See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/269631495

Force Plate Use in Performance Monitoring and Sport Science Testing

Article · January 2014

CITATIONS READS

8 21,232

3 authors:

George Beckham Timothy J. Suchomel

California State University, Monterey Bay Carroll University
35 PUBLICATIONS   341 CITATIONS    110 PUBLICATIONS   877 CITATIONS   

SEE PROFILE SEE PROFILE

Satoshi Mizuguchi
East Tennessee State University
49 PUBLICATIONS   225 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Catch phase force-time characteristics of weightlifting exercises and their derivatives View project

Exploring vertical jump performance in relation to modifiable and non-modifiable factors View project

All content following this page was uploaded by Timothy J. Suchomel on 23 December 2014.

The user has requested enhancement of the downloaded file.

 STUDY

Force Plate Use in © by IAAF

29:3; 25-37, 2014

Performance Monitoring
and Sport Science Testing

by George Beckham, Tim Suchomel and Satoshi Mizuguchi

ABSTRACT AUTHORS
Force plates are useful for examining George Beckham, MA, and Tim Suchomel,
the kinetic characteristics of an athlete’s MS, are PhD students in Sport Physiology
movement. They provide information about and Performance in the Exercise and Sport
the external forces involved in movement Science Department at East Tennessee
that can aid a coach or sport scientist to State University in the USA.
quantitatively evaluate the athlete’s ex-
ecution of a skill or his/her physical de- Satoshi Mizuguchi, PhD, is an assistant
velopment. Obtaining data of the highest professor in the Exercise and Sport Sci-
quality and minimising error requires an ence Department at East Tennessee State
understanding of the inner workings of a University.
force plate, as well as the process by which
data are transferred, processed and anal-
ysed. Knowledge of these helps validate
whether the results produced are represen-
tative of what is actually happening on the Introduction
force plate rather than error. The aim of this
article is to inform coaches and other prac- ll movements are better understood
titioners about the principles of force plate
operation, including the theoretical basis,
A by examining the forces involved.
Measurement of forces applied by an
forces plate design and function, key as- athlete can aid a coach or sport scientist to
pects of data acquisition and technical in- quantitatively evaluate an athlete’s execution of
formation of note. The authors also provide a skill or physical development. For example,
a discussion of their laboratory’s set up and data of the forces applied to the ground in a
their work with force plates as a practical vertical jump provide a picture of the explosive
example of how these tools can be used. abilities of the athlete, as well as an indicator of
his/her progress if measurements are made at
multiple times.

Force plates and other force measure-

ment devices (e.g. S-type load cells, pressure
sensors) are used to assess external forces
generated by athletes. Many force plates can
simultaneously measure external ground re-
action forces (GRFs) in three planes - vertical,

New Studies in Athletics · no. 3.2014 25

Force Plate Use in Performance Monitoring and Sport Science Testing

anterior-posterior, and medial-lateral. These at a constant velocity, i.e. without force, there
measurements provide a detailed picture of will be no change in movement. The second
the interaction between an athlete and the law states that the acceleration or change
ground. GRFs and other variables derived from in velocity of an object is proportional to the
them have been shown to have moderate rela- forces applied to it (represented by the equa-
tionships with a number of other performance tion force = mass x acceleration). The third
measures, including one maximum repetition law explains that when a force is exerted on an
back squat20,22, agility20,22, and sprint perfor- object, the object exerts a simultaneous force
mance20,22.25,31. Further, it is possible to use that is equal to and opposite in direction to the
GRF data to differentiate between high-level original force.
senior and junior rugby players12, power ath-
letes, bodybuilders and recreational athletes28, By combining laws two and three, it can be
as well as younger and older adults14. understood that when an athlete applies verti-
cal force to the ground, the ground applies an
Obtaining data of the highest quality and equal, reactive force back against the athlete. It
minimising error requires an understanding is this reactive force that ultimately results in the
of the inner workings of a force plate, as well vertical propulsion of the athlete off the ground.
as the process by which data are transferred, In other words, forces always occur in pairs -
processed and analysed. Knowledge of these an athlete pushes against the ground and the
processes helps validate whether the data ground pushes back at the same time, with the
produced are representative of what is actually same magnitude but in the opposite direction.
happening on the force plate rather than error.
The reactive forces applied to the athlete
The aim of this paper is to inform coaches are termed ground reaction forces or GRFs
and other practitioners about the principles of (see Figure 1). Key facts that should be under-
force plate operation and, including the theo- stood with this concept are 1) all ground-based
retical basis, forces plate design and function, movement is a function of the forces applied to
key aspects of data acquisition and technical the ground, even during horizontal movements
information of note. In addition, information such as running, and 2) a difference in forces
about our laboratory is presented as a practi- result in differences in how a movement in ex-
cal example of how these tools can be used for ecuted, for example if an athlete jumps two
athlete performance monitoring and research. times with different amounts of total force, the
jump with more total force will be higher.
Theoretical Basis

Movement is described by kinematic and

kinetic characteristics. Kinematics studies the
trajectories of points, lines and other geomet-
ric objects and their differential properties such
as velocity and acceleration and contributes to
a visual analysis of skill. In kinetics, we observe
the torques and forces related to movement.
Kinetic characteristics are ultimately what
cause movement, but must be observed indi-
rectly (e.g. using a force plate).

Newton’s three laws of motion are important

for understanding ground-based propulsion.
The first law explains that an object with no
force acting upon it will either remain at rest or Figure 1: Example of ground reaction force

26 New Studies in Athletics · no. 3.2014

 Force Plate Use in Performance Monitoring and Sport Science Testing

Figure 2: Jumps with variables

Measurement of forces involved with While jump height is an important measure, it

ground-based activities is typically made us- is not necessarily detailed enough to illuminate
ing force plates. Figure 2 is an illustration of some of the differences between jumps. Fig-
some of the many characteristics of interest in ure 3 is an example of a number of jumps over
a countermovement vertical jump. It should be time that shows negligible differences in jump
clear that there is a vast amount of information height, but larger differences in other variables
available from force plate data that is not nec- (e.g. peak force, rate of force development, im-
essarily available from other testing methods. pulse). A sport scientist or coach may not have
For example, vertical jump height is often used seen differences between jumps if jump height
as a performance marker of explosive ability. was the only measure used.

Figure 3: Jump force over time

New Studies in Athletics · no. 3.2014 27

Force Plate Use in Performance Monitoring and Sport Science Testing

Force Plate Design and Function calculation of the applied force. Strain gages
and beam load cells operate on the basis
One of the qualities of force, the tendency that changes in electrical current occur when
to distort, is the reason force plates are able a metal or semiconductor is deformed1,24. A
to measure applied forces. When an athlete thin sheet of metal or semiconductor material
applies any kind of force to the force plate, is bonded to a metal object, providing a solid
sensors within the plate distort, causing mea- structural device on which to apply the force.
sureable voltage changes coming from the Deformations from a force result in chang-
sensors. In a properly constructed force plate, es of the electrical resistance of the bonded
these voltage changes are proportionate to sheet, modifying the current moving through
the magnitude of the forces applied. In many it. Piezoelectric cells, strain gages, and beam
force plates, sensors are configured in differ- load cells all have an excitation voltage that is
ent orientations allowing for the direction and run through them, with the initial input voltage
magnitude (i.e., vector) of the forces to be as- known. Monitoring of the changes in voltage
certained. Important pieces of information can with applied force allows for calculation of the
be obtained from a modern, full-featured force
force applied to the device.
plate include:
• force in the X, Y, and Z directions,
A typical tri-planar force plate is construct-
• the centre of pressure,
ed with four three-component load cells17, with
• the centre of force,
each cell measuring force in the X, Y and Z
• the moment (torque) around each of the
axes. direction. The four cells are arranged in each
corner of the force plate allowing for the calcu-
The data collected allows the investigator lation of the moments about the axes, centre
to calculate a multitude of other variables from of pressure and centre of force from the indi-
the initial measures. For example, using trigo- vidual cells and their location on the force plate
nometry, the magnitude and direction of ap-
plied forces can be monitored. Typically, the forces that are transduced
with individual sensors within force plates are
Force Transducers summed to obtain a single resultant force as
Within a force plate, forces are measured an easy to interpret measure of performance.
through a force transducer, which functions In a force plate with four vertically-oriented
to convert "physical states into electrical sig- load cells, the forces measured by each of the
nals”26. The force applied to the plate is con- sensors is summed to give the total vertical
verted or transduced into a measureable, elec-
force. Likewise, in a force plate with load cells
trical voltage by what are known as load cells.
oriented in the anterior-posterior directions,
There are a variety of load cell types including
the total force in the anterior-posterior direc-
piezoelectric transducers, strain gages, and
tion is the sum of measured forces in all the
beam load cells. Each type of load cell re-
anterior-posterior cells.
ceives an “excitation” voltage input, which in
turn produces a different electrical current pro-
portional to the load experienced by the trans- Signal Flow
ducer1. After that, the operation of the various Figure 4 shows the flow of the force plate
types differs in a few distinct ways. data. Data flows from the initial analog signal
output (continuous voltage) obtained from the
Piezoelectric transducers operate on the load cells, to the final digital input signal lead-
basis that when a force is applied to a piezo- ing into the final analysis by computer soft-
electric material, such as certain crystals, ce- ware. The modifed charge is then sent to an
ramics and even bone, a charge proportional amplifier to be scaled up in voltage (the data
to the force appears on the surface of the ma- acquisition device requires a higher voltage
terial. Recording the resultant voltage allows than what is output by the load cells). Analog

28 New Studies in Athletics · no. 3.2014

 Force Plate Use in Performance Monitoring and Sport Science Testing

signal processing can occur here. After the measurement of various force-time variables.
signal has been amplified, the current goes to Contrary to the previous authors, STREET et
a data acquisition device, where it enters an al27 found that sampling rates of less than 1080
analog-to-digital (A/D) converter to convert Hz could lead to an underestimation of jump
the continuous analog signal to a series of height (calculated via the impulse method) by
discrete, evenly spaced, digital signals. Fur- up to 4.4%. Other authors recommend that
a sampling frequency of 500 Hz or 1000 Hz
ther signal processing can occur on the signal,
ensures greater accuracy, especially when im-
after which the recorded digital signal can be
pact is involved2.
analysed in a software program.
Sampling frequencies must be high enough
to ensure precision of measurement and re-
duction of signal aliasing (where the recorded
digital signal fails to accurately show the origi-
nal signal due to inadequate sampling). The
Nyquist Theorem suggests that the absolute
minimum sampling frequency is two times
the frequency of interest.2 BARTLETT recom-
mends a sampling frequency of at least 500
Hz2, however 1000 Hz is a common choice for
force plate capture of human motion17.

High sampling frequency is important for

providing sufficient resolution with regards to
force-time curves. If the frequency is low (i.e.
200 Hz) during the first 50ms of force appli-
cation, this corresponds to a force-time curve
consisting of only 11 data points. This will not
provide ample resolution with which to exam-
ine the changing forces in this short period.

Signal Amplification and Processing

Figure 4: Signal flow diagram Data collected from the force plate is not
useable until signal amplification occurs. Typi-
cally, analog signals (the raw voltage values)
from the force plate are amplified and sent to an
Data Acquisition A/D converter17. The A/D converts the analog
voltage signal into a scaled digital signal, which
Sampling can then be processed by computer software.
Monitoring the changes in force applied to Even with an ideal setup, there remain many
the force plate requires sampling at regular environmentally-based sources of error in the
intervals. Substantial research has examined collected signal. Such sources may include
sampling frequency in a wide variety of ap- thermal and chemical noise, as well as electri-
plications (e.g., engineering). In sport science, cal interference. Thermal noise is associated
however, less has been done to examine sam- with the temperature of the device in use. Part
pling frequency, although several authors have of the reason for allowing a device to warm up
evaluated the variability of vertical jump perfor- is to minimise or stabilise thermal noise, as ris-
mance at different sampling frequencies2,13,27,29. ing component temperatures result in chang-
Vanrenterghem et al29 found that frequen- ing electrical noise. Chemical noise is random
cies above 100 Hz were adequate while HORI noise existing everywhere, originating from
et al13 indicated that 200 Hz was precise for variations in temperature, humidity, pressure,

New Studies in Athletics · no. 3.2014 29

Force Plate Use in Performance Monitoring and Sport Science Testing

and other sources. Electrical noise results from

devices around the testing area that use elec-
tricity. Electrical noise exists at 60 Hz and is
progressively weaker at its harmonics - 120
Hz, 180 Hz, etc. For example, if a force plat-
form is placed in a room that has fluorescent
lights, air conditioning/heat, and the building is
located near power lines, electrical noise may
be a substantial source of interference.

Customizable software such as LabVIEW TM

(National Instruments, Austin, TX, USA) or
Matlab (Mathworks Incorporated, Natick, MA,
USA) can perform signal processing to clean
up some the noise from the data. Microsoft Ex-
celTM (Microsoft Corporation, Redmond, WA,
USA) has also been used for the smoothing
out of data. However, it should be noted that
many force plates come with proprietary soft-
ware with both filtering and smoothing meth-
ods included. Moreover, most force plates can
perform electronic signal smoothing during the
amplification process.

A method of data processing referred to as

a "moving average" can be calculated with the
use of Excel. Figure 5a represents a counter-
movement jump, with simulated random noise.
Figure 5b is the same countermovement jump
after applying a 5 ms "moving average" in
Excel. There is a direct tradeoff between the Figures 5a & 5b: Samples of moving average
method’s ability to ‘average’ out (relatively)
random noise, and the sensitivity to change
in the signal. If the signal of interest involves curate fitting of the curves, since the force pat-
very high rates of change, then shorter aver- terns of many movements on a force plate are
aging windows are warranted, as longer win- not easily represented by a single polynomial
dows will soften a curve with a sharp change. function. Cubic splines (pieced together 3rd
Conversely, a noisy signal that does not entail degree polynomials) as well as quintic splines
rapid changes may be better processed with (pieced together 5th degree polynomials) are of-
a longer moving average. One criticism of this ten used in force plate research. Quintic splines
technique is it does not distinguish between may better fit signals with very rapid changes,
noise and true signal, thus it may not actually although the problem of polynomial oscillations
isolate the true signal as well as a more ad- can occur (where false oscillations are intro-
vanced processing method6. duced into the smoothed out signal).

Splines are another smoothing method com- The Fast Fourier Transform (FFT) is useful for
monly used in the literature. This method ap- identifying noise that tends to occur at certain
plies sequential polynomial functions to a noisy frequencies, as it transforms data from time-
data signal, creating a "smooth" line closely to frequency- based, indicating the “amount”
representing the true signal2,4. Piecing together of signal at different frequencies in the newly
the polynomial functions allows for a more ac- transformed data9. The FFT shows both the

30 New Studies in Athletics · no. 3.2014

 Force Plate Use in Performance Monitoring and Sport Science Testing

signal of interest and the noise in the signal, situations. In turn, it is not wise to rely entirely
which then allows for relative ease in identify- on previous literature, as not all laboratory situa-
ing in which frequencies the data is noisy. For tions and movements are affected by the same
example, an FFT would reveal a high density of noise and sources of error6,9. However, use of
signal around 60hz (and harmonics of 120hz, prior literature with similar testing procedures
180hz) if there was a significant amount of can be a starting point, but ultimately filter selec-
electrical noise collected into the signal. The tion should be made on an individual basis4,6,9.
FFT can be very useful for identifying consis- Frequency-domain techniques, like the FFT,
tent noise in a signal, so that a particular filter can be very useful in determining which filter or
can be applied to reduce the noise that tends smoothing method to use on signal data.
toward certain frequencies. Further, an FFT
can also help to identify the range of the fre- There are a number of automated selection
quencies of the data of interest, so that noise procedures for choosing a filtering/smoothing
outside of the key range can be filtered out us- method, but the methods are "to some extent
ing another type of filter. 'black box' techniques, which should be used
with caution"4. For more detailed information
Filters are generally termed high-, low-, on signal processing, the reader is directed
band- pass, notch, and stop-band6,9. Filter- to Bartlett2, Challis4, Derrick6 and
ing can be done on either the analog or digital Street et al27. Regardless of the end-user’s
signal, or both. A low-pass filter eliminates fre- final choice on a filter, the general recommen-
quencies in the signal above a certain frequen- dation for filter choice is the one that most ef-
cy, while a high-pass filter eliminates frequen- fectively and accurately isolates the signal of
cies below a cut-off frequency. Sometimes it interest, without sacrificing data accuracy.
is necessary to restrict the data collected to
a range of frequencies, in which case a band- Calibration
pass filter is appropriate. Other circumstances Force plate calibration is necessary to es-
necessitate eliminating a certain range of fre- tablish a regression equation to calculate
quencies from a signal; a notch (small range) GRFs from the output voltage of the force
or stop-band filter (larger range) will be able to plate. Since force plates only provide an out-
accomplish that task. put voltage, a calibration equation is required
to calculate the actual GRFs. Too often, little
Butterworth filter methods are commonly is done by scientists and coaches to address
used in biomechanics, and are considered the proper calibration of force plates or to
well-suited for biomechanical variables6. But- re-evaluate currently used methods11. While
terworth filters operate sequentially through calibration of some of the testable variables of
time-series data, causing a phase-shift, which the force plate can be difficult, it is absolutely
is usually re-run backwards through the data, necessary given the immense error that can be
correcting the shift6,9. The order of the filter re- introduced into variables of interest11.
fers to the "sharpness" of the filter's transition
through the transition band. A higher order fil- The general idea behind calibration is that a
ter will have a sharper cutoff, thus the transition range of known forces is applied to the force
between frequencies that are kept and those plate observing the resultant voltage given by
that are discarded is more distinct. the load cells. This method creates a regres-
sion equation1. For instance, a common meth-
Each of these methods operates differently. od for calibrating in the Z-direction, is to place
When the optimal filter or smoothing technique a range of “dead weights” of known value on
is applied to a signal, the end result will be the top of the plate. This weight is associated with
same i.e., less choppy and smoother, with the an output voltage allowing the researcher to
data being “cleaner” and easier to analyse. Un- calibrate the force plate. Calibration of the hori-
fortunately, it is difficult to provide recommenda- zontal forces, torque, and centre of pressure
tions on filtering methods that would apply to all or centre of force can be a more difficult en-

New Studies in Athletics · no. 3.2014 31

Force Plate Use in Performance Monitoring and Sport Science Testing

deavor. However, researchers have proposed y = maximum deviation from linearity, and Y =
methods of calibration that are possible in the full-scale deflection2. Full-scale deflection refers
laboratory environment, such as a pulley sys- to the voltage output with the highest load with-
tem for X- and Y- direction recommended by in the limit of the force plate. Dividing deviation
HALL et al11. With this system, a regression from linearity by the highest voltage achieved
equation is created from the output voltages gives a relative measure of linearity, which can
in relation to progressively higher (horizontally) be compared to the standard given in Table 1.
applied loads. A pendulum system for dynamic
calibration designed by Fairburn et al8 is also Hysteresis
a possibility for more advanced calibration. Hysteresis is the difference in output values
seen during the loading and unloading of a ma-
terial2. Hysteresis should be minimised, as many
Technical Information of Note
force plate measurements involve both a load-
ing and unloading component (see Figure 6).
Force plate technical reports typically con- For example, large hysteresis in a load cell might
tain a data table with information about some over-estimate the forces in the eccentric portion
or all of the following: linearity, hysteresis, of a squat, while correctly estimating the forc-
crosstalk, and/or natural frequency. Each of es during the concentric portion. Hysteresis is
these items provides valuable information sometimes seen as a result of a mechanical lag
about the characteristics of the force plate, in deformity during the return to normal shape
as each affects the data obtained (for recom- occurring during loading of the force transduc-
mended ranges refer to Table 1). ers. Hysteresis can be calculated with the equa-
tion: (X L–X U )·Z –1·10 0, where XL = output volt-
Table 1: Guideline values for force plate characteristics age for a given load, XU = output voltage for the
same load during unloading, and Z = full scale
deflection2.
Linearity---------------------------- ≤0.5% of
full-scale
deflection

Hysteresis-------------------------- ≤0.5% of
full-scale
deflection

Crosstalk--------------------------- ≤0.5% of
full-scale
deflection

Natural Frequency----------------- ≤800 Hz

Maximum Frequency Ratio------- ≥0.2

Linearity Figure 6: Example of hysteresis of a force plate

Linearity is the maximum deviation of col-
lected force plate data from a straight line2.
Perfect linearity is ideal, but is not necessarily Cross-Talk
a requirement for accurate data collection and Cross-talk refers to the interference of force in
analysis, as it can be calibrated by applying a one component direction with the measurement
higher order polynomial to the data points2. Lin- of force by a component in another direction2.
earity can be expressed as: y·Y –1 ·10 0 (where: Many force plates measure forces in multiple

32 New Studies in Athletics · no. 3.2014

 Force Plate Use in Performance Monitoring and Sport Science Testing

planes; the components required to measure Minimising Error

the different directions generally have at least a The possibility of inaccurate measurements
minor amount of cross-talk. It is important that must be minimised if accurate conclusions are
this quality is minimised, preventing an error in to be drawn. Certain steps can be taken to re-
measurement, falsely attributing forces to an in- duce the possibility of error, although even the
correct source. Bartlett has stated that less most optimal setup will have at least a small
than 3% of full-scale deflection is preferable2. amount of error. First and foremost, if a user
does not know how to properly use the force
Natural Frequency & Maximum Frequency plate and the associated software he or she
Ratio may be the source of error, which is an argu-
When struck, every object has a frequency ment for this article and general education
at which it will tend to vibrate4. Force plates about the basics of force platform use. Adher-
are constructed of multiple materials, and may ence to the recommendations by Bartlett2
contain multiple frequencies. Force plate man- will also ensure accuracy of the data obtained
ufacturers often report this natural frequency in from the plate, as any deviance of the force
order to insure the proper set up. Force plates plate characteristics outside of the recom-
generally have high natural frequencies aiding mendations increases the risk of inaccuracy.
in the ease of isolation through filtering. For ex- For example, Lees & Lake17 and HALL et al11
ample, the Kistler Type 9281E Triplanar force have demonstrated how cross-talk of even 1%
plate (Kistler Group, Winterthur, Switzerland) could introduce a large degree of error in some
has a natural frequency of 1000 Hz. The high gait measurements.
natural frequency of this force plate allows the
measurement of the impact activities of sports, Calibration should occur over a range of
which can surpass 100 Hz2. The ratio of fre- loads, from unloaded to above the highest ex-
quency of the measured skill to natural fre- pected load, within the manufacturer-specified
quency of the force plate (Maximum Frequen- loads (if the expected loads are outside of
cy Ratio) should be less than or equal to 0.2, the range, then a new plate with greater load
so that the information of interest in the signal range is necessary). Without calibration up to
can be effectively isolated from the natural fre- at least the highest expected forces it cannot
quency of the plate (refer to Table 1 and 2). be known if the calibration equation acquired
represents the higher-range forces that will be
Table 2: Frequencies of selected cyclic movements
measured. For example, in an isometric mid-
thigh pull, where measurements of vertical GRF
can exceed 7000 N3, the force plate should be
Study Skill Frequency
calibrated in the Z-axis with loads ranging from
(single limb)
0kg to more than 700kg. For horizontal cali-

bration, a pulley system, similar to the one de-
Mackala18 0-20m
signed by HALL et al11 should suffice. Dynamic
of 100m Sprint 2.16 Hz
calibration is more tricky, requiring expensive

equipment or complex methodology. If such a
70-80m
calibration is required, it is best to enlist the
of 100m Sprint
services of major force plate companies or pri-
(Max Velocity Phase) 2.28 Hz
vate metricians.


In all devices, the measured value is the re-
Ounpuu21 Walking
sult of a "true" plus "error" score. Even though
at 1.2 m/s 1.1 Hz
steps can be taken to eliminate as much error
as possible, a degree of error will always be
Running
present in the collected data. It is up to the tes-
at 2.23 m/s 1.67 Hz
ter to eliminate and reduces as many sources
of error as possible, and to make a theoretical

New Studies in Athletics · no. 3.2014 33

Force Plate Use in Performance Monitoring and Sport Science Testing

Figure 7: Force plate and computer set-up in the sports science laboratory

and logical judgment as to how much error in to only allow collection in the vertical direction,
the collection is acceptable. the plates offer a substantial reduction in cost
compared to other plates offered by KistlerTM
Finally, the methodology and calculation or AMTITM. Furthermore, a number of studies
methods must be sound for the specific tests have demonstrated that vertical forces and
being used (e.g. Dugan et al7), although this is vertical-oriented skills have strong relation-
outside the scope of this article. ships to explosiveness and speed in sport-
ing movements; thus measurement of vertical
forces is of substantial importance10,16,30,31. In
Our Laboratory
addition, we use force plates (0.914 x 0.46m;
Rice Lake Weighing Systems, Rice Lake, WI)
The following is a discussion of the use of situated side-by-side to allow for the collec-
force plates at the Sport Science Laboratory tion of unilateral (i.e., single leg) force data. For
at East Tennessee State University in John- bilateral data collection, the forces from each
son City, Tennessee, USA, as an example of a force plate are summed.
working laboratory.
Each plate in the laboratory is interfaced to
Equipment and Processing an amplifier and conditioner module (Trans-
To reduce contamination of data from ex- ducer Techniques TM0-2, Temecula, CA,
traneous sources, all force plates in the labora- USA). The amplifier provides both the excita-
tory sit on a level concrete pad that is on the tion signal (the initial current going to the load
ground floor of the building. While the labora- cell) and amplification of the analog signal.
tory uses tri-directional force plates, the ma- Situated between the amplifier/conditioner
jority of the force plates we, the authors, use and the A/D converter is a shielded connec-
are unidirectional, and measure only vertical tor block, [BNC-2110 (National Instruments,
forces (see Figure 7). Although it is a drawback Austin, TX, USA)], which transfers the analog

34 New Studies in Athletics · no. 3.2014

 Force Plate Use in Performance Monitoring and Sport Science Testing

Table 3: Example of force plate applications

Study Measure

Kraska, et al.16 Jump Height for Countermovement and Static Jumps

at Various Loads

Leary, et al. Instantaneous Force at 30 ms, 50 ms, 90 ms, 100 ms, 200 ms,
and 250 ms in measured in the isometric mid-thigh pull

Mizuguchi
Evaluated the use of vertical jump net impulse as a variable
for athlete performance monitoring

Sato and Heise23 Investigated weight distribution asymmetry between

right and left legs during the barbell back squat

Jensen and Ebben15 Various rate-of-force development measures across

multiple plyometric exercises

Beckham, et al.3 Isometric peak force, measured at 3 key conventional

deadlift positions and the isometric mid-thigh pull

signal to the A/D converter. In turn, the block collection. This also avoids a potential shift of
is connected to another instrument [DAQCard- voltage output over time. After the warm-up
6063E (National Instruments, Austin TX, USA)], period, force plates are calibrated using loads
which converts the analog to a digital signal. from 0 to 350kg or 500kg, depending on the
specific use of the plate (either jumps or iso-
The acquired signal is analysed with custom metric pulls, respectively). The plates are pro-
software developed in LabVIEW TM (National gressively loaded in 25kg increments, with the
Instruments, Austin TX, USA). This software output voltage recorded with each new plate.
samples the analog signal at 1000 Hz and has A linear regression equation is then applied to
been set up to save the digital signal file and the calibration load. This regression equation
filter the digitised signal using a 4th order low- is saved, and used in the LabVIEW TM program
pass Butterworth filter at 100 Hz. From there, during analysis.
the signal is now ready to be analysed for any
variables of interest. Data Use
The majority of information we obtain from
The Data Collection Process the aforementioned set-up is related to one
Prior to calibration, force plates, amplifiers, aspect of performance monitoring conducted
A/D converters, and computers are turned on a regular basis (2-4 times per year). As we
on so that all of the collection equipment can evaluate variables related to maximal strength,
warm up, thus stabilising thermal and instru- isometric rate of force development, and ex-
mentation noise. Calibration of the force plates plosive performance in static and counter-
is performed immediately before data collec- movement jumps with a variety of loads, we
tion, assuring that the calibration equation have a very good picture of the general abilities
used in data analysis is established under of the athletes. Single assessments provide in-
similar environmental conditions as the data formation for valuable normative comparisons

New Studies in Athletics · no. 3.2014 35

Force Plate Use in Performance Monitoring and Sport Science Testing

against normative data while multiple assess- REFERENCES

ments over time provide a worthwhile view
of changes in performance that an athlete or 1. BAHRA, C. S. & PAROS, J. (2010). Measurement of
group of athletes experience over the training Force, in: Instrumentation reference book. W Boyes, ed.
Amsterdam ; Boston: Butterworth-Heinemann/Elsevier,
process. In both testing cases, this data can
127-133.
help substantially in programme development
and future planning of training. 2. BARTLETT, R. (2007). Introduction to sports biomechan-
ics : analysing human movement patterns. Milton Park,
Abingdon, Oxon ; New York, NY: Routledge.
Conclusion 3. BECKHAM, G. K.; LAMONT, H. S.; SATO, K.; RAMSEY,
M. W.; HAFF, G. G. & STONE, M. H. (2012). Isometric
A full understanding of a testing device and Strength of Powerlifters in Key Positions of the Conven-
tional Deadlift. J of Trainology, 1: 32-35.
its characteristics is an integral part of accu-
racy, validity and reliability of testing. The force 4. CHALLIS, J. H. (2008). Data Processing and Error Esti-
mation, in: Biomechanical evaluation of movement in sport
plate is a rather complex device. The com- and exercise : the British Association of Sport and Exercise
plexities of the device and its peripherals allow Sciences guidelines. C Payton, R Bartlett, eds. Abingdon,
the user to collect a large variety of high-quality Oxon ; New York, NY: Routledge, 53-76.
data for analysis that are difficult to obtain via 5. CORMIE, P.; MCBRIDE, J. M. & MCCAULLEY, G. O.
other means. This diversity of data gives op- (2007). The influence of body mass on calculation of power
tions for a large variety of analysis projects during lower-body resistance exercises. J Strength Cond
Res, 21: 1042-1049.
(see Table 3). While it is somewhat difficult to
master use of the device, the information ob- 6. DERRICK, T. R. (2004). Signal Processing, in: Research
Methods in Biomechanics. DGE Robertson, GE Caldwell, J
tained from a force plate makes the endeavor
Hamill, G Kamen, SN Whittlesey, eds. Champaign, IL: Hu-
worthwhile. man Kinetics, 80-83.

7. DUGAN, E. L.; DOYLE, T. L.; HUMPHRIES, B.; HASSON,

C. J. & NEWTON, R. U. (2004). Determining the optimal
load for jump squats: a review of methods and calculations.
J Strength Cond Res, 18: 668-674.
Please send all correspondence to:
8. FAIRBURN, P. S.; PALMER, R.; WHYBROW, J.; FIELDEN,
George Beckham S. & JONES, S. (2000). A prototype system for testing force
[email protected] platform dynamic performance. Gait Posture, 12: 25-33.

9. GIAKAS, G. (2004). Power Spectrum Analysis and Filter-

ing, in: Innovative Analyses of Human Movement. N Ster-
giou, ed. Champaign, IL: Human Kinetics, 223-258.

10. HAFF, G. G.; STONE, M. H.; O'BRYANT, H.; HARMAN,

E.; DINAN, C.; JOHNSON, R. & HAN, K. H. (1997). Force-
time dependent characteristics of dynamic and isometric
muscle actions. J Strength Cond Res, 11: 269-272.

11. HALL, M. G.; FLEMING, H. E.; DOLAN, M. J.; MILL-

BANK, S. F. & PAUL, J. P. (1996). Static in situ calibration of
force plates. J Biomech, 29: 659-665.

12. HANSEN, K. T.; CRONIN, J. B.; PICKERING, S. L. &

DOUGLAS, L. (2011). Do force-time and power-time mea-
sures in a loaded jump squat differentiate between speed
performance and playing level in elite and elite junior rugby
union players? J Strength Cond Res, 25: 2382-2391.

13. HORI, N.; NEWTON, R. U.; KAWAMORI, N.; MC-

GUIGAN, M. R.; KRAEMER, W. J. & NOSAKA, K. (2009).
Reliability of performance measurements derived from
ground reaction force data during countermovement jump
and the influence of sampling frequency. J Strength Cond
Res, 23: 874-882.

36 New Studies in Athletics · no. 3.2014

 Force Plate Use in Performance Monitoring and Sport Science Testing

14. IZQUIERDO, M.; AGUADO, X.; GONZALEZ, R.; LOPEZ, 22. PETERSON, M. D.; ALVAR, B. A. & RHEA, M. R. (2006).
J. L. & HÄKKINEN, K. (1999). Maximal and explosive force The contribution of maximal force production to explosive
production capacity and balance performance in men of movement among young collegiate athletes. J Strength
different ages. Eur J Appl Physiol Occup Physiol, 79: Cond Res, 20: 867-873.
260-267.
23. SATO, K. & HEISE, G. D. (2012). Influence of weight dis-
15. JENSEN, R. L. & EBBEN, W. P. (2007). Quantifying tribution asymmetry on the biomechanics of a barbell back
plyometric intensity via rate of force development, knee squat. J Strength Cond Res, 26: 342-349.
joint, and ground reaction forces. J Strength Cond Res,
21: 763-767. 24. SCOTT, D. M. (2008). Industrial process sensors. Boca
Raton: CRC Press.
16. KRASKA, J. M.; RAMSEY, M. W.; HAFF, G. G.; FETHKE,
N.; SANDS, W. A.; STONE, M. E. & STONE, M. H. (2009). 25. SLEIVERT, G. & TAINGAHUE, M. (2004). The relation-
Relationship between strength characteristics and un- ship between maximal jump-squat power and sprint accel-
weighted and weighted vertical jump height. Int J Sports eration in athletes. Eur J Appl Physiol, 91: 46-52.
Physiol Perform, 4: 461-473. 26. SOLOMAN, S. (2010). Sensors handbook. New York:
17. LEES, A. & LAKE, M. (2008). Force and Pressure Mea- McGraw-Hill.
surement, in: Biomechanical evaluation of movement in 27. STREET, G.; MCMILLAN, S.; BOARD, W.; RASMUS-
sport and exercise. CJ Payton, R Bartlett, eds. New York, SEN, M. & HENEGHAN, J. (2001). Sources of error in de-
NY: Routledge, 53-76. termining countermovement jump height with the impulse
18. MACKALA, K. (2007). Optimisation of performance method. J Appl Biomech, 17: 43-54.
through kinematic analysis of the different phases of the 28. UGRINOWITSCH, C.; TRICOLI, V.; RODACKI, A. L.;
100 metres. New Studies in Athletics, 2: 7-16. BATISTA, M. & RICARD, M. D. (2007). Influence of training
19. MCBRIDE, J. M.; KIRBY, T. J.; HAINES, T. L. & SKIN- background on jumping height. J Strength Cond Res 21:
NER, J. (2010). Relationship between relative net vertical 848-852.
impulse and jump height in jump squats performed to 29. VANRENTERGHEM, J.; DE CLERCQ, D. & VAN CLEV-
various squat depths and with various loads. Int J Sports EN, P. (2001). Necessary precautions in measuring correct
Physiol Perform, 5: 484-496. vertical jumping height by means of force plate measure-
20. NIMPHIUS, S.; MCGUIGAN, M. R. & NEWTON, R. U. ments. Ergonomics, 44: 814 - 818.
(2010). Relationship between strength, power, speed, and 30. WISLØFF, U.; CASTAGNA, C.; HELGERUD, J.; JONES,
change of direction performance of female softball players. R. & HOFF, J. (2004). Strong correlation of maximal squat
J Strength Cond Res, 24: 885-895. strength with sprint performance and vertical jump height
21. OUNPUU, S. (1994). The biomechanics of walking and in elite soccer players. Br J Sports Med, 38: 285-288.
running. Clin Sports Med, 13: 843-863. 31. YOUNG, W.; MCLEAN, B. & ARDAGNA, J. (1995). Re-
lationship between strength qualities and sprinting perfor-
mance. J Sports Med Phys Fitness, 35: 13-19.

New Studies in Athletics · no. 3.2014 37

View publication stats