Documente Academic
Documente Profesional
Documente Cultură
opez-Martınez
Department of Engineering,
University of Almerıa,
Almerıa 04120, Spain
e-mail: javier.lopez@ual.es
Design of Three New Cam-Based
Daniel Garcıa-Vallejo
Department of Mechanical Engineering
Constant-Force Mechanisms
and Manufacturing, Constant-force mechanisms are designed to keep a constant or nearly constant input
University of Seville, force along a prescribed stroke of the mechanism. The implementation of this kind of
Seville 41092, Spain mechanisms has been approached in literature using compliant mechanisms or through a
e-mail: dgvallejo@us.es certain combination of springs and nonlinear transmissions. In this work, three new
constant-force mechanisms based on the use of springs, rollers, and cams are presented
Francisco Manuel Arrabal- and analyzed. The rolling friction forces between the rollers and the cam are included in
Campos the force equilibrium equations and considered in the integration of the cam profile. The
Department of Engineering, influence of the friction force on the input force as well as the design parameters involved
University of Almerıa, is studied based on numerical techniques and simulations. In fact, the results evidence
Almerıa 04120, Spain that to obtain a precise constant-force mechanism, rolling friction forces must be consid-
e-mail: fmarrabal@ual.es ered in the cam profile definition. The main design guidelines for the three constant-force
mechanisms proposed are described. [DOI: 10.1115/1.4040174]
Jose Manuel Garcia-Manrique
Department of Civil Engineering, Keywords: constant-force mechanism, cam profile, rolling friction
Materials and Manufacturing,
University of Malaga,
Malaga 29071, Spain
e-mail: josegmo@uma.es
Fig. 1 Sketches of the three constant-force mechanisms proposed: (a) CFM-1, (b) CFM-2, and (c) CFM-3
of springs arranged in parallel (Fig. 2 shows four spring in parallel along the surfaces, only rolling friction occurs [40]. These friction
but a larger number is possible). forces can be expressed as the product of the coefficient of the
rolling friction lr and the normal force at the contact
3 Constant-Force Condition Equations
Ff ¼ lr N (1)
From the operating principle of the CFMs described in Sec. 2,
one can define the shape of the cam in order to achieve a constant From the force diagram at Fig. 3, the equilibrium equations in
force at the input link. Based on a quasi-static force analysis of the horizontal and vertical directions can be written as
the mechanisms, the mathematical conditions that must be ful-
filled will be used to define the cam profiles. In the following, it is FK þ Ff 1 þ Ff 2 sin b ¼ N2 cos b (2)
assumed that inertial forces due to the moving masses are negligi-
ble. This assumption is accurate enough for low accelerations and N1 ¼ N2 sin b þ Ff 2 cos b (3)
in static position. Also, when the weights of the springs and the
rollers are small compared with the input force, as will be usual, where b is the angle between the tangent to the imaginary curve
their contribution in the dynamics of the mechanism will be negli- that describes the axis of the roller and the vertical axis. The slope
gible even for higher acceleration. of this curve is defined as
F
N1 ¼ (7)
2
Finally, substituting Eq.(7) in Eq. (6), the input force F is
expressed as a function of the stiffness coefficient K, the spring
elongation Dx, the coefficients of rolling friction, lr1 and lr2 , and
the slope of the curve
1
1 lr2 tan b
F ¼ 4K Dx lr1 (8)
tan b þ lr2
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl
ffl}
Fig. 3 Forces diagram for the rollers assembly of CFM-1 A
Finally, when friction is ignored, the force equation reduces to When friction is ignored, the Eq. (16) simplifies to
0
F ¼ 4KDx tan b (11) 1
1
F0 ¼ 4KDx tan a (18)
3.2 Analysis of Constant-Force Mechanism-2. In a similar tan b
way, the equilibrium equations and the constant-force condition
are obtained for CFM-2. The free-body diagram of the roller is
shown in Fig. 4. The equilibrium equations in the horizontal and 3.3 Analysis of Constant-Force Mechanism-3. Figure 5
vertical directions can be written as shows the forces acting on the roller for CFM-3. The equilibrium
equations in the radial and circumferential directions are written
FK þ Ft sin a þ Ff sin b ¼ N cos b (12) as
Fig. 4 Forces diagram for the roller of CFM-2 Fig. 5 Forces diagram for the roller of CFM-3
Finally, simplifying Eq. (23) the input force F for CFM-3 is where tan b in Eq. (9) has been substituted by dx=dy in agreement
defined as follows: with the definition of axes in Fig. 3. The following nonlinear first-
order differential equation can be obtained from the previous one
F ¼ K Drðsin b þ lr cos bÞ (24)
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} as follows:
C
dx 1 lr2 f ð xÞ
The condition for a constant-force mechanism requires that the ¼ (30)
dy lr2 þ f ð xÞ
term in parenthesis be constant
where
C ¼ Drðsin b þ lr cos bÞ constant (25)
1
f ð xÞ ¼ ð x x0 Þ þ lr1 (31)
Finally, if friction is neglected, Eq. (24) can be simplified to A
2 !12
1 x x
f ð xÞ ¼ ð x x0 Þ þ 1 (37)
B lt lt
1 lr f ðxk Þ
xk xk1 h ¼0 (38)
lr þ f ðxk Þ
0 1
dx
B 1 l r C
B dy C
Dx ¼ BB tan aC (34)
@ dx A
þ lr
dy
2 !12
x x
tan a ¼ 1 (35)
lt lt
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi C
1 cos2 bk þ lr cos bk ¼ 0;
rk r0
rk
cos bk ¼ sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 2 (45)
rk rk1 rk rk1
rk sin hk þ cos hk þ rk cos hk þ sin hk
hh hh
The previous nonlinear problem is solved for an initial value of The test is not shown again for not to unnecessarily increase the
r (rðh0 Þ ¼ R0 with h0 ¼ 0) by using an iterative algorithm for text length.
solving nonlinear equations. Note that R0 is different to r0,
meaning that the spring may be elongated at h ¼ 0. As previ- 5 Influence of the Rolling Friction and Design
ously mentioned, Newton–Raphson method [41] is used to find
the trajectory of the roller center in terms of cylindrical coordi- Considerations
nates, r and h. Since rolling friction affects the value of the force at the input
Figure 10 shows different trajectories of the roller center link, it is important to study its influence in the response of the
obtained for a certain combination of target force, F, and spring mechanism concerning the constant-force condition. In a first
properties, see figure caption, and different values of the rolling approach to the design problem, one can define the trajectory of
friction coefficient. In the figure, it has been shown in the same the roller center by two different ways. In the first case, the trajec-
color curves corresponding to the same rolling friction value, but tory of the roller center is derived ignoring friction effects. As far
with opposite direction of the motion. For each pair of curves in as authors know, this is the procedure used in all previous works.
the same color, the one with the smallest radius corresponds to the Liu et al. [26,27], for a friction analysis, once designed the cam
rotation of the roller in counterclockwise direction according to profile without friction consideration, studied the deviation of the
Fig. 5. Again, a nonrealistic excessively large value of the rolling force applied to the input link due to the friction force. In the
friction (lr ¼ 0:02) has been included to remark the different second case, the cam profile can be designed accounting for
shapes required for the cam depending on the motion direction. the friction forces as shown in Sec. 4. This strategy is useful when
A convergence test similar to the one shown in Fig. 7 was per- the requirement of constant-force condition is preferable only in
formed to select an appropriate value of the integration step size. one movement direction of the mechanism stroke. Note that since
the friction force is opposed to the movement of the roller, as pre- the center of the roller for a desired constant force F of 30 N when
viously showed, the cam profile that meets constant-force condi- the rollers move in positive direction of y-axis (with a rolling
tions will be different for the loading movement than for the friction coefficient of 0.002) and for different spring stiffness
unloading movement of the mechanism. Also, to obtain a precise coefficients K; while Fig. 12(b) shows the corresponding forces
cam profile, the coefficient of rolling friction must be known. ratio F0 =F for coefficients of rolling friction lr1 ¼ lr2 ¼ 0:002
Since such rolling friction coefficient is difficult to be known (this value of rolling friction may correspond to a hard material as
accurately, and may change over time due to wear or changes in hardened steel [40]). According to Figs. 12(a) and 12(b), to reduce
operational conditions, the study of the influence of the coefficient the influence of the rolling friction, given a target constant force,
of rolling friction in the performance of the mechanism is of inter- it is preferable to use springs with small stiffness values. This
est. In this section, the influence of the friction forces and the leads to cam profiles with small slope values, see the cam profile
design parameters in the performance of the three mechanisms depicted in black in Fig. 12(a). The choice of the spring-cam pair
proposed is analyzed. is a critical factor regarding the friction effects, especially when
For CFM-1, from Eqs. (8) and (11), the ratio of the force with- the mechanism requires a long stroke. Although it is not clearly
out friction F0 and the force with friction F results a function of appreciable in Fig. 12(b), the ratio F0 =F has a maximum for values
the slope of the cam profile and of the coefficients of rolling fric- of y around 1 mm; for smaller values in the y-axis, the ratio F0 =F
tion as decreases rapidly, as can also be understood from Fig. 11 when
the angle b approximate to p/2.
F0 1 lr2 tan b Other factors must be taken into account in the design process,
¼ tan b lr1 (46) as space limitations that may limit the cam width, or the admissi-
F tan b þ lr2
ble elongation of the spring. The spring elongation is directly
The ratio F0 =F as a function of the angle b is plotted in Fig. 11 related with the stroke of the mechanism. Thus, when the largest
(continuous lines) for different values of the coefficient of rolling possible stroke is required, one must select springs with values of
friction. Note that these curves are independent of any other stiffness coefficient K and maximum admissible elongation 2Dx
design parameter of the mechanism. Values of angle b close to that allows the maximum stroke of the mechanism (coordinate y
zero or to p/2 lead to a higher influence of the friction forces. Dis- in Fig. 12(a)). Spring elongation may be a limitation for the use of
continuous lines represent the ratio F0 =F when the rollers move in these mechanisms in applications where a large stroke is required.
the opposite direction (negative direction of y-axis in Fig. 3) and A design solution to increase the stroke of the mechanism is to
then Eq. (10) instead of Eq. (8) is used. Figure 11 is useful to use, instead of two springs, a larger number of softer springs
know the deviation in the target force F0 , for an estimated rolling arranged in parallel, with the same resultant force (see Fig. 2(b)).
friction coefficient, when a cam profile without friction considera- In this way, since softer springs have larger allowable maximum
tions is designed. In the same way, similar curves can be obtained elongation, the stroke of the mechanism can be increased. The
to estimate the variations in the force value, due to changes in the possibility to add and remove springs in parallel allows to modify
friction coefficient value, when a cam profile is designed for a the resultant equivalent stiffness K, giving the mechanism the
given rolling friction coefficient. This study is of interest when the ability to easily change its input force F, as can be seen from
working conditions of the mechanism are subject to significant Eq. (8).
changes in the coefficient of rolling friction. In any case, it is Next, in a similar way, from Eqs. (16) and (18), the ratio of the
recommended to use materials for the cam and the rollers with a force without friction F0 and the force with friction F for CFM-2
rolling friction coefficient as low as possible. results in
As described in Sec. 3, for a required constant force value and a 1
given spring stiffness constant, there exists a cam profile that F0 1 lr tan b 1
¼ tan a tan a (47)
meets the constant force condition. Then, in the design process, a F tan b þ lr tan b
pair of cam profile and spring stiffness must be selected. To visu-
alize the influence of the friction depending on the cam profile, where, in addition to the slope of the cam profile and the
Fig. 12(a) shows in solid lines the trajectory that must describe coefficient of rolling friction, the angle a between the brace and
the y-axis appears. In this case, to obtain the ratio F0 =F, some geo- seen that the influence of the friction forces is somewhat lower in
metric parameters of the mechanism must be known. For a brace CFM-1. The differences between CFM-1 and CFM-2 curves will
length lt ¼ 250 mm, a spring rest length 2x0 ¼ 180 mm and an ini- decrease for larger brace lengths of CFM-2; also the trajectories
tial condition X0 ¼ 90 mm, Fig. 12 shows in dashed lines the of the roller center of CFM-2 in Fig. 12(a) will approximate to
required trajectory of the roller center for a desired constant force CFM-1, as was shown in Fig. 9.
F of 30 N when the rollers move in positive direction of y-axis Finally, for CFM-3, from Eqs. (24) and (26), the ratio of the
(with a rolling friction coefficient of 0.002) and the corresponding force without friction F0 and the force with friction F results as
forces ratio F0 =F. Similarly to CFM-1, for softer springs, a wider
cam profile is preferred to decrease influence of the friction force. F0 1
¼ (48)
Also, space considerations and the required stroke for the F 1 þ lr
mechanism are limitations for the cam width and spring selection tan b
(stiffness coefficient and admissible elongation). In addition, for
CFM-2, the brace length lt must be selected, where the angle a in where the ratio F0 =F depends on the angle b and on the coefficient
Eq. (47) depends on the brace length lt (see Fig. 4). Figure 13 of rolling friction lr.
shows the effects of the brace length in the ratio F0 =F. For smaller The ratio F0 =F in function of the angle b is plotted in Fig. 14
values of lt, the angle a will be larger and the tension force at the for different coefficients of rolling friction. The friction force
brace increases (see Eq. (15)); this results in higher values of the increases with lower values of the angle b, i.e., when the spring
normal force and therefore of the friction forces. Then, a large elongation, and therefore the normal force between the roller and
brace length is preferable instead a shorter one. Comparing the
forces ratio F0 =F for CFM-1 and CFM-2 in Fig. 12(b), it can be
the cam, increases. Note that these curves of F0 =F versus b are force, spring stiffness, spring natural length, and brace length for
independent of the size of the cam and of any other design param- CFM-2, are equal to those used in Sec. 4. Additionally, rollers
eter of the mechanism. It is convenient to remember here that the with a diameter of 30 mm have been used. The roller center trajec-
force F in Eq. (48) represents the force acting on the roller axis tory used in the construction of the cam profile was obtained
instead the force at the final end of the roller chain, i.e., the fric- according to the method described in Sec. 4, where an upward dis-
tion between the roller chain and the cam has not been included. placement of the rollers (positive direction of the y-axis) and a
This means that the ratio F0 =F, in case that F is the force at the rolling friction coefficient of 0.002 were considered in the condi-
end of the roller chain, will differ from the curves shown in tion of constant force. Figure 16 shows the trajectory described by
Fig. 14, mainly for angles b close to p/2, where the ratio F0 =F the roller axis and its corresponding cam profile. Due to the proce-
would be less than one. dure used to obtain the cam profile from the roller center trajec-
Figure 15(a) shows the trajectory of the center of the roller for tory, it may happen to find arcs of the cam profile that are locally
a desired constant force F of 30 N (with a rolling friction coeffi- concave. In particular, this happens for the arc of roller center tra-
cient of 0.002) and for different spring stiffness coefficients K, jectory between points 0 and 1 in Fig. 16. Since such concave arcs
while Fig. 15(b) shows the corresponding forces ratio F0 =F along of the cam profile are not usable, the simulation has been per-
the traveled arc length when the spring rotates counterclockwise formed for fully convex arcs of cam profile. Therefore, an initial
for a coefficient of rolling friction lr ¼ 0.002. The traveled arc coordinate y1 different to zero has been defined for each simula-
length, Dl, has been numerically integrated by using the following tion. Table 1 summarizes the minimum and maximum values of
formula: the roller axis coordinate y (y1 and y2) for an input displacement
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi (stroke) of 100 mm, and its corresponding angles b (b1 and b2).
ð lf ð h f 2 For CFM-3, values of minimum and maximum angles h are given
dr
Dl ¼ dl ¼ þ ðr Þ2 dh (49) as more representative values instead of the coordinate y.
l0 h0 dh In the simulations, the contacts between the solids have been
defined as no penetration but separation allowed. A value of fric-
where the polar coordinates have been used for simplicity tion coefficient lr of 0.002 has been used for the contact between
(x ¼ r cos h; y ¼ r sin h). Figure 15(b) shows larger values of the the rollers and the cams. The same value has been used for the
force ratio than those in Fig. 12(b) for the same spring stiffnesses. friction lr2 between the roller and the horizontal bar in CFM-1.
As for CFM-1 and CFM-2, a soft spring is preferable in order to Additionally, deformation of the solids due to the contact forces is
minimize the influence of the friction and to increase the stroke of calculated during the simulations, where a contact stiffness of
the mechanism. Again, the use of soft springs leads to small slope 100 GPa is used to approximate the material stiffness at the
cam profiles. boundary of interaction between the two solids in contact
elastic contact behavior was not taken into account in the develop-
ments of Sec. 3. Figure 18 shows, for CFM-1, the force results for
different contact stiffness values, where larger differences with
respect to the desired force are drawn for softer materials. These
results show that, in addition to their beneficial lower rolling fric-
tion coefficient, hard materials such as steel are also preferable
due to their lower deformation during contact. It should be pointed
out that, although the hysteresis loop for CFM-3 in Fig. 17 is
smaller than for CFM-1 and CFM-2, this force curve for CFM-3
has been obtained without friction forces in the contact between
the roller chain and the cam. The influence of the roller chain fric-
Fig. 16 Roller axis trajectory and its equidistant curve that tion will depend on the length where the roller chain is in contact
define the cam profile with the cam, which varies with the displacement.
larger for the cusps located near to the top of the cam due to the Latter, a comparison with existing designs is summarized at the
higher spring force. Figure 19(b) shows curves for two cam pro- second part of this section.
files with the same desired constant force but for two different
spring stiffness constant. These curves show that the effect of the 7.1 Comparison of the Three Constant-Force Mechanisms.
cusps decreases for small slope cam profiles (i.e., small spring In spite of the similarity of the three CFMs proposed, where all of
stiffness). This is attributed to the fact that the increase in the them are based on the use of a cam profile, springs, and rollers,
spring force due to the cusps is smaller in soft spring than in stiffer Table 2 resumes the main differences between them. None of the
ones. three design solutions presents special mechanical complexity,
Finally, another manufacturing or assembly error has been where CFM-2 stands out for its simplicity and reduced number of
simulated. Figure 20 shows the force hysteresis loops for different components. As pointed in Sec. 2, simple modification of the
deviations of the spring rest length (from 1 to 1 mm) from its mechanisms can be done to allow CFM-1 and CFM-2 to work
theoretical value. These simulations may reproduce the effect of a with constant compression forces. On the contrary, there is no
manufacturing or assembly error that modifies the distance direct modification of the CFM-3 that allows it to work in this
between the rollers, or an error in the spring rest length estimation way.
for preloaded extension springs. For a deviation of 1 mm, results For all of the designs proposed, the stroke of the mechanism is
show a variation of 3 N (10%) in the force at the beginning of the limited by the admissible spring elongation. Small differences
displacement, while this difference is reduced to 1 N at the end. have been found between the three CFMs. CFM-1 and CFM-2
Then, a precise assembly without gaps and an accurate value of show almost the same maximum displacement for the same
the spring rest length is required for a good response of the spring, while CFM-3 reaches a slightly greater stroke.
mechanism. An important aspect of the performance of the mechanisms is
the influence of the rolling friction and the force hysteresis loop.
7 Performance Comparison As showed in Fig. 17, CFM-1 and CFM-2 hysteresis loops are
quite similar for a brace length lt of 250 mm. Short brace lengths
Based on the results of Secs. 5 and 6, a performance compari-
of CFM-2 should be avoided so as not to increase the friction
son between the three constant force mechanisms is discussed.
effects. Hysteresis loop increases with the rolling friction coeffi-
cient and the stroke of the mechanisms. Regarding CFM-3, the
friction forces and hysteresis loop will depend on the length of the
roller chain in contact with the cam.
Common characteristics of the three CFMs for a better design
are the use of soft springs together with small slope cam profiles
to reduce the friction force influence and the effects of occasional
defects, and the use of stiff and low rolling friction coefficient
materials, as hardened steel, for the rollers and cams.
that our three designs belong to) is that can be designed for large Funding Data
stroke, also, in general, the mathematical model and structure is
simpler than in compliant mechanisms. Compliant constant force This work has been partially funded by the Spanish
mechanisms are usually of compact size and do not have rolling “Ministerio de Economıa y Competitividad” under the pro-
or sliding components, avoiding the problem of friction and ject DEP2016-80296-R (AEI/FEDER, UE).
backlash.
Inside the group of conventional rigid-link CFMs, five kinds of
CFMs are differentiated [2]. In this case, our three CFMs must be References
included in that kind of CFMs named as “Curved surface constant [1] Wall, A., 1963, Mechanical Springs, 2nd ed., McGraw-Hill, New York.
force mechanism.” Next, a comparison between our CFMs pro- [2] Wang, P., and Xu, Q., 2018, “Design and Modeling of Constant-Force Mecha-
nisms: A Survey,” Mech. Mach. Theory, 119, pp. 1–21.
posed and the existing ones of the same group is discussed, where [3] Bidgoly, H., Ahmadabadi, M., and Zakerzadeh, M., 2016, “Design and
two devices, designed by Liu et al. [26,27], have been found in Modeling of a Compact Rotational Nonlinear Spring,” IEEE International Con-
the literature. In the same way to our three CFMs, they are actu- ference on Intelligent Robots and Systems (IROS), Daejeon, South Korea, Oct.
ated with a linear force, and comprise rollers, springs, and a cam. 9–14, pp. 4356–4361.
[4] Howell, L., and Midha, A., 1995, “Parametric Deflection Approximations for
However, the design solutions lead to some performance differen- End-Loaded, Large-Deflection Beams in Compliant Mechanisms,” ASME J.
ces. The main observed advantages of the CFMs proposed in this Mech. Des., 117(1), pp. 156–165.
paper are listed as follows: (i) simpler design, including a minor [5] Howell, L., 2001, Comliant Mechanisms, Wiley, New York.
number of components (specially in CFM-2); (ii) the cam profile [6] Gallego, J., and Herder, J., 2010, “Classification for Literature on Compliant
Mechanisms: A Design Methodology Based Approach,” ASME Paper No.
has been derived including the friction effects, which improve the DETC2009-87334.
constant force condition; (iii) the use of traction springs, instead [7] Tolman, K., Merriam, E., and Howell, L., 2016, “Compliant Constant-Force
of compression springs, avoid buckling problems and may Linear-Motion Mechanism,” Mech. Mach. Theory, 106, pp. 68–79.
increase the allowable stroke; (iv) it is easy to add or remove [8] Lamers, A., Gallego Snchez, J., and Herder, J., 2015, “Design of a Statically
Balanced Fully Compliant Grasper,” Mech. Mach. Theory, 92, pp. 230–239.
springs arranged in parallel to modify the force value, or to [9] Boyle, C., Howell, L., Magleby, S., and Evans, M., 2003, “Dynamic Modeling
increase the stroke using several soft springs with the same total of Compliant Constant-Force Compression Mechanisms,” Mech. Mach.
stiffness. Theory, 38(12), pp. 1469–1487.
[10] Pham, H.-T., and Wang, D.-A., 2011, “A Constant-Force Bistable Mechanism
for Force Regulation and Overload Protection,” Mech. Mach. Theory, 46(7),
8 Conclusions pp. 899–909.
[11] Meaders, J., and Mattson, C., 2010, “Optimization of Near-Constant Force
Three new constant-force mechanisms have been presented in Springs Subject to Mating Uncertainty,” Struct. Multidiscip. Optim., 41(1), pp.
this paper. All CFMs are based on the use of springs, cams, and 1–15.
roller, but each one with a different configuration, what leads to [12] Prakashah, H., and Zhou, H., 2016, “Synthesis of Constant Torque Compliant
Mechanisms,” ASME J. Mech. Rob., 8(6), p. 064503.
certain performance differences. First, from a quasi-static force [13] Pedersen, C., Fleck, N., and Ananthasuresh, G., 2006, “Design of a Compliant
analysis the mathematical conditions that must comply the cams Mechanism to Modify an Actuator Characteristic to Deliver a Constant Output
profiles have been obtained. Latter, the shape of the cams profiles Force,” ASME J. Mech. Des., 128(5), pp. 1101–1112.
have been obtained by numerical integration, where the rolling [14] Chen, Y.-H., and Lan, C.-C., 2012, “An Adjustable Constant-Force Mechanism
for Adaptive End-Effector Operations,” ASME J. Mech. Des., 134(3), p.
friction has been considered. It is shownd that different cam pro- 031005.
files are required depending on the direction of motion of the [15] Liu, Y., Zhang, Y., and Xu, Q., 2017, “Design and Control of a Novel Compli-
mechanism to ensure a constant force. When the mechanism is ant Constant-Force Gripper Based on Buckled Fixed-Guided Beams,” IEEE/
intended to work in both directions, an intermediate profile may ASME Trans. Mechatronics, 22(1), pp. 476–486.
[16] Wang, P., and Xu, Q., 2017, “Design of a Flexure-Based Constant-Force Xy
be recommended. The mathematical models of the CFMs have Precision Positioning Stage,” Mech. Mach. Theory, 108, pp. 1–13.
been validated with the help of computational simulation in well- [17] Xu, Q., 2017, “Design of a Large-Stroke Bistable Mechanism for the Application
known commercial software. A performance study of the mecha- in Constant-Force Micropositioning Stage,” ASME J. Mech. Rob., 9(1), p. 011006.
nisms shows that the rolling friction influence increases with the [18] Qiu, J., Lang, J., and Slocum, A., 2004, “A Curved-Beam Bistable Mecha-
nism,” J. Microelectromech. Syst., 13(2), pp. 137–146.
slope of the cam profile, with the stroke and with the spring stiff- [19] Chen, G., Gou, Y., and Zhang, A., 2011, “Synthesis of Compliant Multistable
ness. Then, for a required input force, it is preferable the use of Mechanisms Through Use of a Single Bistable Mechanism,” ASME J. Mech.
soft springs, that leads to small slope cam profiles. The use of soft Des., 133(8), p. 081007.
springs is also advantageous in the presence of occasional small [20] Starostin, E., 1987, “Calculating a Cam Profile for a Constant-Force Mecha-
nism,” Sov. Mach. Sci., 4, pp. 69–76.
manufacturing errors as cusps or any other kind of irregularities in [21] Duval, E., 2010, “Dual Pulley Constant Force Mechanism,” U.S. Patent No.
the cam surface. A comparative study of the three proposed mech- 7,677,540.
anisms has been included. [22] Riley, R., and Carey, D., 1980, “Exercise Machine With Spring-Cam Arrange-
ment for Equalizing the Force Required Through the Exercise Stroke,” U.S.
Patent No. 4,231,568.
Acknowledgment [23] Schepelmann, A., Geberth, K., and Geyer, H., 2014, “Compact Nonlinear
Springs With User Defined Torque-Deflection Profiles for Series Elastic
The authors would like to thank Celia Nun~ez Torres for her Actuators,” IEEE International Conference on Robotics and Automation
collaboration in the 3D design of the mechanisms. (ICRA), Hong Kong, China, May 31–June 7, pp. 3411–3416.