Available online at www.sciencedirect.

com

ScienceDirect
Procedia CIRP 21 (2014) 367 372

24th CIRP Design Conference

Analysis of vibration damping in machine tools

Giacomo Bianchia*, Stefano Cagnab, Nicola Caua, Francesco Paoluccia
a

Istituto di Tecnologie Industriali e Automazione - CNR, Milano, 20133, Italy

b
Universit degli Studi di Bergamo, 24127, Italy

* Corresponding author. Tel.: +39-0223699911; fax: +39-0223699941. E-mail address: g.bianchi@itia.cnr.it

Abstract

The dynamic behavior of a machine tool structure directly influences key metal cutting performance like being able to quickly
remove hard workpiece material during roughing or minimize unwanted oscillations during high speed movements in finishing.
While structure conception is still funded on designer experience and inventiveness, Finite Element models are very effective in
analyzing the conceived structure, allowing its optimization, in term of stiffness increase and/or mass reduction.
While today FE models provide a satisfying description of structure distributed stiffness and inertia, machine damping is usually
not represented or is approximated as a uniform viscous damping, with no precise reference to the actual dissipation phenomena
occurring in the structure. The corresponding incertitude in the estimation of the overall dynamic behavior often strongly limits
the possibility of delivering accurate absolute estimations of machine performance. In order to overcome this limitation, this
work aims at adding key energy dissipation mechanisms into numerical structural models: the velocity loop of the axis position
controller, the frictional forces acting on the axis kinematic chain and guide ways and a distributed modal damping. Experimental
tests have been performed on a machine tool axis equipped with tunable roller plus plain friction guide ways. The proposed
model shows how different components and phenomena contribute into increasing machine performance, in term of material
removal capacity. Given that the resulting models are essentially non-linear, appropriate methodologies are also suggested to
integrate the proposed analysis into the usual machine development design cycle.
2014
2014 Published
The Authors.
Published
by This
Elsevier

by Elsevier
B.V.
is anB.V.
open access article under the CC BY-NC-ND license
Selection and peer-review under responsibility of the International Scientific Committee of 24th CIRP Design Conference in the person of
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
the Conference
Chairs Giovanni
Moroni and Tullio
Tolio.
Selection
and peer-review
under responsibility
of the International
Scientific Committee of 24th CIRP Design Conference in the person of the
Conference Chairs Giovanni Moroni and Tullio Tolio
Keywords: machine tools, damping, friction

1. Introduction
Several key machine tool performance, like the accuracy
while running at high speed along complex trajectories and the
capability of removing a large volume of workpiece material
in a short time depend on machine dynamic behavior. A
machine tool in operation is a complex system, composed by
the mechanical frame, electrical motors, sensors, control and
the cutting process. In order to support an effective and
optimal design, in the recent years numerical models have
been more and more used, moving toward a Virtual
Prototyping approach. A well know modeling methodology
is based on a Finite Elements representation (FEM) of the
mechanical frame: such models, widely diffused in industry
today, allow a sophisticated and effective design of complex

structures, equipped with 3D net of stiffening elements. A

standard FEM produces an accurate representation of
distributed stiffness and inertia properties but usually no
estimation of damping, which typically has a great influence
on dynamic performance, starting from detailed material
physical properties: an overall damping has to be defined
directly by the analyst, basing on his/her experience and
experimental tests performed on similar machines. Because of
those limitations, FEM are typically used, in industrial
contexts, to compare alternative design solutions, under the
same hypotheses for structural damping. Such methodologies
are well suited for incremental design, e.g. comparing new
solutions to previous machine design.
A more general and effective goal, especially in a scenario
of custom made machines produced in small series, would be

2212-8271 2014 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Selection and peer-review under responsibility of the International Scientific Committee of 24th CIRP Design Conference in the person of the
Conference Chairs Giovanni Moroni and Tullio Tolio
doi:10.1016/j.procir.2014.03.158

368

Giacomo Bianchi et al. / Procedia CIRP 21 (2014) 367 372

to evaluate the proposed design against customer required

performance. In this context, numerical models should provide
sufficient absolute accuracy, in respect to experimental results.
A more detailed description of energy dissipation is then
required: for this goal, friction in moving contacts and position
control should be taken into consideration. Several works have
analyzed the machine as a mechatronic system [1] and a large
literature study friction, mainly focused on its effect on motion
control, also proposing model based compensation strategies
[3]. This work instead analyzes how energy dissipation
influences the dynamic compliance of the structure at points
where process forces are applied, i.e. at the work piece or tool.
Chapter 2 describes the reference case study and its
specification, Chapter 3 the numerical models adopted for
mechatronic components and friction, Chapter 5 some analytic
simplified guidelines, Chapter 6 experimental and numerical
results, Chapter 7 conclusions and further developments.

Being this study focused on dynamic behavior, two main

requirements must be taken into account: being able to move
accurately at high speed along complex trajectories (tracking
accuracy) and to cut large volumes of material in a short time
(material removal capability).
ONENTS
1
1
2
3
4
5
6
7

fixture

of 1
(Elems)
(Elems)
(Elems)
(Elems)
(Elems)
(Elems)
(Elems)

Y
Z

AUG 2
work piece
15
point

machine base

belt
guide ways
with friction

K0
K0

Nomenclature
LuGre friction model:
z:
state variable representing the average deflection
of the bristle in the friction contact [m];
v:
relative velocity [m/s]
0, 1:
stiffness [N/m] and viscous damping [Ns/m] of
the bristle
2:
viscous coefficient related to slide speed [Ns/m]
Vs, E:
transition speed [m/s] and speed exponent
Fs, Fc, Ff: static, coulomb and total friction force [N]
simplified structural model:
M1,M2:
lumped masses [kg]
D:
MWP/ MGW
K:
stiffness [N/m]
C:
viscous damping [Ns/m]
equivalent damping analysis:
Ceq:
equivalent viscous damping [Ns/m]
area:
area of the hysteresis cycle (displ, force) [J]
::
pulsation of the harmonic excitation [rad/s]
A:
amplitude of the resulting displacement [m]
others:
dof:
degree of freedom
FEM:
Finite Element Model
FRF:
Frequency Response Function
TMD:
Tuned Mass Damper
2. The case study
The proposed methodology can be applied to a full
machine tool but in this study, to better illustrate the concept,
has been applied on a single machine axis, that moves the
work piece in the X direction (Fig 1). The mechatronic model
is composed by:

the base, fixed to the ground;

the kinematic chain (motor, belt and ball screw);

the sliding guide ways (with pads coated in

Turcite(TM), a low friction, self lubricated material);

a slide with the moving table;

a fixture, used to machine large work pieces by the

horizontal spindle (not considered in this study).

K0

linear
encoder

ball screw
and nut

motor,
encoder

Fig 1 the analyzed structure, modeled by Finite Elements.

Tracking accuracy is basically related to the reachable

bandwidth of the position control loop and can be directly
evaluated measuring the tracking error along reference
trajectories. This error is affected by axis friction, that
produces typical errors when the velocity sign is inverted
(often indicated as quadrature glitches or inversion errors)
and can reduce workpiece surface accuracy in finishing
operations. Several studies on friction modeling have tackled
those issues, with the goal of adopting model-based control
schemes that forecast the motor torque required to
compensate friction.
In order to reach a high material removal capability, a
machine must withstand process forces with limited
displacements at the tool and the work piece. A tool cutter
undergoing large displacements produces a waved machined
surface, that, when cut by the subsequent cutter, will provoke
oscillating forces. This regenerative process can became
unstable, as studied by the classic chatter theory in metal
cutting operations [6]. Stability lobes diagrams can be
computed basing on the tool work piece dynamic
compliance and cutting process parameters, indicating the
system vibration modes that mainly limit the material removal
capability.
3. The numerical model
The mechanical structure has been modeled by a Finite
Element software (by Ansys Inc..), representing the whole
kinematic chain (motor with rotary encoder, belt, ball screw
with support bearings, guide ways, linear encoder) with the
corresponding inertia and compliances (Fig 1). A 3% modal
damping has been considered for all vibration modes. As nonlinear friction models will ask for several time simulation, to
reduce the computational load while preserving a good
accuracy, the original FEM, described by 65945 nodes, has
been reduced by a Craig Bampton approach [5][4], with the

369

Giacomo Bianchi et al. / Procedia CIRP 21 (2014) 367 372

following boundary dofs: relative X translation at the guide

way pads, absolute XYZ translations at the work piece point,
rotor/stator rotation at the motor. The reduced model has 41
mechanical dofs.
After being translated into the Matlab Simulink
environment, the axis control is added, composed by the
motor electrical model and the cascaded current, velocity and
position controllers, depicted as continuous time elements.

4. Simplified analyses
Simplified analyses are proposed to provide a better
understanding of experimental tests and/or non linear
numerical simulations and to guide machine structure design.
4.1. Damper interaction with the structure dynamics
The effect of additional damping on structure vibrations
can be investigated on two basic layouts, reported in the
following figure: in both a spring and two masses realize
an oscillating system, while friction is represented as a
viscous damper. Layout (A) qualitatively explains the
effect of guide ways damping on certain structural modes
(see Fig 9). Layout (B) is analyzed analytically because it
can be mapped onto the analyzed axis: M1 depicts the slide
inertia near the guide ways, M2 the structure inertia near
the work piece and C the guide ways viscous friction.

The model has been designed to be compared to tests

executed on a machine axis in operational conditions, using
the motor itself to measure the applied torque: as single
components cannot be investigated on separate test benches,
in this study all friction forces occurring along the axis (at the
guide way, bearings, ball screw) have been lumped as an
equivalent friction located at the guide ways pads. Given the
numerical complexity of the overall system and the limited
available identification experiments, the LuGre model, well
known in the literature [6], has been chosen, to reproduce
friction forces during large motion and energy dissipation in
the stiction region, assuring continuity of the friction force
and a small parameter set to be identified, compared with
other models:




M2
x1

x2

f2

C
M1
f1

M2
x1

GW

x2

f2
WP

Fig 3 lumped parameter models for the analytic analysis

From the dynamic equilibrium of system (B), substituting

M2=DM1 and applying the Fourier transform:

 M1 :2  jC:  K

j:  j D M :  D M1C :2  j M1 K 1  D :  C K

x2
f2

>

2
1

This dynamic compliance depends on guide ways damping

C, but, for two different pulsation values, its module does not
vary with C. This behavior is typically encountered with
devices like Tuned Vibration Dampers (TMD, [7]): the so
called fixed points can be identified equating to zero the
derivative of the FRF module in respect to pulsation :,
getting:

4
Motor Current [A]

f1

(B)

Experiment 15: circular path r=0.01[mm] Feed=1[mm/min], Kv=1

2
0
-2
-4
-6

-8
-0.015

M1

(A)

The internal state z can be interpreted as the deformation of

an elastic bristle in the contact region that transfers the
friction force between the two bodies up to a limit value that
depends on the relative velocity v.

-0.01

-0.005
0
0.005
Slide Displacement [mm]

0.01

0.015

Fig 2 Motor torque in the stiction and low speed region.

The previous figure shows a test on the examined axis,

comparing the motor torque from the model with LuGre
friction to the measured torque while executing a sinusoid
positioning command with amplitude 10 Pm and a period of
3.7s.

: fixed 1

:0

1 D  1 D 2
; : fixed 2
2D

:0

1 D  1D 2
2D

where is the natural frequency of the system

without damping and with mass MWP fixed to ground.

As in TMD theory, the most desirable response is obtained
when C is tuned to bring the resonance peak at the second
fixed point. Mathematically, Copt is found when the derivative
of the module, in respect to the pulsation, is null in : fixed 2
(the transcendental function f(D) is not reported for brevity):

370

Giacomo Bianchi et al. / Procedia CIRP 21 (2014) 367 372

Copt

connect M2 to M1: in this case, the resonance frequency will

decrease.


K M1 f D  

: fixed 2 and Copt in Eq (4), the

Finally, substituting

corresponding peak height is computed, as reported in the

following Eq :
x2
f2

1  D  1  D 2 1  D   D 2 1  D  1  D 2
1
1

{ g D 
K D 1  D  1  D 2 1  1  D  1  D 2  1  D 2 K

opt

the g(D) function plotted in the following figure expresses the

fact that, even when its possible, in a given structure, to choose
the optimal C value, the maximum attenuation of the resonance
peak at the work piece depends on D: if D is small the resonance
is anyhow characterized by a high dynamic amplification (i.e.
the mass near the work piece is smaller than the mass near the
guide ways: this is the case of eigen modes that move the work
piece significantly more than the guide ways).
X2F2 Cfixed g(alpha)
30

4.2. Energy dissipation in the stiction region

Analyzing vibrations in non linear systems is useful to
introduce the concept of equivalent viscous damping, defined
in order to dissipate the same amount of energy in one
oscillation cycle as the real friction force. On a viscous
damper, friction forces would tend to zero when displacement
vanishes, instead, when static friction acts, even small
displacements generate significant forces: this difference will
produce a very large equivalent viscous damping. To
investigate this effect, the response to an harmonic excitation
force can be analyzed by the describing function approach,
where the produced displacement is represented only by its
fundamental sinusoidal term (i.e. the first harmonic in its
Fourier expansion). It is possible to demonstrate [6] that the
equivalent damping is given by:

area 
: S A2
where area is the area of the hysteresis cycle. To
compute area an oscillation cycle is considered that brings
the friction contact across the stiction region and into the
sliding region. The LuGre exponential transition between
stiction and sliding (e.g. in Fig 2) can be approximated by a
linear function (i.e. neglecting the force terms that depends on
derivative elements in Eq (3)), obtaining the simplified cycle
depicted in the following figure (force applied to the sliding
body).
 C

25

eq

g(alpha) []

20
15
10
5
0

0.1

0.2

0.3

0.4
0.5
0.6
alpha=Mwp/Mgw []

0.7

0.8

0.9

Fig 4 dynamic amplification factor g(D)at the work piece with optimum
damping Copt, from eq (7).
FRF x2 f2 MOD phase

x2/f2 mod [m/N]

Cof

Copt

2Fs

C|0

10

fixed point 1

2Fs

ffrict

V0

-6

V0

Fs
fixed point 2

-8

10

x
x2/f2 phase [deg]

0
4.912e+003
4.912e+004
4.912e+006
4.912e+007
4.912e+005

-50
-100
-150
-200

10

15

20

-Fs

Fig 5 dynamic compliance at the work piece mass as a function of guide

ways damping C.

To illustrate those results, considering a model with

parameters listed in the following table, the previous figure
shows the dynamic compliance at mass M1, where the cutting
process acts, for different values of the damping coefficient C.
Its clear that there is an optimal damping value and the
resonance frequency decreases as C is augmented.
M1
K
D

25

freq [hz]

5000 kg
16E6 N/m
0.3

The same approach could be applied to case (A) of Fig 5.

Here its sufficient to note that increasing C will progressively

Fig 6 schematic representation of the hysteresis cycle with stiction and

sliding.

The area value can be than computed by simple

geometrical considerations and substituted in Eq (8):

 C

eq

2F
1 2 FS
FS 2 A  S  2

V0 2 V0

: S A2

F
2 FS A  S
V 0 

: S A2

This formula assumes that some sliding occurs, i.e.

2FS : when equality holds, Ceq reaches its maximum
At
V0
value:
 C

eqMAX

V 0 

2: S

371

Giacomo Bianchi et al. / Procedia CIRP 21 (2014) 367 372

Eq (9) shows that Ceq decreases toward zero as A

increases. It is therefore confirmed that static friction
produces a larger damping effect, in term of equivalent
viscous damping, when oscillation amplitude is near the
displacement occurring during the transition from static to
sliding friction.
5. Results
5.1. Friction during motion
Parameters related to sliding in the LuGre model have been
identified and validated measuring the slide displacement with
the linear encoder and the motor displacement and torque
during axis large movements at different constant speeds. As
illustrated in the following figure, the correspondence with the
LuGre model is good, with a clear Stribeck effect at low
speeds.

5.2. Optimal damping

The effect of guide ways friction on axis dynamics has
been evaluated computing, on the FEM model without
position control (i.e. the axis is free to move), the dynamic
compliance at the work piece with varying values of viscous
damping at carriages (i.e. without static friction). As discussed
for cases (A) and (B) in chapter 4.1, increasing C moves mode
(A) to a higher frequency and mode (B) to a lower frequency.
Another typical case is (C): some vibration modes arent
influenced at all by guide ways damping because their shape
doesnt involve a significant displacement at the guide way. It
can be noticed that different modes ask for different optimal
damping values: it is therefore necessary to choose the best
tradeoff, taking into account how each mode influences
machine performance.
-70
-80

Turcite
6000

Sperimental tests
tests
Experimental

[m/N]
[Db]
MAGN

4000

2000

-100
-1

-110
-1

-120

-2000

-130

-4000

-140

C = 5e3[Ns/m]
C = 5e4[Ns/m]
C = 5e5[Ns/m]
C = 5e6[Ns/m]
C = 5e7[Ns/m]

-0.4

-0.3

-0.2

-0.1
0
0.1
velocity [m/s]

0.2

0.3

0.4

Phase (deg)

FrictionForce [N]

--90

Fitted curve

-6000
-0.5

FRF, FORCE ON WP X DIRECTION, DISPLACEMENT IN THE SAME POINT AND DIRECTION

0.5

Fig 7 Measured and fitted friction force in sliding state.

The stiction parameters of the LuGre model have been

identified fitting several tests where the axis executed small
motions in the stiction and sliding regions, as those illustrated
in Fig 8. The gains of the speed and position regulators have
been set at the corresponding values used on the real machine.

-45
-90
-135
-180
0
10

10

Frequency (Hz)

10

Fig 9 dynamic compliance in X direction at the Work Piece point with

increasing viscous damping at the guide ways (no static friction,
no motion control).

5.3. Vibration damping

Exp 25 double circle, Kv=1, "TURCITE"

0.045
0.04
0.035

Slide Position [mm]

0.03
0.025
0.02
0.015
0.01

Measured
Simulated
Reference

0.005
0
-0.005
0

10

15
Time [s]

20

25

30

Fig 8 Measured and simulated X position while executing circles with

20 and 5 Pm radii.

FC
E
V0
V2

nominal

var

2198 N
1
5.9E9 N/m
5578 Ns/m

1.17E9 N/m
-

nominal
FS
VS
V1

var

6013 N
258 N
0.01 m/s
-1.6E7 Ns/m 1.3E6 Ns/m

The previous table reports all the identified coefficients:

for the stiction parameters: the variance between the values
identified on each of the six executed tests is also listed.

Preliminary tests have been executed to investigate the

structural damping: an electro-dynamic shaker applied a
sinusoidal force at the work piece point in the X direction.
The excitation frequency was swept from 10 Hz to 270 Hz in
30s, while the amplitude went from 100N (the shaker
maximum value) to 50N, depending on the structure
compliance. It has been noticed that the linear optical encoder,
placed aside the linear guide ways, wasnt able to detect any
motion exceeding its resolution of 0.1 Pm: the available force
level and excitation frequencies werent able to provoke
sliding.
The numerical model has been used to investigate how
structural dynamics, control and friction interacts under
external forces. To inject energy in the vibrations modes that
mostly limit machining performance, the excitation frequency
was swept from 10 Hz to 100 Hz in 20s, with different
amplitudes, reported in Table 3. The table lists:
1) experiment setup: friction model: lg: full LuGre
non linear damping; vd = viscous damping only
(no static friction component); peak-to-peak applied
force; imposed motion speed (note: a speed of 30mm/s
allows exceeding the Stribeck speed for all values of
external force).

372

Giacomo Bianchi et al. / Procedia CIRP 21 (2014) 367 372

still. At the contrary, if the average velocity avoids

motion inversions, the difference between viscous and
LuGre friction is reduced.

reson. freq.
[Hz]

WP oscill/
max force
[mm/N]

lg

50

-0.017

-0.005

57

5.96e-5

vd

50

-0.001

-0.028

32

6.14e-5

lg

50

30

-1380

1380

32

6.641e-5

lg

100

-0,07

-0,019

57

5.89e-5

vd

100

-0.004

-0.112

32

6.14e-5

lg

100

30

-1380

1380

32

6.64e-5

lg

200

-0.281

-0.073

57

5.73e-5

vd

200

-0.016

-0.483

32

6.14e-5

lg

200

30

-1380

1380

32

6.643e-5

10

lg

600

-3.02

-0.581

57

4.7e-5

Acknowledgements

11

vd

600

-0.14

-3.94

32

6.14e-5

12

lg

600

30

-1380

1380

32

6.65e-5

energy
motor [J]

vel [mm/s]

The proposed approach allows to investigate the complex

interaction, in a machine tool, between structural dynamics,
motion control and friction forces. Simplified analytical
relationships illustrate potential and limitations of vibration
damping due to guide ways friction.
The adopted LuGre model confirms an accurate
reproduction of frictional load during motion, offering more
articulated results compared to typically used models with
viscous damping only.
The work done indicates also possible improvements:
1) model distinct friction torques and forces along the
kinematic chain, in guide ways, ball screw, bearings
2) test alternative friction models (e.g. the KUL model
with hysteresis), focusing on the transition between
stiction and sliding
3) experiment with higher excitation forces, at stand still
and during motion
4) propose new design methodologies, to manage more
articulated specification

energy
friction [J]

WP force
[N p-p]

6. Conclusions and further developments

damping
model

2) results: energy injected by friction and by the motor in

a fixed interval of 18s, to cut out the initial transient
(energy dissipated by distributed damping is not
reported). The last column reports the maximum
oscillation amplitude at WP divided by the force
amplitude.

13

lg

2000

-40.5

-4.16

38 1.9e-5
56

14

vd

2000

-1.60

-43.8

32

6.14e-5

The authors kindly acknowledge the Italian National

Research Project Michelangelo Industria 2015 (D.M. n.
00052MIOl) that supported the described activity and the
company Mandelli Sistemi Spa (Piacenza, Italy) for
participating to the study and for providing the machines used
for the experiments.

15

lg

2000

30

-1290

1340

32

6.8e-5

References

Table 3: simulated trials. In column 8: vd = viscous

damping only; lg: LuGre non linear damping
Remarks:
A) standing still axis: most of the energy injected by the
external force at the WP is absorbed: by the motor if
only viscous damping is considered; by friction, when
stiction is modeled with LuGre. For low excitation,
dynamic amplification occurs near the resonance
frequency computed with locked guide ways: 57 hz.
B) moving axis: energy of the frictional force is nearly
balanced by the energy that the motor injects to
prevails sliding friction. If speed is high enough to
avoid velocity inversion at the frictional contact, there
is little, even if not negligible, difference between
LuGre and viscous damping models. Dynamic
amplification occurs near the resonance frequency
computed with viscous damping: 32 hz.
C) when the excitation force is large enough to bring the
contact more deeply into the sliding region (excitation
2000N), oscillations are more clearly reduced by the
static friction (ref eq.10), when the axis is standing

[1] N. Cau, M. Leonesio, E. Zanotti, P. Parenti, G. Bianchi, Integrated

Machine Tool Design, Proc of the 2nd Int Researchers Symposium 2009
on Innovative Production Machines and Systems (IPROMS). Ischia
(Italy), 22-24 July, Napoli ISBN/ISSN: 978-88-95028-38-5
[2]

G. Bianchi, F. Paolucci, P. Van den Braembussche, H. Van Brussel,

Towards Virtual Engineering in Machine Tool Design, Annals of CIRP,
Vol. 45/1/1996

[3] G. Ferretti, G. Magnani, G. Martucci, P. Rocco, Friction Model

Validation in Sliding and Presliding Regimes with High Resolution
Encoders, Experimental Robotics VIII, Springer Tracts in Advanced
Robotics Volume 5, 2003, pp 328-337, doi 10.1007/3-540-36268-1_29
[4] Y. Altintas, M. Weck, Chatter Stability of Metal Cutting and Grinding,
CIRP Annals - Manufacturing Technology, Volume 53, Issue 2, 2004,
Pages 619-642, ISSN 0007-8506, doi 10.1016/S0007-8506(07)60032-8.
[5] Craig R.R., Bampton M.C.C, Coupling of Substruc-tures for Dynamic
Analysis, AIAA Journal, Vol 6, No 7, July 1968, pp. 1313-1319
[6] F. Al-Bender,W. Symens, J. Swevers, H. Van Brussel, Theoretical
analysis of the dynamic behavior of hysteresis elements in mechanical
systems, Int Journal of Non-Linear Mechanics 39 (2004) pp 1721
1735, doi:10.1016/j.ijnonlinmec.2004.04.005
[7] J Dayou, The fixed-points theory revisited with new applications, Proc.
IMechE Vol. 224 Part C: . Mechanical Engineering Science, Pages 16271634, doi: 10.1243/09544062JMES1895