Excitation Techniques IMAC 30 by Pete Avitable

Excitation Techniques
Do’s and Don’ts

Peter Avitabile Marco Peres
UMASS Lowell The Modal Shop
1 Dr. Peter Avitabile

Excitation – Do’s and Don’ts Structural Dynamics & Acoustic Systems Lab
Excitation Considerations
Objectives of this lecture:
• Overview impact testing considerations – part 1

• Overview shaker testing considerations – part 2
• Identify some typical concerns

• Provide some examples
IMAC 27 presentation covered shaker excitation techniques

IMAC 29 presentation covered shaker testing considerations

MODAL SPACE – In Our Own Little World
Could you explain and how is it

modal analysis used for solving
dynamic problems?
Illustration by Mike Avitabile Illustration by Mike Avitabile Illustration by Mike Avitabile
http://sdasl.uml.edu/umlspace/mspace.html
Series of articles on various
aspects of modal analysis
currently in its 15th continuous
year of publication
Measurement Definitions - Refresher
ANALOG SIGNALS
INPUT
ANTIALIASING FILTERS
OUTPUT
Actual time signals

AUTORANGE ANALYZER
ADC DIGITIZES SIGNALS
Analog anti-alias filter
INPUT
APPLY WINDOWS
OUTPUT
Digitized time signals

INPUT OUTPUT
Windowed time signals
COMPUTE FFT
LINEAR SPECTRA
LINEAR
INPUT
SPECTRUM
LINEAR
OUTPUT
SPECTRUM Compute FFT of signal
AVERAGING OF SAMPLES
COMPUTATION OF AVERAGED
INPUT/OUTPUT/CROSS POWER SPECTRA
INPUT
POWER
SPECTRUM
CROSS
POWER
SPECTRUM
OUTPUT
POWER
SPECTRUM Average auto/cross spectra
COMPUTATION OF FRF AND COHERENCE
Compute FRF and Coherence

FREQUENCY RESPONSE FUNCTION COHERENCE FUNCTION

Measurements - Linear Spectra - Refresher
x(t) h(t) y(t) TIME
INPUT SYSTEM OUTPUT FFT & IFT

Sx(f) H(f) Sy(f) FREQUENCY
x(t) - time domain input to the system

y(t) - time domain output to the system
Sx(f) - linear Fourier spectrum of x(t)
Sy(f) - linear Fourier spectrum of y(t)
H(f) - system transfer function

h(t) - system impulse response

Measurements - Power Spectra - Refresher
Rxx(t) Ryx(t) Ryy(t) TIME
INPUT SYSTEM OUTPUT FFT & IFT

Gxx(f) Gxy(f) Gyy(f) FREQUENCY
Rxx(t) - autocorrelation of the input signal x(t)
Ryy(t) - autocorrelation of the output signal y(t)
Ryx(t) - cross correlation of y(t) and x(t)
Gxx(f) - autopower spectrum of x(t) G xx ( f )  Sx ( f )  S*x ( f )
Gyy(f) - autopower spectrum of y(t) G yy ( f )  Sy ( f )  S*y ( f )
Gyx(f) - cross power spectrum of y(t) and x(t) G yx ( f )  Sy ( f )  S*x ( f )

Measurements - Derived Relationships - Refresher
Sy HSx
H1 formulation
- susceptible to noise on the input
Sy S*x G yx
- underestimates the actual H of the system
Sy S*x HSx S*x H 

Sx Sx G xx
*
Other
H2 formulation formulations
- susceptible to noise on the output
for H exist
Sy S*y G yy
- overestimates the actual H of the system
Sy S*y HSx S*y H 

Sx Sy G xy
*
COHERENCE
(Sy S*x )(Sx S*y )
   
G yx / G xx H1
(Sx S*x )(Sy S*y )
2
xy
G yy / G xy H2
Typical Measurements - Refresher
Measurements - Auto Power Spectrum Measurements - Cross Power Spectrum
x(t) y(t)
AVERAGED INPUT
POWER SPECTRUM
AVERAGED OUTPUT
POWER SPECTRUM
G xx (f) G yy (f)
INPUT FORCE OUTPUT RESPONSE
G xx(f) G yy (f)
AVERAGED INPUT AVERAGED OUTPUT
AVERAGED CROSS
POWER SPECTRUM POWER SPECTRUM POWER SPECTRUM
G yx (f)
Measurement Definitions 12 Dr. Peter Avitabile Measurement Definitions 13 Dr. Peter Avitabile
Modal Analysis & Controls Laboratory Modal Analysis & Controls Laboratory
Measurements - Frequency Response Function Measurements - FRF & Coherence
Coherence
1
Real
AVERAGED INPUT AVERAGED CROSS AVERAGED OUTPUT
POWER SPECTRUM POWER SPECTRUM POWER SPECTRUM

0
0Hz AVG: 5 200Hz

G xx(f) G yx(f) G yy(f)
COHERENCE
Freq Resp
40
dB Mag
-60
0Hz AVG: 5 200Hz
FREQUENCY RESPONSE FUNCTION

FREQUENCY RESPONSE FUNCTION
H(f)
Measurement Definitions 14 Dr. Peter Avitabile Measurement Definitions 15 Dr. Peter Avitabile
Modal Analysis & Controls Laboratory Modal Analysis & Controls Laboratory

IMAC 30 - Jacksonville, FL - 2012


Impact Excitation
• Overview impact excitation techniques

• Review hammer/tip characteristics
• Review special DSP considerations
• Identify areas of concern and things to consider

Impact Excitation
An impulsive excitation which is very short in the time window

usually lasting less than 5% of the sample interval.
ADVANTAGES
- easy setup
- fast measurement time
- minimum of equipment
- low cost
CONSIDERATIONS
- poor rms to peak levels
- poor for nonlinear structures
- force/response windows needed
- pretrigger delay needed
- double impacts may occur
- high potential for signal overload and underload of ADC

Practical Modal Impact Test Checklist
• General
• Range settings for channels
• Frequency range – bandwidth – BW
• Hammer
• Pre-trigger settings
• Hammer tip selection
• Windows
• Response
• Windows
• FRF / Coherence
• Measurement considerations

Impact Excitation - Hammer Tip Selection
The force spectrum can be customized to some extent

through the use of hammer tips with various hardnesses.
A hard tip has a very short pulse and will excite a wide
frequency range. A soft tip has a long pulse and will excite
a narrow frequency range.
However, the hammer tip alone does not totally determine

the frequency range excited. The local flexibility of the
structure must also be considered.

Impact Excitation - Hammer Tip Selection
METAL TIP HARD PLASTIC TIP
Real Real
-976.5625us 123.9624ms -976.5625us 123.9624ms

TIME PULSE TIME PULSE
dB Mag dB Mag
0Hz 6.4kHz
0Hz
FREQUENCY SPECTRUM 6.4kHz
FREQUENCY SPECTRUM
SOFT PLASTIC TIP RUBBER TIP
Real Real
-976.5625us 123.9624ms
-976.5625us
TIME PULSE 123.9624ms
TIME PULSE
dB Mag dB Mag
0Hz 6.4kHz
0Hz
FREQUENCY SPECTRUM 6.4kHz FREQUENCY SPECTRUM

Impact Test – Pretrigger Delay – Sometimes Confusing
If the leading portion of the time pulse is not captured then

there will be a distortion of the measured input spectrum
t=0
NO PRETRIGGER USED
t=0
PRETRIGGER SPECIFIED

Impact Test – Double Impact
Double impacts can occur due to a sloppy hammer swing or

many times due to the responsive nature of many
structures. They should be avoided wherever possible.
DOUBLE IMPACT DOUBLE IMPACT
Real Real
-976.5625us 998.53516ms -976.5625us 998.53516ms

TIME PULSE TIME PULSE
dB Mag dB Mag
0Hz 800Hz
0Hz
FREQUENCY SPECTRUM 800Hz
FREQUENCY SPECTRUM

Impact Excitation - Windows May Be Necessary
If response does not die out then a window may be required
ACTUAL TIME SIGNAL
SAMPLED SIGNAL
WINDOW WEIGHTING
WINDOWED TIME SIGNAL

Impact Excitation - Exponential Window
If the signal does not

naturally decay within the
sample interval, then an T=ND t
exponentially decaying
window may be necessary.
However, many times
changing the signal
processing parameters such
as bandwidth and number
of spectral lines may
T=ND t
produce a signal which
requires less window
weighting

Impact Excitation – Force & Exponential Window

Impact Excitation - Right Hammer for the Test
Measurement adequacy depends on what is required
40
COHERENCE
dB Mag
FRF
INPUT POWER SPECTRUM
-60
0Hz 800Hz
40 COHERENCE
FRF
dB Mag
INPUT POWER SPECTRUM
-60
0Hz 200Hz

Impact at One Point – Listen at Another – What FRF?
H out/in = H row/col

Impact at One Point – Listen at Another – What FRF?
1 2 3 1 2 3 1 2 3
1 1 1
2 2 2
3 3 3
1 2 3

Drive Point Measurements
Drive point measurement

Same input and output location – in the same direction
1 2 3

Reciprocity - H out/in = Hij
Reciprocity is an underlying necessity for modal theory
1 2 3

Reciprocity - H out/in = Hij
Reciprocity is an underlying necessity for modal theory
H out/in
Hij

Reciprocity - H out/in = Hij - What can go wrong?
1 2 3

Impact Test – Multiple Reference Impact Test
Either a row or column of the

FRF matrix is needed to
estimate mode shapes
Ref#1
Ref#2 Ref#1 Ref#2 Ref#3
Ref#3

Shaker Test vs. Impact Test – What is the difference?
Typical Shaker Test Typical Impact Test
1 2 3
h 13
1 2 3 1
3
2
h 23
3
h 33
h 31
h 33 h 32

Measurement Definitions
June 1998 – Modal Space Articles

FRF from Impact or Shaker Data
Impact Data Shaker Data

ANALOG SIGNALS ANALOG SIGNALS
INPUT OUTPUT INPUT OUTPUT
ANTIALIASING FILTERS ANTIALIASING FILTERS
AUTORANGE ANALYZER AUTORANGE ANALYZER

ADC DIGITIZES SIGNALS ADC DIGITIZES SIGNALS
APPLY WINDOWS APPLY WINDOWS
COMPUTE FFT COMPUTE FFT

LINEAR SPECTRA LINEAR SPECTRA
LINEAR LINEAR LINEAR LINEAR
SPECTRUM SPECTRUM SPECTRUM SPECTRUM
AVERAGING OF SAMPLES AVERAGING OF SAMPLES
COMPUTATION OF AVERAGED COMPUTATION OF AVERAGED

INPUT/OUTPUT/CROSS POWER SPECTRA INPUT/OUTPUT/CROSS POWER SPECTRA
INPUT CROSS OUTPUT INPUT CROSS OUTPUT

POWER POWER POWER POWER POWER POWER
SPECTRUM SPECTRUM SPECTRUM SPECTRUM SPECTRUM SPECTRUM
COMPUTATION OF FRF AND COHERENCE COMPUTATION OF FRF AND COHERENCE
FREQUENCY RESPONSE FUNCTION COHERENCE FUNCTION FREQUENCY RESPONSE FUNCTION COHERENCE FUNCTION


FFT Reference vs. Modal Reference – Confusing Nomenclature
SHAKER TEST Reference means different

things to different people
That is why there is a “swap”
ROVING IMPACT TEST
HP35665 FFT Analyzer
002Z006Z.DAT
OUT
PUT USB
Accelerometer
DYNAMIC
SIGNAL 6
ANALYZER 5
Impact Hammer 4
z
1 2 3 4 3
2 y x
OUT 1
FILTER
IN
Hij
Why Do Initial Conditions Need to be Zero?
Laplace Domain Equation of Motion
(ms 2 csk) x (s)  f (s)  (ms  c) x 0  mx 0
Characteristic Applied Initial Initial

Portion Force Displacement Velocity
Assuming that initial conditions are zero
(ms 2 csk) x (s)  f (s)

ACTUAL IMPACT RESPONSE
USER PERCEPTION
SAMPLE CAPTURED
RESPONSE SAMPLE
CAPTURED
WINDOWED SAMPLE
CAPTURED

SAMPLE
CAPTURED
SAMPLE CAPTURED WITH

RINGING OF FIRST SAMPLE
SAMPLE CAPTURED WITH

RINGING OF FIRST AND
SECOND SAMPLE

Too Hard a Hammer Tip Can Cause Problems
Energy is imparted to the structure beyond the frequency

range of interest and may overload or saturate the response
40
128 HZ BW
dB Mag
-30
-60
dB Mag
0Hz 800Hz
50 HAMMER TIP
EXCITES MODES
dB Mag -80
OUTSIDE BAND
0Hz 400Hz
50
OF INTEREST
-50
0Hz 200Hz
dB Mag
40
dB Mag
-50
0Hz 400Hz
-60
0Hz 200Hz

Impact Spectrum Considerations
Selecting the right impact tip to excite the right frequency

range is critical to optimizing the measured response
128 HZ BW INFORMATION BEYOND BW
-30
STRONG RESPONSE WEAK RESPONSE
dB Mag
40
1.0 VOLT 1.5 VOLT

dB Mag
ENERGY ENERGY 4.0 VOLT 0.1 VOLT
-80 ENERGY ENERGY
0Hz 128Hz
50 -60
0Hz 800Hz
dB Mag 40
dB Mag
1.0 VOLT 3.0 VOLT
-50 ENERGY ENERGY
0Hz 128Hz
9.0 VOLT 0.5 VOLT
-60 ENERGY ENERGY
0Hz 800Hz

Exponential Window – Can It Be a Problem?
3.5
While a window V
TIME PULSE
may be ultimately Real
required, never
start with the -1.5
window applied
V
-25.391ms 973.633ms
before the raw 2.5

V
WINDOWED RESPONSE
measurement is Real
reviewed.
-2.5
V
-25.391ms 973.633ms
How many peaks 50
are observed in dB Mag
the measured FRF.

-50 FREQUENCY RESPONSE FUNCTION
0Hz 400Hz

2.5
Here is a V RAW TIME RESPONSE
measurement Real
where a significant
amount of damping -2.5
is applied to the
V
-25.391ms 973.633ms
measurement. 1.2
V
WINDOWED RESPONSE
Real
How many peaks

are observed in -800
mV
-25.391ms 973.633ms
the measured FRF. 25
dB Mag
0Hz 400Hz

2.5
V RAW TIME RESPONSE
Picking a longer Real
time block allows

the response to -2.5
naturally decay
V
-25.3906ms 1.9736s
and lessens the 2.5

V
need of the Real

WINDOWED RESPONSE
exponential
window.
-2.5
V
-25.3906ms 1.9736s
50
How many peaks

are observed in dB Mag
the measured FRF.

0Hz 400Hz
TWO CLOSELY SPACED MODES

Window should only be applied once it is deemed necessary
2.5
2.5
3.5 V RAW TIME RESPONSE
V RAW TIME RESPONSE
V
TIME PULSE
Real
Real
Real
-2.5
-2.5
-1.5 V
V
V -25.3906ms 1.9736s
-25.391ms 973.633ms
-25.391ms 973.633ms
2.5
2.5 1.2 V
V V
WINDOWED RESPONSE WINDOWED RESPONSE WINDOWED RESPONSE
Real
Real Real
-2.5
-2.5 -800 V
V mV -25.3906ms 1.9736s
-25.391ms 973.633ms -25.391ms 973.633ms
50 50
25
dB Mag dB Mag
dB Mag
-50 FREQUENCY RESPONSE FUNCTION FREQUENCY RESPONSE FUNCTION -50 FREQUENCY RESPONSE FUNCTION
-25
0Hz 400Hz 0Hz 400Hz

0Hz 400Hz
TWO CLOSELY SPACED MODES

Double Impacts – A Problem – Or is it ???
Picking a poor measurement location avoids the double impact

but does the measurement look better?

Double Impact – Common Difficulty
October 2008 – Modal Space Articles

If you can’t avoid double impact – what about multiple impacts

So if you can’t avoid double impact – what about multiple

So if you can’t avoid double impact – what about multiple

Multiple Impacts – A Possibility !!!
Single Impact Multiple Impacts

Should I look at all the measurements ?
This measurement looks fine – but do all look this good?
COHERENCE
FORCE
SPECTRUM
IMPACT EXCITATION
TIME RESPONSE
ACCELEROMETER RESPONSE
DRIVE POINT FREQUENCY RESPONSE FUNCTION

Should I look at all the measurements ?
Here’s a measurement that doesn’t look as good as the rest.
COHERENCE
IMPACT EXCITATION FORCE

SPECTRUM
TIME RESPONSE
FREQUENCY
RESPONSE
FUNCTION

All measurements should be reviewed
IMPACT EXCITATION
TIME RESPONSE
COHERENCE
FORCE
SPECTRUM COHERENCE
FORCE
SPECTRUM
FREQUENCY
DRIVE POINT FREQUENCY RESPONSE FUNCTION RESPONSE
FUNCTION

Filter Ring
Sometimes there can be some ringing on the impact input.

This is referred to as “filter ring”
Depending on the bandwidth and impact spectrum, this may

or may not appear on the measured data
The following slide shows the effects of this phenomena

Filter Ring 400 HZ BANDWIDTH SETTING 1600 HZ BANDWIDTH SETTING
RED AIR CAPSULE RED AIR CAPSULE
BLUE PLASTIC BLUE PLASTIC
WHITE PLASTIC WHITE PLASTIC
BLACK METAL BLACK METAL

2KHz excitation for 500 Hz BW
Sometimes data may be collected for multiple purposes.

One group wants data to 500Hz and another needs 2KHz.
Can a test be constructed with one set of accelerometers to
acquire the data for both test ranges?
Difficult to achieve unless you have infinite resolution and

infinite spectral resolution.

2KHz excitation for 500 Hz BW
AVERAGED INPUT AVERAGED CROSS AVERAGED INPUT AVERAGED CROSS

POWER SPECTRUM POWER SPECTRUM POWER SPECTRUM POWER SPECTRUM
COHERENCE COHERENCE
FUNCTION FUNCTION
ALLOWS MORE SENSITIVE

LOW FREQUENCY ACCELEROMETER
TO BE USED TO MEASURE SYSTEM
FREQUENCY HIGH FREQUENCY ACCELEROMETER FREQUENCY

RESPONSE RESPONSE
REQUIRED TO MEASURE SYSTEM
FUNCTION FUNCTION
1 KHz 2 KHz 1 KHz 2 KHz
COHERENCE
FUNCTION
FREQUENCY
RESPONSE
FUNCTION

Excitation – Do’s and Don’ts Structural Dynamics &2 KHz
1 KHz Acoustic Systems Lab
Impact Location Effects – Skewed and Same Point
When performing impact testing it is important

to impact the same point in the same direction
for all averages.
One case will be presented to show the effects

of having a skewed input, that is different for
each average of the measurement.
Another case is presented to show the effects

of impacting close to the same point, but not
exactly the same point, for all averages.

Impact Location Effects
Notice that the coherence for the skewed input is not as

good as the measurement with consistent input excitation
Good Measurement
Skewed/Angle Impact

Impact Location Effects
Notice that the coherence for the impact around point is not
as good as the measurement with consistent input excitation
Good Measurement
Impact Around Point

Control the Location of Excitation
HAMMER
STRAW
SLEEVE
SWIVEL
JOINT
CLAMP
Adaptor for small impact hammer enables easy

orientation of hammer to impact structure using
swivel joint on small tripod fixture.
Current design uses a straw sleeve adapted to
connector to swivel joint on tripod.

Impact Location
Control of input point

and direction is very
important

Accelerometer Saturation – But No Overload
Sometimes the response transducer may be too sensitive

which generally may cause an overload.
But there are times when the accelerometer and the signal
conditioner may not overload the data acquisition system
BUT may be distorted due to saturation of the signal
conditioner.

Accelerometer Saturation – But No Overload
Accelerometer too sensitive Accelerometer with proper sensitivity

Analyzer ICP / External ICP / DC Accelerometer Comparison
DC 200mv/g DC 1V/g ICP 1V/g

Analyzer ICP / External ICP / DC Accelerometer Comparison
70.00 1.00
F FRF 200 mv DC:+Z/Force - 75%:+Z
F FRF 1 V ICP:+Z/Force - 75%:+Z
F FRF 1 V DC:+Z/Force - 75%:+Z
Amplitude
70.00 1.00
g/lbf
F FRF 200 mv DC:+Z/Force - 75%:+Z

dB
F FRF 1 V ICP:+Z/Force - 75%:+Z

F FRF 1 V DC:+Z/Force - 75%:+Z
20.00 0.00
0.00 Hz 16.00
Amplitude
g/lbf
dB
20.00 0.00
0.00 Hz 1.00

How Hard Should I Hit – Air Capsules
The hammer kits normally have the ability to use a variety of

different tips to customize the input spectrum.
But what happens if some impacts are harder and some are
softer? Does this affect the input excitation spectrum?
Depending on the hammer tip, this can be significant.

Air Capsule Plastic Cap on Hard White Tip Hard White Tip
How Hard Should I Hit – February2010

Time Pulse Time Pulse Time Pulse
S OFT HI T
INCREASI NG I MP ACT HAMM ER FO RCE L EVE L

--------- - 45 dB --------- - 48 dB --------- - 49 dB
20 dB 20 dB
20 dB
2500 Hz
220 Hz
240 Hz

MED IUM HIT
--------- - 41 dB --------- - 43 dB --------- - 41 dB

20 dB
20 dB 20 dB
340 Hz 2700 Hz
300 Hz

HA RD HIT
--------- - 37 dB --------- - 40 dB --------- - 34 dB
8 dB 20 dB
10 dB
400 Hz 2010 – Modal Space Articles

February 400 Hz 3000 Hz

INCREASI NG HAM MER T IP HARDNE SS
Selection of Measurement Locations
So what are the chances that you would pick 9 of the worst
possible measurement locations for a plate ???
August 1998 – Modal Space Articles

Selection of Measurement Locations

The Modal Question
Do these two test yield the same modal information?
SETUP 1
STATIONARY
TRI-AX AT 9 8 4
7
5
1 STATIONARY
IMPACT AT 9
IN Z ONLY
SETUP 2
2
9
6 ROVING IMPACT
IN Z ONLY 8 4
3 1
7
5
2
9
6 ROVING TRI-AX
3

The Modal Question
SETUP 1 SETUP 2
ROVING IMPACT
IN Z ONLY ROVING TRI-AX
STATIONARY z
TRI-AX AT 9 8 STATIONARY
4 8 x
1 IMPACT AT 9 4
z 7 IN Z ONLY 1 y
5 7
x y 5
2
9 2
6 9
1x
3 1y
6 1x
1z 3 1y
2x 1z
2y 2x
THREE 2z 2y
PARTIAL 3x
2z
ROWS 3y
3x
OF 3z ONE
3y
FRFS FULL
3z
COLUMN
9x
OF
9y
FRFS 9x
9z
9y
1x 1y 1z 2x 2y 2z 3x 3y 3z 9x 9y 9z
9z
1x 1y 1z 2x 2y 2z 3x 3y 3z 9x 9y 9z


Reference Selection – Where Should Reference Be Located?
Random Point Selection Organized Point Selection

Composite Plate Pseudo-Repeated Root Example
A plate structure with suspected pseudo-repeated roots was

tested to determine the appropriate reference locations
13 Z
15 Z
3Z

A summation plot and typical drive point FRFS are shown

Using all 3 references, TRIP identifies a repeated root
Note: plot only to 500 Hz

Using references 3 Z & 15 Z, TRIP identifies repeated root

Using references 13 Z & 15 Z, TRIP identifies repeated root

Using references 3 Z & 13 Z, does not ! ! ! ! ! ! !


Shaker Excitation
• Overview shaker testing considerations

• Identify some typical set up concerns
• Provide some examples
IMAC 27 presentation covered excitation techniques

IMAC 29 presentation covered shaker testing considerations

Excitation Configuration
Shaker
force
Stinger structure
sensor
Test Signal
-random
-burst Random
-pseudo-random AUTORANGING AVERAGING
-periodic-random
-Chirp
Power
Amplifier 1 2 3 4
AUTORANGING AVERAGING WITH WINDOW
AUTORANGING AVERAGING
1 2 3 4
1 2 3 4

Reason for Stinger
Purpose of Stinger
• Decouple shaker from test structure
• Force transducer between stinger and structure decouple
forces acting in the axial direction only
• Forces acting in any other direction will be unaccounted
for creating error in the measurements
Bending
Modal Force Gage
Shaker Axial
Stinger
Structure

Possible Problems with Stinger
• Location of stinger on structure may be affected by the

local stiffness and/or structure deformation
Axial stiffness Axial and bending stiffness

Stinger Configuration with Through Hole Shaker
2-part
Force
chuck
sensor
assembly
Test Structure
Modal Exciter
collet
armature stinger

Multiple Input Shaker Excitation
Objectives of this part of lecture:
• Identify some basics of MIMO testing

• Discuss several practical aspects of
multiple input multiple output shaker testing

Multiple Input Shaker Excitation
• Provide a more even distribution of energy

• Simultaneously excite all modes of interest
• Multiple columns of FRF matrix acquired
• More consistent data is collected
• Same test time as SISO case

Excitation Considerations - MIMO
Multiple referenced FRFs are

obtained from MIMO test
Energy is distributed
better throughout the
structure making better
measurements possible
Ref#2 Ref#3
Ref#1

Large or
complicated
structures
require special
attention

Multiple shakers are

needed in order to
adequately shake the
structure with
sufficient energy to
be able to make good
measurements for
FRF estimation

Frequently Asked Questions
Objectives of this part of lecture:
• Provide some measurements to illustrate issues

• Revisit reciprocity
• Compare impedance head vs force/accelerometer
• Compare MIMO measurements

SISO vs MIMO
Excitation technique is one necessary step to acquire better

measurements (random/hann vs burst random).
But using MIMO instead of SISO is another important
consideration.
And mass loading effects are also important
(a) (b) (c) (d)
SISO SISO MIMO MIMO

RANDOM WITH HANNING BURST RANDOM RANDOM WITH HANNING BURST RANDOM

Shaker Mass Loading Effects


MIF
MIF


 
SUM BLOCKS & STABILITY DIAGRAM

SUM BLOCKS & STABILITY DIAGRAM

 
 

Shaker Mass Loading Effects
Three Measurement Setups – Compare Repeated Root:

No Mass Compensation Mass Compensation All Accels Mounted
Accelerometer Equivalent
and Mounting Cube Mass

What’s an impedance head? Why use it? Where does it go?
An impedance head is a transducer that measures both force and

response in one device.
This is a critical measurement for the structure and it is strongly
advised that impedance heads be used in all cases.
A combination of a separate force gage and accelerometer is often
used but time and time again this measurement has been seen to
never be better than that obtained with an impedance head.
The force gage or impedance head needs to be mounted on the
structure side of the stinger arrangement.
2-part chuck Force sensor
assembly
collet
armature stinger

Test Set Up
Measurements taken to show difference in set up
Incorrect Correct
X √
Shaker Shaker
Impedance Quill
Quill Impedance
Structure Structure

What is the proper mounting technique for the force transducer?
X Drive-point √
FRFs
10.00 1.00 10.00 1.00
√
Amplitude
Amplitude
g/N
g/N
dB
dB
-90.00
0.00 Hz
X 400.00
0.00 -90.00
0.00 Hz 400.00
0.00
Distinct difference in drive point FRF based on force configuration !

Drive Point FRF – Stinger Effects
No sleeves
With Sleeves

Drive Point FRF – Stinger Effects
No sleeves
With Sleeves

Differences in Reciprocal Measurements – Impedance vs Accel
Top View
• Offset Accelerometer
• Accelerometer on Other Face of Structure
• Impedance Head
Pt. 2
Measurement
locations
Bottom View Pt. 1
• All reciprocity measurements are between

points 1 & 2 with respect to force from
impedance heads

Reciprocal Measurements
Offset Accelerometer
What if I can
only put the
accelerometer
next to the
force gage?
Accel
Force

Accelerometer on Other Face of Structure
What if I can
only put the
accelerometer
on the face of
the structure
that is opposite
the force gage?
Accel
Force

Impedance Head
What if I have
an impedance
head that
measures force
and acceleration
at the same
place?
Force &
Accel

What is the correct amplitude level for modal testing applications?
The excitation levels for modal testing are usually very low.
There is no need to provide large force levels for conducting a
modal test especially if appropriate response transducers
(accelerometers) are selected with good sensitivity.
The level only need be sufficient to make good measurements.

The excitation levels for modal testing are usually very low.
There is no need to provide large force levels for conducting a
modal test especially if appropriate response transducers
(accelerometers) are selected with good sensitivity.
The level only need be sufficient to make good measurements.
In fact large force levels tend to overdrive the structure and can
excite nonlinear characteristics of the structure and provide overall
poorer measurements than with lower level force tests.

30.00 30.00
(m/s 2)/N
(m/s 2)/N
dB
dB
FRF 2:+Z/2:+Z MIMO FRF 2:+Z/2:+Z MIMO
FRF 2:+Z/2:+Z SIMO FRF 2:+Z/2:+Z SIMO
-50.00 -20.00
1.00 1.00
Amplitude
Amplitude
/
/
0.00 0.80
0.00 Hz 800.00 549.46 Hz 688.14
High excitation level degrades drive point FRF quality!!!

40.00
(m/s 2)/N
dB
FRF 4:-Z/2:+Z MIMO

FRF 4:-Z/2:+Z SIMO
-60.00
1.00
Amplitude
/
0.00
0.00 Hz 800.00
High excitation level degrades drive point FRF quality

and measurements across components may be worse!!!

How many shakers should I use in my modal test?
The number of shakers is often a difficult one to answer.

Basically there are never enough shakers when conducting a large
modal test.
Often we are limited by the total number of shakers available in
the test lab for modal testing.
Usually two shakers are sufficient for many tests. Sometimes
three or four shakers are needed for much larger structures.
But generally more than five shakers are rarely used.
The main point is that there needs to be enough shakers acting as
reference locations that are positioned so that all of the modes of
the structure are adequately excited and good frequency response
measurements are obtained.

How many shakers should I use in my modal test?
10.00 0.00
10.00 1.00
( m/s 2)/N
(m/s 2)/N
Amplitude
dB
dB
dB
SISO MIMO
-70.00 0.00
0.00 Hz 800.00 -80.00 -100.00
0.00 Hz 800.00
Single input may not be able to provide accurate FRFs

Why bother with MIMO testing? Why not run a SISO instead?
Single shaker testing is adequate providing all the modes of the

structure can be sufficiently excited and measured.
In component testing this can often times be sufficient.
However, when structures have several components, then the ability
to provide sufficient excitation to acquire good measurements
across the whole structure may be difficult.
Tests can be conducted with a single shaker that is moved to
different reference locations but generally this does not provide
consistently related measurements.
When this is the case (as it often is), then MIMO is needed.

FRFs Using SIMO vs

FRFs Using MIMO
Blue Shaker is Reference for SIMO & MIMO

20.00
30.00 10.00
FRF 1:+Z/2:+Z MIMO
FRF 1:+Z/2:+Z SIMO
(m/s 2)/N
(m/s 2)/N
(m/s 2)/N
dB
dB
dB
FRF 2:+Z/2:+Z MIMO

FRF 2:+Z/2:+Z SIMO
FRF 3:+Z/2:+Z MIMO

FRF 3:+Z/2:+Z SIMO
-60.00
0.00
Am
-80.00
0.00 -50.00
0.00
Am
Am

15.99 0.96
SIMO
( m/s 2)/N
Amplitude
dB
-9.05 0.71
607.85 Hz 696.46

Why bother with MIMO testing? Why not run 3 SISO instead?

Why bother with MIMO testing? Are modal results better?
SIMO Tests Combined – Overall Stabilization Plot
100th order
polynomial

SIMO Tests Combined Stabilization Plot (Close-up)
100th order
polynomial

Individual SIMO Tests Stabilization Plots (Close-up)
Reference Point 1 Reference Point 2
100th order
polynomial
Reference Point 3
SIMO Tests Combined – Overall Stabilization Plot
100th order
polynomial

SIMO
1 Volt
MIMO
0.1 Volt 0.1 Volt

0.1 Volt

MIMO Test – Overall Stabilization Plot
100th order
polynomial

MIMO Test Stabilization Plot (Close-up)
100th order
polynomial

Experimental Modal - Considerations and Use
Several items are very important
• Test Setup – boundary conditions

• Excitation Methods – accurate measurements
• Parameter Extraction – accurate parameters
• Dynamic Model Development
(some personal notes from experience)

Things to Consider
Test Setup
• Pre-Test helps in so many ways but be careful to not fully

rely on the model to be correlated
• Do everything possible to make the best possible
measurements (or the original sin results)
• Make sure that all measurements are consistent
• Be aware of all boundary conditions such as support
structure, shaker stinger interaction, instrumentation
effects on structure

Things to Consider
Test Setup
• Check for overloads and underloads of transducer

• Check for saturation of signal conditioning
• Check linearity of structure
• Check mass loading effects of transducers
• Check frequency shifts due to support condition
• Check frequency resolution for measurements

Things to Consider
Measurements
• Check every measurement including input/output time

traces, power spectrum, frequency response function and
coherence
• Check reciprocity where possible
• Repeat drive point measurements on test that require
multiple sets of data to completely describe all points on
the structure

Things to Consider
Impact Technique
• Check FRF with different tips, over different frequency

ranges with different resolutions
• Maintain consistent force level for measurements
• Impact the same point in the same direction for each
measurement
• Compare different number of averages to determine
convergence to FRF

Things to Consider
Shaker Excitation Technique
• Check force/quill alignment to prevent any overturning

moments on force gage
• Check reciprocity on MIMO tests
• Try multiple excitation techniques to determine what
technique works best
• Check drive point FRFs when multiple banks of data are
collected

Excitation Techniques
Do’s and Don’ts
Peter Avitabile Marco Peres
UMASS Lowell The Modal Shop
Impact Animation


Excitation Techniques IMAC 30 by Pete Avitable

Uploaded by

Copyright:

Available Formats

You might also like

Excitation Techniques IMAC 30 by Pete Avitable

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Excitation Techniques IMAC 30 by Pete Avitable

Uploaded by

Copyright:

Available Formats

Excitation Techniques

Do’s and Don’ts

1 Dr. Peter Avitabile

Objectives of this lecture:

• Overview impact testing considerations – part 1

• Identify some typical concerns

IMAC 27 presentation covered shaker excitation techniques

2 Dr. Peter Avitabile

Could you explain and how is it

Illustration by Mike Avitabile Illustration by Mike Avitabile Illustration by Mike Avitabile

Actual time signals

Digitized time signals

Compute FRF and Coherence

4 Dr. Peter Avitabile

x(t) h(t) y(t) TIME

INPUT SYSTEM OUTPUT FFT & IFT

x(t) - time domain input to the system

Sx(f) - linear Fourier spectrum of x(t)

Sy(f) - linear Fourier spectrum of y(t)

H(f) - system transfer function

5 Dr. Peter Avitabile

Rxx(t) Ryx(t) Ryy(t) TIME

INPUT SYSTEM OUTPUT FFT & IFT

Ryy(t) - autocorrelation of the output signal y(t)

Ryx(t) - cross correlation of y(t) and x(t)

Gxx(f) - autopower spectrum of x(t) G xx ( f )  Sx ( f )  S*x ( f )

Gyy(f) - autopower spectrum of y(t) G yy ( f )  Sy ( f )  S*y ( f )

Gyx(f) - cross power spectrum of y(t) and x(t) G yx ( f )  Sy ( f )  S*x ( f )

6 Dr. Peter Avitabile

Sy S*x HSx S*x H 

Sy S*y HSx S*y H 

Measurements - Frequency Response Function Measurements - FRF & Coherence

AVERAGED INPUT AVERAGED CROSS AVERAGED OUTPUT

POWER SPECTRUM POWER SPECTRUM POWER SPECTRUM

0Hz AVG: 5 200Hz

0Hz AVG: 5 200Hz

FREQUENCY RESPONSE FUNCTION

8 Dr. Peter Avitabile

9 Dr. Peter Avitabile

10 Dr. Peter Avitabile

Objectives of this lecture:

• Overview impact excitation techniques

11 Dr. Peter Avitabile

An impulsive excitation which is very short in the time window

12 Dr. Peter Avitabile

13 Dr. Peter Avitabile

The force spectrum can be customized to some extent

However, the hammer tip alone does not totally determine

14 Dr. Peter Avitabile

METAL TIP HARD PLASTIC TIP

-976.5625us 123.9624ms -976.5625us 123.9624ms

SOFT PLASTIC TIP RUBBER TIP

15 Dr. Peter Avitabile

If the leading portion of the time pulse is not captured then

16 Dr. Peter Avitabile

Double impacts can occur due to a sloppy hammer swing or

DOUBLE IMPACT DOUBLE IMPACT

-976.5625us 998.53516ms -976.5625us 998.53516ms

Sy Sx HSx Sx H 

Sy Sy HSx Sy H 