Characterization of laser doppler vibrometers using acousto-optic modulators


ACTA IMEKO | www.imeko.org December 2020 | Volume 9 | Number 5 | 361 

ACTA IMEKO 
ISSN: 2221-870X 
December 2020, Volume 9, Number 5, 361 - 364 

 
 

CHARACTERIZATION OF LASER DOPPLER VIBROMETERS USING 

ACOUSTO-OPTIC MODULATORS 
 

Michael Gaitan1, Jon Geist2, Benjamin J. Reschovsky3, Ako Chijioke4 

 
1 NIST, Gaithersburg MD, USA, michael.gaitan@nist.gov  

2 NIST, Gaithersburg MD, USA, jon.geist@nist.gov 
3 NIST, Gaithersburg MD, USA, benjamin.reschovsky@nist.gov 

4 NIST, Gaithersburg MD, USA, akobuije.chijioke@nist.gov 

 

Abstract: 

We report on a new approach to characterize the 

performance of a laser Doppler vibrometer (LDV). 

The method uses two acousto-optic modulators 

(AOMs) to frequency shift the light from an LDV 

by a known quantity to create a synthetic velocity 

shift that is traceable to a frequency reference. 

Results are presented for discrete velocity shifts and 

for sinusoidal velocity shifts that would be 

equivalent to what would be observed in an ideal 

accelerometer vibration calibration. The method 

also enables the user to sweep the synthetic 

vibration excitation frequency to characterize the 

bandwidth of an LDV together with its associated 

electronics. 

Keywords: Laser Vibrometer; Vibration 

Calibration; Acousto-Optic Modulator; 

1. INTRODUCTION 

Following ISO Standard 16063-41 Method for 

the calibration of vibration and shock transducers – 

Part 41 Calibration of laser vibrometers [1], laser 

Doppler vibrometers (LDVs) are calibrated by a 

comparison-type measurement to a laser homodyne 

interferometer that is defined as the primary 

standard. Not covered by ISO 16063-41, but for the 

case where all the components of the LDV system 

and their associated uncertainties are known, 

methods can be employed for the direct 

determination of measurement uncertainty of the 

LDV [2], [3] or by using a combination of a 

heterodyne with a homodyne-quadrature 

configuration [4]. 

The technology for manufacturing commercial 

LDV systems has matured as well as their use in 

commercially available primary vibration 

calibration systems. These systems require 

calibration by the manufacturer over a periodic time 

interval that is typically one year and are traceable 

to the Système International d'Unités (SI) through 

the manufacturer. From the end user perspective, 

the commercially manufactured LDV system is like 

a “black box”, meaning that the design and internal 

components of the LDV are not known in detail by 

the user. Therefore, following an uncertainty 

determination approach described in [2] or [3] is not 

possible for such commercial black box systems 

especially if their internal workings are considered 

proprietary by the manufacturer. The only 

possibility provided by the standard for calibrating 

a such commercial LDV systems is therefore to 

follow ISO 16063-41 and compare it to a primary 

heterodyne system, resulting in the LDV system 

being considered a secondary system. 

A challenge therefore remains in the adoption of 

cost-effective commercial LDV systems by 

National Measurement Institutes (NMIs) who are 

responsible for direct determination of uncertainty. 

Towards this end, we have recently reported on the 

calibration of laser heterodyne velocimeters using 

shock excitations and total distance travelled [5], [6]. 

One advantage of that method is that it characterizes 

the entire measurement system under the same 

conditions that would be used in an accelerometer 

shock calibration and could as well be included as 

part of the accelerometer shock calibration. 

However, a drawback of the method is that it does 

not characterize the frequency response and 

bandwidth of the LDV. The bandwidth of the 

excitation must not exceed the bandwidth of the 

LDV in order to produce accurate accelerometer 

shock calibrations. This drawback motivated us to 

develop a new method to characterize the 

bandwidth of the LDV system and resulted in the 

method that we present in this report. 

2. EXPERIMENTAL DESIGN 

Figure 1 shows a diagram and photograph of the 

acousto-optic modulator (AOM)-based LDV 

characterization system that we have developed. 

The laser light is first collimated using 300 mm and 

30mm lenses to create a beam diameter that is 

compatible with the aperture of the AOMs. The 

http://www.imeko.org/
mailto:michael.gaitan@nist.gov
mailto:jon.geist@nist.gov
mailto:benjamin.reschovsky@nist.gov
mailto:akobuije.chijioke@nist.gov


ACTA IMEKO | www.imeko.org December 2020 | Volume 9 | Number 5 | 362 

beam passes through the 1st AOM where it is 

downshifted in frequency f. Next it passes through 

the 2nd AOM where it is upshifted by a frequency 

f+. The beam is then reflected back along its path 
with a mirror, doubling the effect of the AOMs, and 

delivering light to the LDV that is shifted in 

frequency by: 

2𝛿 = 2 (𝑓 + 𝛿 − 𝑓). (1) 

The reason why we use two AOMs to produce 

the frequency shift is that a single AOM cannot 

generate frequency shifts of order 1 MHz or less as 

required. The velocity v reported by the LDV is 

related by the Doppler equation with a frequency 

shift of 2𝛿 and the wavelength of the laser by the 
relationship: 

𝑣 =  ½ 𝜆 (2𝛿) =  𝜆 𝛿 , (2) 

where the laser wavelength 𝜆 = 632.81 nm [7]. 

3. RESULTS 

The experimental results that we report were 

obtained using [8] a commercially available LDV 

system that includes a Polytec OFV-503 Sensor 

Head, a OFV-5000 Vibrometer Controller, and a 

Polytec Data Management System that were 

interfaced with the design shown in Figure 1. The 

AOMs (Brimrose TEF-110-50-633) were driven by 

two National Instruments PXIe-5650 sinusoidal 

radio frequency (RF) signal generators. An Agilent 

3458A Digital Multimeter was also used to measure 

root mean squared (RMS) voltage for sinusoidal 

excitations. The signals from the RF signal 

generators were amplified using Mini-Circuits 

ZHL-2010+ RF amplifiers connected to each of the 

AOMs. The base frequency f for our results was 

selected to be 110 MHz, corresponding to the center 

frequency of the AOMs. The zero frequency (DC) 

and transient velocity readings that we report were 

obtained using the Polytec Data Management 

System. The RMS readings that we report were 

obtained using the RMS multimeter. 

3.1. Results for fixed frequency shift 

Figure 2 shows the relationship between the 
frequency shift 𝛿, the reported velocity from the 
LDV, and the calculated velocity using the Doppler 

equation (2). These results were obtained with 

vibrometer controller set to VD-09 with a 

corresponding amplification factor of 0.5 m/s/V.  

Figure 1: Diagram (upper) and photograph (lower) of the acousto-optic modulator (AOM)-based LDV characterization system. 

http://www.imeko.org/


ACTA IMEKO | www.imeko.org December 2020 | Volume 9 | Number 5 | 363 

 
Figure 2: Plot of the velocity reported by the LDV as a function 

of frequency shift 𝛿 in blue and the corresponding calculated 
values using the Doppler equation (2). 

The resulting dc voltage was sampled 2048 times 

at 204800 samples/s with no filtering using the 

Polytec Data Management System. The data were 

averaged, and the standard deviation was 

determined. The LDV exhibited a 0.0012 m/s offset 

when directed onto a non-moving surface without 

the AOMs in the beam path as well as when the 

frequency shift 𝛿 was set to zero. This offset was 
subtracted from the measured results depicted in the 

figure. 
The data in Figure 2 is replotted in Figure 3 in 

terms of the percent difference between the velocity 

reported by the LDV (with the offset subtracted) and 

the calculated velocity from the Doppler shift 

equation (2). The data show a maximum percent 

difference of  0.04 % over the frequency range that 

was tested.  

 
Figure 3: Percent difference between the velocity reported by 

the LDV (with the offset subtracted) and the calculated velocity 

using the Doppler shift equation (2). The manufacturer of the 

LDV reports a 1% uncertainty in their instrument specifications 

while the frequency shift that we can produce with the signal 

generators is orders of magnitude smaller in uncertainty.  

 

3.2. Results for sinusoidal excitation 

In this experiment the 2nd AOM was excited 

using the National Instruments PXIe-5650 RF 

signal generator with a sinusoidal frequency 

modulation to create a synthesized vibration 

measurement. The goal of this experiment was to 

characterize the bandwidth of the LDV system. In 

this experiment, the root mean squared (RMS) 

voltage from the analog output of the Polytec OFV-

5000 Vibrometer Controller was measured using 

the RMS multimeter and converted to RMS velocity 

using the VD-09 gain factor of 0.5 m/s/V. The 

sinusoidal modulation at the 110 MHz base 

frequency was swept from 100 Hz to 3 MHz. Figure 

4 shows that the LDV Vibrometer Controller has a 

uniform response up to 1 MHz and drops off beyond 

that frequency.  

 
Figure 4: Frequency response of the LDV using sinusoidal 

frequency modulation of the 2nd AOM to create a synthesized 

vibration measurement condition. 

3.3. Results for velocity step function excitation 

In this experiment the 2nd AOM was excited 

using an HP 83650B 10 MHz to 50 GHz RF Swept 

Signal Generator to provide a capability to 

frequency modulate an arbitrary analog signal. An 

Agilent 33250A Arbitrary Waveform Generator 

signal generator was used to produce a 1 Hz square 

wave alternating from 0 mV to 300 mV for 

frequency modulation to simulate a step function for 

synthesized velocity. The analog velocity signal 

from the Vibrometer Controller was digitized using 

the Polytec Data Management System set at its 

maximum sampling rate of 204800 samples/s. The 

resulting response shown in Figure 5 includes the 

effects of the LDV as well as the digital acquisition 

system, which would be expected to have a 

maximum bandwidth of 102400 Hz. 

http://www.imeko.org/


ACTA IMEKO | www.imeko.org December 2020 | Volume 9 | Number 5 | 364 

4. SUMMARY 

Our results show that this approach is 

immediately useful as a tool for characterizing DC, 

sinusoidal steady state, and transient response of an 

LDV as a system as a whole, together with its data 

acquisition and control electronics and amplifiers. 

The DC response we reported exhibited a maximum 

of  0.04 % difference between the measured value 

and the calculated value based on the Doppler 

equation, which is in good agreement to what we 

had reported earlier using shock excitations and 

total distance travelled [5], and is well within the  

1 % accuracy specified by the manufacturer. The 

LDV system bandwidth of 1 MHz determined by 

sinusoidal excitations is in good agreement to what 

the manufacturer specifies. Lastly, the velocity step 

function experiment serves as an example that it is 

possible to create complex velocity profiles to test 

the response of the LDV together with its data 

acquisition, control electronics, and amplifiers. Our 

future work is focused on further improvements on 

the system design and carrying out a full uncertainty 

analysis. One improvement that we envision in the 

design is measure the frequency shift 𝛿 with a 
photodiode rather than reading it from the signal 

generators driving the AOMs. This could capture 

any offsets or fluctuations between the signal 

generators and the light entering the LDV, e.g. due 

to refractive index fluctuation. We anticipate that 

after further investigation this method can be used 

as a tool for primary calibration of Laser Doppler 

Vibrometers.  

5. REFERENCES 

[1] ISO Standard 16063-41 Method for the calibration 
of vibration and shock transducers – Part 41 

Calibration of laser vibrometers, ISO 16063-

41:2011(E). 

[2] G. Siegmund, “Sources of Measurement Error in 
Laser Doppler Vibrometers and Proposal for 

Unified Specifications”, Proc. of SPIE Vol. 7098, 

70980Y-1, 2008. 

[3] N. Vlajic, A. Chijioke, “Traceable dynamic 
calibration of force transducers by primary means,” 

Metrologia 53, S136–S148, 2016. 

[4] T. Bruns, F. Blume, A. Täubner, “Laser Vibrometer 
Calibration at High Frequencies using 

Conventional Calibration Equipment”, XIX 

IMEKO World Congress, Lisbon, Portugal, 

September 6-11, 2009. 

[5] M. Gaitan, M. Afridi, J. Geist, “On the Calibration 
of Laser Heterodyne Velocimeters Using Shock 

Excitations and Total Distance Travelled”, IMEKO 

XXII World Congress, Belfast, 2018. 

[6] M. Afridi, J. Geist, M. Gaitan, “Primary Calibration 
of the Low Frequency Response of a Laser 

Heterodyne Velocimeter Used in a Pendulum-

Based Shock Excitation System by SI Traceable 

Distance Measurement”, J Res Natl Inst Stan, In 

Press.  

[7] ISO Standard 16063-11 Methods for the calibration 
of vibration and shock transducers – Part 11 

Primary vibration calibration by laser 

interferometry. 

[8] Certain commercial equipment and instruments are 
identified in this article in order to describe the 

experimental procedure adequately. Such 

identification is not intended to imply 

recommendation or endorsement by the National 

Institute of Standards and Technology, nor is it 

intended to imply that the materials or equipment 

identified are necessarily the best available for the 

purpose.

 
 

Figure 5: Response of the LDV system to a synthesized velocity 

step produced by frequency modulation of a 1 Hz square wave 

on the 2nd AOM. The time depicted on the x-axis has been 

magnified to observe the transient of the velocity step. 

 

http://www.imeko.org/