00078 Lin Hu - Timber Design Society

CONTROLLING CROSS-LAMINATED TIMBER (CLT) FLOOR
VIBRATIONS: FUNDAMENTALS AND METHOD
Lin Hu 1 and Sylvain Gagnon 2
ABSTRACT: Cross-Laminated Timber (CLT) is proving to be a promising solution allowing wood to compete in
building sectors where traditionally steel and concrete have predominated. The dynamic behaviour of CLT floor
systems differ from that of the traditional lightweight wood-joisted floors and heavy concrete slab floors. The existing
standard vibration-controlled design methods for lightweight and heavy floors may not be applicable to CLT floors. A
new design method was developed based on the understanding of the fundamentals of floor vibrations and laboratory
study of CLT floors with various construction details. The design method predicted vibration performance of CLT
floors well and matched with the subjective ratings. The new design method is simple that the vibration-controlled
spans can be directly calculated from CLT stiffness and mass. The vibration-controlled spans of CLT floors predicted
by this new design method were almost the same as the spans determined by “CLTdesigner” software that was
developed in Austria. It is concluded that the proposed design methodology to determine vibration-controlled
maximum spans of CLT floors is promising.
KEYWORDS: CLT floor, Vibration, Normal Walking, Design Method
1 INTRODUCTION 123
Cross-Laminated Timber (CLT) is proving to be a
promising solution allowing wood to compete in
building sectors where traditionally steel and concrete
have predominated. Figure 1 shows a typical crosssection
of a CLT floor.
Previous studies conducted at FPInnovations have found
that bare CLT floor systems differ from traditional
lightweight wood-joisted floors that have a typical mass
around 20 kg/m 2 and a fundamental natural frequency
greater than 15 Hz. Figure 2 illustrates the conventional
North American lightweight wood joisted floor built
with joists and 15mm-18mm thick wood composite
panels. Various toppings and ceilings are common in
multi-family wood buildings.
Figure 1: Cross-section of a bare CLT floor
1 Lin Hu, FPInnovations, 319, rue Franquet, Quebec, QC G1P
4R4, Canada. Email: lin.hu@fpinnovations.ca
2 Sylvain Gagnon, FPInnovations, 319, rue Franquet, Quebec,
QC G1P 4R4, Canada. Email:
sylvain.gagnon@fpinnovations.ca
Figure 2: Conventional North American Lightweight
wood floor built with joists and subfloor

Furthermore, even if non-joisted CLT floor looks similar
to the concrete slab floor, the CLT floor systems differ
from heavy concrete slab floors in terms of construction
details. The typical concrete slab floors have a mass
greater than 200 kg/m 2 and a fundamental natural
frequency less than 9 Hz. Based on FPInnovations’ test
results, bare CLT floors were found to have a mass
varying between approximately, 30 kg/m 2 to 150 kg/m 2 ,
and a fundamental natural frequency greater than 9 Hz.
Due to CLT floor’s unique dynamic behaviour, the
existing standard vibration-controlled design methods
for lightweight and heavy floors may not be necessarily
applicable to CLT floors. Many of the manufacturers
recommend using a uniform distribution load (UDL)
deflection method for CLT floor control vibrations by
limiting the static deflections of the CLT panels under a
UDL. Using this approach, success in avoiding excessive
vibrations in CLT floors relies mostly on the engineer’s
judgement. A new design methodology is needed to
determine the vibration-controlled spans for CLT floors.
This paper describes the development of a new design
method to control CLT floor vibrations.
2 METHOD
2.1 KNOWLEDGE OF THE FUNDAMENTALS
OF FLOOR VIBRATIONS
The new design method was developed based on the
understanding of the fundamentals of floor vibrations.
FPInnovations’ previous study [1] on wood-joisted
floors found that for floors with a fundamental natural
frequency above 9 Hz, the vibrations induced by normal
walking exhibit a transient nature. The transient
vibrations can be controlled through controlling the
combination of floor stiffness and mass. Simply by
controlling the combination of the fundamental natural
frequency and 1 kN static deflection, it is possible to
successfully control the vibration of wood-joisted floors.
This led to the proposal of a new design method to
control wood joisted-floor vibrations. SINTEF’s
extensive field CLT floor vibration study further proved
this understanding of the fundamentals of wood floor
vibrations [2]. SINTEF found that with the
FPInnovations’ new design method using 1 kN static
deflection and fundamental natural frequency as design
parameters to control wood joisted-floor vibrations,
predictions of the field CLT floor vibration performance
was in alignment with occupants’ expectations.
2.2 LABORATORY STUDY
Laboratory tests and subjective evaluations were
conducted on CLT floors with certain variables such as
CLT element thickness, floor spans, type of the betweenelement
joints, connections and support conditions.
2.2.1 CLT Floor Specimens
The floor specimens were built using three individual
pieces of 2.0 m wide CLT panels with different
thicknesses (i.e. 230 mm, 182 mm and 140 mm). The
spans varied from 8.0 m to 4.5 m and with two types of
joint details for connecting panels together. The joint
details are showed in Figures 3 and 4 below.
Figure 3: Step joint (half-lapped)
Cross-laminated LVL
Figure 4: Spline joint
For the “step joint” detail, 8 mm diameter Würth screws
were used to connect two CLT panels at a spacing of 320
mm o.c. In the case of the spline joint detail, normal
wood screws No. 10 (diameter of 4.83 mm.) were used
to connect the continuous strip of cross-laminated-LVL
to the CLT panels. The spacing was 200 mm o.c.
The ends of each CLT floor assembly were supported on
190 mm thick and 685 mm high Glulam walls connected
using Würth screws.
The floor assemblies were tested under two types of
supporting conditions at the floor edges, i.e. free and
simple support along the longitudinal direction of the
panels. When the floor edges were supported, the 38
mm x 89 mm wood stud wall panels of 2.0 m long were
used as supports. The wall panels were spaced at 2.0 m
or less.
The measured performance parameters were natural
frequencies, modal damping ratios, static deflection
under 1 kN static load, velocity and accelerations due to
a 1 N-S impulse using the test protocols developed at
FPInnovations [3]. This study was to identify the
constructions and design parameters that significantly
affected the CLT floor vibration performance measured
by the natural frequencies, static deflections, velocities
and accelerations, and human perceptions.

2.2.2 Subjective Evaluation
The key objective of the subjective evaluation of
vibration performance is to define the maximum
annoying vibration level that can be acceptable to the
majority of occupants of residential floors.
To more closely mimic normal living conditions during
subjective performance evaluations, the CLT floor
assemblies were carpeted and furnished with a cabinet
and two vases filled with water and flowers, as shown in
Figures 5 and 6. The reason for decorating the china
cabinet using flowers and water is because these objects
are good indicators of floor excessive vibrations. If the
floor is vibrating at high level when a person is walking
by the cabinet, he/she will notice the flower and the
water moving. In Figure 6, you also can see some kind
of china cabinet with glassware. It has been observed
that the glassware in the china cabinet is another good
indicator of excessive vibrations. Typically, with poor
floors, one can easily notice the rattling of glassware in
the room. In these Figures, you could also see a chair
located in the centre of the floor. During the subjective
evaluation, if the person is sitting on the chair and feels
the floor vibrating when someone is walking by the
chair, then the floor vibration is most likely not
acceptable.
Figure 6: Subjective evaluation while the evaluator is
sitting, feeling, and observing the floor movement
The evaluator was then asked to sit on the chair, while
another person would walk on the floor according to a
pre-designated pattern (Figure 6). The walking pattern is
such that the person walks at least two times along the
two diagonal directions from one corner of the floor to
the other, then walks at least two times along the middle
lines in the parallel and perpendicular directions of the
subfloor panels. Again the evaluator was observing the
three clues, i.e. seeing, hearing and feeling regarding
floor vibration performance. Immediately after the
evaluation, the evaluator is asked to fill out a
questionnaire which provided an overall performance
rating for the floor as well as a score for the three key
performance-related clues.
2.2.3 Static Concentrated Load Test
This test was conducted to determine the maximum
static deflection of the floor under a 1 kN concentrated
static load. This is a measure of the entire stiffness of the
floor.
Figure 5: Subjective evaluation while the evaluator is
walking, feeling, and observing the floor movement
At least twenty persons were asked to evaluate the
performance of a floor subjectively. Only one evaluator
was allowed on the floor at a time. He or she first walked
freely on the floor while observing clues related to floor
performance (Figure 5). The clues included movement of
the flowers and the water, rattling of the china cabinet,
and feeling the movement of the floor.
The basic elements needed to measure static deflection
under a concentrated load are: 1) a stable reference from
which to measure floor movement, 2) an accurate and
sensitive deflection measuring device, and 3) a mobile
loading system. In this study, the concrete ground floor
was used as the reference. Two electronic gauges having
a resolution of 0.001 mm were used as the deflection
measuring devices as shown in Figure 7. The deflection
gauges were mounted to the free ends of two rods in
contact with the concrete ground floor surface. The end
of one deflection gage was fixed to the bottom of the
middle point of the centre CLT panel to measure the
static deflection of the floor centre. The end of the
second deflection gauge was fixed to the bottom of the
middle point of the joint of two CLT panels to measure
the static deflection of the joint. The concentrated static
load was applied by a person standing over each CLT
panel’s centre, floor edges and the joints in turn while
recording the measurement at the gauge location. Figure
8 shows the static loading process using the tester’s
weight. The deflection profile of the floor was generated
from a complete set of measurements. Three

measurements were taken at each loading location to
ensure that stable results were obtained. The average of
three sets of the deflection profiles were used to plot the
deflection profile of the test floor under the person’s
weight. The deflection measurements were normalized
to 1 kN load.
to determine the natural frequencies, modal damping
ratios and mode shapes.
Figure 9: Modal test on a CLT floor specimen
Figure 7: Static deflection measurements using two
deflection gages under the CLT floor
Figure 8: Static loading using the tester’s weight
2.2.4 Modal Test
Modal test was conducted to determine the natural
frequencies, modal damping ratios, and vibration modes
of the CLT floor specimens.
Modal testing followed a standard procedure specified in
FPInnovations’ protocols [3]. Hammer excitation was
selected because of its simplicity and reliability. The
hammer impact was applied at the top of the floor by a
person sitting on a beam supported on the ground so that
the tester’s weight was not added to the floor. The
hammer impact was located on the side panel and was
offset from the mid-span of the test floor areas. At such
a location, it was unlikely that a nodal point of the first
three modes would occur. The floor vibration was
measured on each panel at one quarter of the span of the
test floor areas. Figure 9 shows a typical modal test setup
and the locations for the hammer and accelerometers.
The force and acceleration signals were recorded by a
multi-channel analyzer. The signals were post-processed
2.2.5 Forced vibration test
Forced vibration testing was conducted to determine the
dynamic responses including acceleration, velocity, and
displacement responses of the CLT floors to an impulse
similar to the heel impact force of the foot steps of
human normal walking.
Forced vibration testing followed the procedure
described in the FPInnovations’ test protocols [3]. The
excitation was triggered by dropping a 5 kg medicineball
on a force plate instrumented with the force
transducer. The ball drop impact was performed by a
tester while sitting on a stool. The ball was caught when
it rebounded to avoid multiple impacts. The ball drop
force was applied at the floor centre. An accelerometer
was also located at the impact location to measure the
maximum response of the floor. The impact force and
acceleration signals were recorded by the multi-channel
analyzer. The acceleration signals were then integrated
to obtain velocity or displacement responses. The impact
force signal was processed to determine the impulse,
which was further used to normalize the dynamic
responses to a 1 N-s impulse. Figure 10 shows the ball
drop impact test and the measurement setup.
Figure 10: Forced vibration test on a CLT floor specimen
using the ball drop as the excitation

3 RESULTS
3.1 KEY CONSTRUCTION AND DESIGN
PARAMETER
Based on data analysis, it was found that the
combination of fundamental natural frequency with 1 kN
static deflection, or with acceleration, or with velocity
was well correlated to human perception. It is
understood that the fundamental natural frequency is
mainly controlled by the longitudinal stiffness and the
mass, while the 1 kN static deflection is determined by
the entire stiffness of the CLT floor in both the
longitudinal and lateral directions. The acceleration or
velocity is determined by the excitation, the entire
stiffness and by damping ratio. It has been wellrecognized
that determining the damping ratio accurately
and to a certain reliability level of is not easy.
Reproducibility is another issue. Besides, damping is
largely controlled by the floor constructions details such
as the joints, connections, and support conditions, and
the non-structural components such as potions, flooring,
toppings, insulation materials, furniture, etc. Moreover,
the modal test results showed that, for the bare CLT
floors tested, the measured damping ratio did not vary
from one assembly to another as it was quit constant
with a value of around 1%.
Based on the laboratory study results, it has been found
that in meeting the safety requirements for the supports
and joints, the vibration performance of CLT floors were
largely controlled by the CLT stiffness along the
longitudinal direction and by the mass. The type of joints
between the CLT panels did not significantly affect the
measured fundamental natural frequencies, the 1 kN
static deflections, and the subjective ratings. Therefore,
it was decided to use the stiffness in the longitudinal
direction and the mass as the key parameters in the
design method to control CLT floor vibrations through a
combination of fundamental natural frequency and 1 kN
static deflection.
3.2 PROPOSED DESIGN METHOD
The proposed design method to control CLT floor
vibrations consists of a design criterion and the relevant
equations to calculate the design parameters.
3.2.1 Scope
At this point, the scope of the proposed design method to
control vibrations of CLT floors covers the following
1. bare floors with finishing, partitions and
furniture, but without heavy topping,
2. vibrations-induced by normal walking,
3. well-supported floors,
4. well-jointed CLT panels, and
5. inclusion of the self weight of CLT panels only
(i.e. without live load).
cover various types of toppings and ceilings, and other
floor design options.
3.2.2 Advantages
The proposed design method is focused on target
features, which include, among others
1. simple for hand calculation,
2. user-friendly,
3. mechanics-based using the design values CLT
panels available in producer’s specification,
4. reliable to prevent CLT floors from excessive
vibrations induced by normal walking.
3.2.3 Design Criterion
Based on the understanding of the fundamentals of floor
vibrations and the special features of CLT floor
vibrations, and following the laboratory test results, a
proposed simple design criterion using fundamental
natural frequency and 1 kN static deflection of a simple
1-m wide CLT panel as design parameters has been
developed.
The design criterion is expressed in Equation (1).
f
d
0.7
13.0
Or
1.43
f
d
39
where f = fundamental natural frequency calculated
using Equation (2) in Hz, and d = 1 kN static deflection
calculated using Equation (3) in mm.
3.2.4 Equations to Calculate the Design Parameters
f
3.142
2
2l
EI
1m
eff
A
(1)
(2)
where, f = fundamental natural frequency of 1m CLT
panel simply supported in Hz, l= CLT floor span in m,
EI 1
m
eff
= effective apparent stiffness in the span direction
which is published by the producers for 1m wide panel
in N-m 2 , = density of CLT in kg/m 3 , and A = crosssection
area of 1-m wide CLT panel, i.e. thickness*1m
width in m 2 . The static deflection under 1 kN load can be
calculated using Equation (3) below.
3
1000Pl
(3)
d
1m
48EI eff
where, d = static deflection at mid-span of the 1m wide
simply supported CLT panel under 1 kN load in mm,
and P = 1000 N.
However, because of the mechanics-based feature, it is
possible to expand its scope to include other construction
details. A study has been planned to extend the scope to

1kN static deflection calculated
using Eq.2 (mm)
3.2.5 Verification
Figure 11 shows a comparison of predictions using the
proposed new design methodology to predict the CLT
floor vibrations and corresponding subjective ratings by
evaluators.
2.00
1.50
1.00
0.50
Predicted CLT floor vibration performance vs.
subjective ratings
criterion ( f/d^0.7>13.0)
Unacceptable
Marginal
Acceptable
0.00
0 5 10 15 20
Fundamental Natrual frequency calculated using Eq. 1 (Hz)
Figure 11: Predicted CLT floor vibration performance by
the proposed design method vs. subjective rating by
participants
3.2.6 Simple Form of the Design Method
Inserting Equations (2) and (3) into the design criterion,
i.e. Equation (1), it is possible to obtain a simple form of
the design method which can be expressed by Equation
(4) as given below.
l
1
9.15
( EI
1m
0.293
eff
0.123
( A)
)
(4)
Using Equation (4), it is possible to determine the
vibration controlled spans for CLT floors directly from
the effective apparent stiffness in the span direction,
density and cross-section area of 1m wide CLT panels.
3.2.7 Impact Study
The vibration controlled CLT floor spans determined
using the proposed design method were compared with
the spans determined by CLTdesigner developed at
university of Graz in Austria [4]. Table 1 provides the
comparison.
Table 1: Vibration controlled CLT floor spans determined
using the new design method vs. the spans determined
by “CLTdesigner”
CLT
Thickness
(mm)
Span
Determined
by the
Proposed
Method
(m)
Span
Determined by
“CLTdesigner”
for 1%
Damping
100 3.58 3.53
120 3.76 3.75
140 4.50 4.43
160 4.80 4.76
180 5.16 5.14
200 5.68 5.67
220 5.84 5.89
240 6.09 6.17
As given in Table 1, the vibration controlled spans of
bare CLT floors predicted by this new design method
were very close to those spans determined by
CLTdesigner.
4 CONCLUSIONS
It is concluded that the proposed design methodology to
determine vibration-controlled maximum spans of bare
CLT floors is promising. This methodology uses only
the design values of CLT mechanical properties. The
method is simple, user-friendly, and reliable.
Wide acceptance of the proposed design method relies
on the use and evaluation of the method by productmanufacturers
and designers. FPInnovations is open to
feedback and ready to adopt and further refine the design
method according to the needs of the producers and
designers. The current form of the design method is for
CLT floors without heavy topping. A study of the effect
of heavy topping on the vibration performance of CLT
floors is under way.
ACKNOWLEDGEMENT
FPInnovations would like to thank its industry members
and Natural Resources Canada (Canadian Forest
Service) for their financial support of this work.
The authors wish to thank KLH for providing CLT
panels for this study and the guidance on CLT floor
construction. Thanks are also extended to Mr. Thomas
Orskaug of Moelven Massivtre AS (Now of KLH Solid
Wood Scandinavia AB) and Dr. Anders Homb of
SINTEF Byggforsk for sharing their experience of
massive wood slab non-joisted floor systems with us and
for providing the opportunity to visit CLT buildings in
Norway. Thanks are also extended to Dr. Gerhard
Schickhofer of Graz Institut für Holzbau und
Holztechnologie, Austria for conducting the comparison
given in Table 1.

REFERENCES
[1] Hu J.: Design guide for wood-framed floor systems.
Final report No.32 for Canadian Forestry Service.
FPInnovations, Quebec, 2007.
[2] Homb A.: Vibrasjonsegenskaper til dekker av
massivtre (in Norwagien). Report of Prosjektrapport
24, Sintef Byggforsk, Norway, 2008.
[3] Hu J.: Protocols for field testing of wood-based
floor systems. Appendix V in Report of
Serviceability design criteria for commercial and
multi-family floors. Report No. 3 for Canadian
Forest Service, FPInnovations, Quebec, 1998.
[4] Schickhofer G.: Comments on FPInnovations new
design method for CLT floor vibration control.
Private e-mail provided the link to access to
CLTdesigner,2010:
http://www.cltdesigner.at/webstart/testversion/cltdes
ignertestversion.jnlp