. 5

+0

Let us work through a specic example. Consider the red dot marked in the landscape of Figure A.2. Suppose we want to place our spherical joint there. We go to the frame of the animation where the vertical black line goes through that point. Here is that frame. (See Figure A.3.)

8:1

7)

one for each position of x -coordinate for the spherical joint. Each animated frame has a vertical black line (the one that is moving) that represents the x -coordinate of the spherical pivot.

A.4. EXAMPLE VISUALIZATIONS FOR OPTIMAL PIVOT PLACEMENT OF THE WEPS-GEN2 97

WEPS-Gen2

DR

AF

1.0

If you understood all that, congrats! Sorry if it is so complicated. Im unsatised with this visualization, but it is the best I can come up with for now. Here are the animated videos of the contours. The red points marked in the contour plots are the same pivot locations as used in Section 6.7 and correspond to Figure 6.15, Figure 6.17, and Figure 6.19.

be

ta1

A.4

Example Visualizations for Optimal Pivot Placement of the

7(

So what are we looking for? We want to search all the frames for a horizontal gray line which crosses as few contours as possible, and which is also the closest tosea level as possible. Once we nd such a gray line, we nd its intersection with the moving vertical black line. Our optimal pivot placement is at intersection of this horizontal gray line and the vertical black line.

20

14

In this example, we have placed the spherical joint at the position of the red dot. When we do this, the sharpening angle near the tip of the knife is almost constant. That is, as we walk to the right along the gray line, we cross very few contour lines. We cross one, maybe two lines, which means a change of 0:1 (maybe 0:2 ) per side. However, near the heel of the knife on the left, we cross many contour lines. From the plot, we can see that the sharpening angle decreases as we cross 7 contours. So our sharpening angle decreases by 0:7 per side. Finally, notice the vertical contour below the calibration point. Of course this must be there! This is because we are setting the WEPS-Gen2 to sharpen at 15 per side at the calibration point, for every choice of (x,y) position of the spherical joint. So we will always have a vertical contour line below the calibration point, and it will have an altitude of zero degrees per side. That means, zero degrees per side deviation from our target angle (which is 15 per side).

.2. 5

+0

8:1

7)

Each time we cross a contour, our sharpening angle has changed by 0:1 per side.

98APPENDIX A. VISUALIZATION OF OPTIMAL PIVOT PLACEMENT FOR THE WEPS-GEN2

DR

AF

1.0

Figure A.4: Visualization for a Chefs Knife on the WEPS-Gen2. Coordinates are in inches. Target sharpening angle = 15 per side at the calibration point. Contour lines every 0.1 per side.

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

A.4.1

A Chefs Knife on the WEPS-Gen2

A.4. EXAMPLE VISUALIZATIONS FOR OPTIMAL PIVOT PLACEMENT OF THE WEPS-GEN2 99

DR

AF

Visualization for a Khukuri on the WEPS-Gen2. Coordinates are in inches. Target sharpening angle = 10 per side at the calibration point. Contour lines every 0.1 per side. The contour plot goes a bit crazy in the upper left corner. Please ignore these artifacts; these are caused by my software which treats +90 degrees as the same as -90 degrees. So when the sharpening angle goes to 90 deg per side, it can rapidly ip between +90 and -90 in the plot, which causes Matlab to draw crazy contours.

Figure A.5:

1.0

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

A.4.2

A Khukuri on the WEPS-Gen2

100APPENDIX A. VISUALIZATION OF OPTIMAL PIVOT PLACEMENT FOR THE WEPS-GEN2

Figure A.6: Visualization for the Spyderco LionSpy on the WEPS-Gen2. Coordi-

DR

AF

1.0

nates are in inches. Target sharpening angle = 10 per side at the calibration point. Contour lines every 0.1 per side.

be

ta1

7(

20

14

.2. 5

+0

8:1

7)

A.4.3

The Spyderco LionSpy on the WEPS-Gen2