Improving Dynamic Performance of Cantilever Boring Bars

E. I . Rivin, H. L. Kang; Wayne State University - Submitted by L. Kops ( I ) , McGill University

Received on January 6,1989

Key words: Boring bar, Vibration absorber, Cutting force,

Random excitation.

Summary:
The paper deals with improving dynamic performance of cantilever
boring bars by using so-called combination structure with tuned
dynamic vibration absorber. Absorber parameters were synthesized
using analytical results for cases of self-excitation and random
excitation of the main system. Combination structure was designed
to achieve an optimal combination of stiffness and absorber
effectiveness (damping). Test results are given for the boring bar
with length-to-diameter ratio L/D=15.

1. Introduction
Slenderness and low damping of boring bars and other with
cantilever tools make them prone to chatter during
cutting. Chatter vibrations affect part geometry and
surface finish, as well as productivity. Two basic ways
to improve performance of a boring bar are enhancement Cl cz
of its stiffness and damping. Stiffness can be increased .'fi251; \lKznl,-252
by using high Young modulus materials such as sintered
carbides. Both such materials and fabrication fromthem HereX1, X2 arevibrationamplitudes of mainand absorber
are very costly. To increase damping, dampers and masses, w 1, w 2 and c 2 are partial natural frequencies
vibration absorbers are frequently used. In damper and damping ratios of main and absorber subsystems,
designs [1][4][5][6][7], the optimumdamping is reached respectively. kt is absorber mass ratio.
by optimizing clearance between cavity wall inside the
boring bar and inertia weight. A boring bar with damped If in ( 2 ) F(t)=F,, the characteristic equation for
vibration absorber is designed in [ Z ] . Length-to- the set (2), ( 3 ) becomes:
diameter ratios L/D<10 are typical for above designs,
with one [7] L/D=12.7.
Another approach is application of a combination
structure [lO][ll]. The root segment of the cantilever
is made of a material with high Young modulus to increase
its effective stiffness and the free end segment is
designed to be light to reduce its effective mass. In
this design natural frequencies are increased and
effective mass is reduced without a significant
reduction of stiffness. The latter factor leads to
increase in the mass ratio of the absorber, especially
considering limited space inside the bar and thus a
limited size of the inertia weight, and results in an Here w is chatter frequency which is close to but
improved absorber effectiveness. different from w , (due to addition of Kcx) and 5. is
In this paper, performance of optimized combination damping ratio ofthemainmass subsystemduring cutting,
boring bars with an optimally tuned dynamic vibration which combines <,(alwayspositive) and effective damping
absorber (DVA) is discussed.
from the dynamic cutting force (1). The latter can be
2. Optimal Tuning Conditions for DVA positive, thus assuring an unconditional dynamic
Classical optimization for DVAS' parameters [ 3 ] is stability of the system, or negative, which then should
based on a harmonic external excitation applied to the be compensated by the positive E l and by the stabilizing
main mass. Frequently these optimal parameters are effects of the absorber in order to achieve stable
considered as universal for DVAs' design. However, conditions. Thus, effectiveness of the absorber can be
cutting processes are characterized by random excit- judged by the critical value of 5. which corresponds
ations and self-excited vibrations [9]. Optimal to the stability boundary of the system. And the maximum
parameters of DVAs for such systems are quite different. effectiveness of the absorber can be characterized by
a. System under dynamic cutting force excitation the maximum magnitude of negative critical value of 5 ,
In theunstable dynamic cutting process (self-excited which the absorber can still compensate.
vibrations) the alternating force that sustains the Routh stability criterion requires that Bo, B1, B2,
motion is created or controlled by the motion itself: B3 must be positive and satisfy inequality [ 3 ] :
when the motion stops the alternating force disappears.
The general expression of the dynamic cutting force can B,B2B,> B:+BzBo (5)
be written as [13]: For the system (4):
S.>-r52(l+p) (6)

where Kcx, Ccx are "effective cutting stiffness" and

"effective cutting damping", respectively. At some
combinations of parameters, force (1) can lead to c e > - - 52
self-excitation of vibrations.
A DVA can be modeled as a mass (inertia weight) M2 r
attached by a spring K2 and a damper C2 to the main
system whose mass MI is subjected to the force F(t) as where frequency ratio r = z .
shown in Fig.1 where K1 and C1 are stiffness and damping
of the main mass. The critical value of <.can be obtained by replacing
Equations of motion of this system can be written all (6)-(8) with equalities. The largest value 5. is
as :
then checked with @ ) . If the latter is not satisfied,
+ ~ ~ ~ l ~ l + s 2 ~ 2 ~ ~ x , + ~ ~ ~ + ~ : the
~ ~critical
~ l value of E.can be determined by iterations
to satisfy (5). Critical values of <.at various mass,
frequency, and damping ratios of the absorber can be
.Y2 + 2 c 2 w 2I 2 + w :x 2 - 2c2w2s , - w :u I = 0 (3)
thus found. At a given mass ratio 11, there exist optimal
(tuning) frequency ratio w d w a n d optimal damping ratio
t20f absorber, at which critical negative value of 5,
has the maximum magnitude (maximum effectiveness of the
absorber).

Annals of the CIRP Vol. 38/1/1989 377

 The optimal frequency ratio at the optimal absorber self-excitation and random excitation exist during
damping condition will be called the global optimal cutting process [9], the optimal tuning values for the
frequency ratio. Since the absorber damping in practical frequency ratio and damping could be chosen in between
designs may not be optimal, the optimal frequency ratio of the optimal values shown in Figs.2,3 for cases Of
at this situation is locally optimal, which means that self-excitation and random excitation.
if a not optimum value of absorber damping is used, the 3 . Optimal Structural Design
local optimal frequency ratio has to be chosen for
absorber to be the most effective at this damping. Both damping and stiffness are important for cutting
Optimal absorber damping and global optimal frequency stability. Their influences can be described as a
ratio for a given mass ratio are shown in Fig.2 and criteria1 "performance index" [14]1151 A = K t , where K
Fig.3 (lines 2). is stiffness and 5 damping ratio. Higher value of the
b. System under random excitation performance index corresponds to better stability.
Thus, it seems to be the natural optimization criterion
Optimal tuning parameters will be synthesized for for a combination boring bar with absorber.
a case of random excitation having white noise char- A combination bar of length L with sintered carbide
acteristics with a constant spectral density function in the root segment (length L1, E1=551600 MN/rn2, and
S O and zero memory. In real circumstances random signal
is frequently constant over a wide, but not infinite, specific gravity p1=14281 Kg/rn3) and aluminum in the
frequency band. Thus, white noise excitation is an free end segment (solid part of length Lz, hollow part
approximation. of length L3, E2=68950 MN/m2, and p2=2597 Kg/m3), shown
In order to get the frequency response functions in Fig.4, was analyzed. The absorber made of heavy
H1(u) and Hz(U) [ a ] , first let: machinable tungsten alloy (p0=17982 Kg/m3) is located
in the hollow part of the free end also shown in Fig.4.
The Rayleigh expression for fundamental natural
frequency of the system without absorber was used,

Substituting above expressions and their derivatives

into (2) and (3), frequency response functions become:
where E ( z ) is Young's modulus, I(z) is moment of inertia
of the cross section, m(z) is mass per unit length, and
all these parameters are considered as function of z
(coordinate along the axis of the cantilever bar). Mt
is the tool head mass at the free end (0.085 Kg), x 1s
vibration amplitude of cantilever bar (a function of
where z ) , and xi is vibration amplitude at the free end of
cantilever bar.
A,- 2 ( t l w i + Ezw2( +rr)) The effective mass at the length L is:
A, = w:+ w;( 1 11)' 4 ~ l ~ z w l w z
+

4 I = 2(5, w I w: + t,w,w:,
A,=W:W;

For the main system without DVA, the frequency The effective stiffness at L then is:
response function is:
k , = w ;f\.I I (19)
The approximate fundamental mode shape of the
cantilever bar to be used in the Rayleigh formula (17)
is assumed to be:

The mean square response of the mass MI under the

white noise excitation is then:
s(z)=l-cos
I:;( -
which satisfies the boundary conditions for this mode.
Three values were chosen for length L3 (=101,127,152
.
mm) By calculating stiffness values at tool end Kt and
which represent the total energy of the main system critical value 5. of the combination bar with DVA under
with attached DVA. optimal tuning and damping conditions, the performance
A non-dimensional normalized mean square response index q, vs Ll/L ratios can be obtained as shown in
of the main subsystem, i.e. the ratio of mean square Fig.5. Since both higher stiffness Kt and more negative
response of M 1 with absorber to mean square response critical value of 5. give better cutting process sta-
of MI without absorber, which reflects effect of the
absorber on the main mass, is defined as: bility, higher magnitudes of the performance index are
corresponding to a better stability of the system. It
can be seen that Ll/L in the range 0.45-0.6 correspond
to the best stability of the boring bar. It can be shown
[12] that Ll/L=O.45 corresponds to the highest natural
where frequency of the combination bar, and at Ll/L=O. 6, the
A = - A , A , -A0A,(45:w:-2W:)+ w;(.il - A , A , ) combination bar with optimized DVA has the minimum
response under harmonic excitation (partly, due to
B = A,(A,A:+ A:- A , A,A,) higher structural stiffness).
Since the mean square response represents the total 4. Cutting Test Results
energy of the system over the entire frequency range, Combination boring bars of 31.75 mm diameter and
the normalized mean square response gives the total 476.25 mm long (L/D=15) with structural damping ratio
energy ratio of the response of the main mass MI but SI=O.02 were designed with two length ratios L1/L=O.6
not the response itself. The larger the normalized mean
square response, the larger is response of the main and 0.45; length of absorber cavity L3=152 mm; and
mass and the lesser effect of the absorber on the main absorber mass m2=0.787 Kg. Absorber frequency tuning
system behavior. Influence of the mass ratio on the can be done by adjusting the screw shown in Fig.4, thus
global optimal frequency, and absorber damping ratios preloading the rubber elements and increasing their
is similar to the case of self-excited vibration. stiffness.
Computed optimal global frequency ratio and optimal Cutting test were done on a lathe with a rotating
absorber damping ratio are plotted in Fig.2 and Fig.3 workpiece. Carbide tool tips with zero rake angle were
(lines 3 ) . used; workpiece material was 1045 steel. Vibration
A mass with DVA under sinusoidal excitation [ 3 ] was displacements of the boring bar in horizontal (x)
considered for zero main system damping which is a good direction at the distance 343 mm from the clamp for
approximation for the cantilever boring bar system Ll/L=0.6 bar and 3 8 1 mm for Ll/L=O.45 bar were measured
(lines 1 in Fig.2,3). by LVDT. These locations were selected sincemeasurement
It can be seen from Fig.2 and Fig.3 that at the of the tool end vibration is impossible during cutting.
same mass ratios, the required optimal absorber damping Static measurements have shown that deflection at the
values are the lowest for the case of white noise tool end is about 1.91 and 1.5 times of deflections
excitation, and the highest for the case of sinusoidal measured at the points where LVDTs were located in these
excitation. The required optimal frequency ratios are two cases. These data were used to calculate the
the lowest for the case of sinusoidal excitation and vibration amplitudes at the tool end from LVDT readings
the highest for the case of self-excited vibrations. during cutting.
For the case of cantilever boring bar, since both The first design of combination boring bar has
Ll/L=O.6 with equivalent mass at Lo (at the midpoint 12. Rivin, E. I. and Kang, H. L., 1989, "Optimized
of the absorber cavity) about m1=0.735 Kg, which gives Boring Bar Design with Damped Vibration Absorber",
the mass ratio p=1.07. Natural frequency of the boring Proceedings ofthe 15thNationalScience Foundation
bar is fl=173 Hz. The second design of combination Conference on Production Research and Technology
boring bar has the length ratio of Ll/L=O.45 with (in print)
equivalent mass m1=0.397 Kg at LO: for this case p=1.97 13. Tobias, S . A , , 1965, "Machine-Tool Vibration1',John
and fl=207 Hz. The second design of combination boring Wiley & Sons, Inc., New York.
bar has the length ratio of L1/L=0.45 with equivalent 14. Tlusty, J., 1981, Criteria for Static and Dynamic
mass m1=0.397 Kg at LO: for this case p=1.97 and fl=207 Stiffness of Structures" in Report of The Machine
HZ. Tool TaskForce 1 1 , Machine ~ o o l
Mechanics, Lawrence
Fig.6 gives the maximum vibration peak-to-valley Livermore Lab, Vol. 3 , pp. 8.5.1-8.5.54.
(p-v) values at the tool end for boring bars of the
first design without and with DVA having various 15. Takeyama H., Lijima N., Nishiwaki N., and Komoto
absorber damping and tuning adjustments. The results K., 1984, "Improvement of Dynamic Rigidity of Tool
show substantial effects of DVA. The absorber tuning Holder by Applying High Damping Material", Annals
frequencies are: 84 Hz incase of sinusoidalexcitation; Of the CIRP, Vol. 3 3 , pp. 249-252
105 Hz in case of random excitation: and 117 Hz in case
of dynamic cutting force. The results show that if
tuning at local optimal frequency ratio at given damping
of rubber was realized according to the dynamic cutting
force and random white noise excitation case, smaller
vibrations were observed (10-30% lower p-v values) than
in case of tuning for sinusoidal excitation. It was
also observed that when damping of the absorber is
closer to the optimal damping values, the vibration
amplitudes are smaller.
Machining results for combination boring bar with
L1/L=O.6 and 5 ~ 0 . 4 5are given in Table-1, and with
Ll/L=O.45 and 5 ~ 0 . 2are given in Table-2.
5. Conclusions
(1). The stability analysis of boring bar with DVA
under dynamic cutting force and random white noise
excitation has shown that the optimal tuning conditions
and damping values are different from those in the
classical case of sinusoidal excitation. Optimal
tuning/damping parameters for actual cutting are shown
to be between these two sets of optimal values.
( 2 ) . Optimization of the combination boring bar
using performance index as a criterion results in a
recommended length ratio Ll/L=O.45-0.6.
( 3 ) . A stable cutting performance of a boring bar
with L/D=15 equipped with adjustable DVA has been
demonstrated.
Acknowledgements
Support fromthe Institute for Manufacturing Research
(IMR) of Wayne State University and NSF Grant DMC-
8718911 is gratefully appreciated.

References
1. Au, Y. H. J., Ng, K. W. and New, R. W., 1978, "The
Lanchester Damper-A Design Procedure for Opti-
mizing the Damping Ratio for a Cylindrical Slug
Damper Fitted to a Machine Element", Journal of
Mechanical Design, Transactions of the ASME, pp.
1-7.
2. Donies,J. andVandenNoortgate, L., 1974, "Alesage
Des Trous Profonds Sans Broutage", Note Technique
10, Crif, Belgium.
3. DenHartog, J. P., 1956, "Mechanical Vibrations",
McGraw-Hill, New York. x2
4. Hahn, R. S . , 1951, "Design of Lanchester Damper
for Elimination of Metal-Cutting Chatter", Trans.
ASME, Vol. 73, No.3. Fig.1 Model of boring b a r w i t h aarnpea vibration
absorber under a n external excitztion
5. Kato, S . , Marui, E. and Kurita, H., 1969, llSome
Considerations on Prevention of Chatter Vibration
.
-
in Boring Operations", ASME Journal of Engineering Under c u t t i n g f o r c e e x c i t a t i o n
for Industry, pp. 717-730. v Under random e x c i t a t i o n
6. Ng, X. W. and New, R . W., 1976, "Profile Boring Undzr s i n u s o i d s l e x c i t a t i o n
Operations, Results of Recent Research with New
Designs of Damped Boring Bars" International
Journal of Production Research, Vol. 14, No. 2,
pp. 149-169.
7. New, R. W. and Au, Y. H. J., 1980, "Chatter-Proof
Overhang Boring Bar Stability Criteria and Design
Procedure for a New Type of Damped Boring Bar",
Journal of Mechanical Design, Vol. 102, pp.
611-618.
8. Newland, D. E., 1975, "An Introduction to Random
Vibration and Spectral Analysis", Longman.
9. Pandit, S . M., Subramaian, T. L., and Wu, S . y.,
1975, "Modeling Machine Tool Chatter by Time
Series", Transactions of the ASME, Journal of
Engineering for Industry, pp. 211-215. 0.2 '
a!,
1

20 o!. l!O 1: I!, 1!0 1.n

1 ' 1
2.0
10. Rivin, E. I., 1983, "Chatter-Resistant Cantilever
Boring Bar", Proceedings of NAMRC-XI, SME, pp. Mass ratio /u
403-407.
11. Rivin, E. I. and Wu, X . , 1986, "An extra-Long F i g . L Optlrnal damping ratlo of absorber vs mass ratio
Cantilever Boring Bar With Enhanced Chatter (optlmal tuning condltion)
Resistance", Proceed. of xv NAMRC.

379
 - .
v
Under c u t t i n g f o r c e e x c i t a t i o n
U n d e r random axzitstion
m-u Under s i n u s o i d a l e x c i t a t i o n

<3
0.1
Spin d i e
-
Max. p - v ?,fax. p-r- Surface
-
Speed x-dir. y-dir. Finish Ra
(rpin) (mm) (mm) (,urn)
.-

;
2

5
u
[r
L. 0.4-
0.6

0.: -
42
80
130
17M
0.0"9"1
0.03073
0.02692
0.02641
0 .o 6 6 a o
0.08JI) I
U.O'J601
0.08280

s = O . O 5 l Inrn/rev, t=O.254 inin

3.33
4.06
3.9n
5.23

Table-1 T e s t r e s u l t s for combination boring b a r

0.2 , 1 ' 1 ' with D V A a t L ~ i L z 0 . 6g , = O (5)

,1 Rotation
Speed
(rpm)
Cutting
Dcpth
(mmi
Cutting
Feed
(mm, rev)
Surface
Finish Ra
hm)
I
4d 0.254 0.0762 4,44
/ 68
80
0.254
0.254
0.076'7
0.0762
4.19
4.08
Carbide 130 0,254 0.076' 4.95
I
150 0.254 0.0762 4.80
'710 0.254 0.0ib1 4.b9
80 0.381 0.Oi62 4.19
Rubber Machinable tungsten Tool head 80 0.501 IJ. 0 7 6 2 3.13
80 0.2.54 0 1016 4.52

Table-? T e s t r e s u l t s f o r combination boring b a r

with D V A a t L I / L = 0 . 4 5 Q 2= 0 2)

F i g 4 Combination boring bar i n d

dynaioic v i b r s t i o n absorber
.--. L 3 = 15 2 mi
H L ~ - 1 2 7 nim
-L3=101 mm

X
T)
.- -2E+OS -
0
c

-4E+(U
0.0 ' of1 ' of2 ' a!! ' a,! 015 ' old I 7 7
Length ratios L l / L

-
C Fig.5 Performance index of combination
.---0 boring bar with DVA
0
;Jo D V A
.- t! -Tuned for sinusoidal 2x2.
U
H Tuned f o r c u t t i n g f o r c e exc.
t
X M Tuned f o r r n n d c m 9:cc.
.-
c
C
.-
0

z-;.
ad

'5 E 0.10-
.l--
0

0.06 -
>
I
b.
0.m 1 1
2 60 do lbo 1 0 l$ !A0 lbo 2 220
0
I
Spindle speed (rprn)
Fig.6 Vibrations of borin bar with and without
DVA (s=0.0762 rnrnyrev, t=0.254 rnrn)

380