Anda di halaman 1dari 10

Optics and Lasers in Engineering 54 (2014) 187196

Contents lists available at ScienceDirect

Optics and Lasers in Engineering


journal homepage: www.elsevier.com/locate/optlaseng

Recent advances in dental optics Part II: Experimental tests for a new
intraoral scanner
Silvia Logozzo a,n, Ari Kilpel b, Anssi Mkynen b, Elisabetta M. Zanetti a,
Giordano Franceschini a
a
b

Department of Industrial Engineering, University of Perugia, Via Duranti, 931 06125 Perugia (Italy)
Department of Electrical Engineering, University of Oulu, PO Box 8000, FI-90014 Oulu, Finland

art ic l e i nf o

a b s t r a c t

Article history:
Received 21 June 2013
Received in revised form
27 July 2013
Accepted 30 July 2013
Available online 14 September 2013

The object of this paper is testing the performance of a new device for 3D oral scanning: a two channel
PTOF (pulsed time-of-ight) laser scanner, designed for dental and industrial applications in the
measurement range of zero to a few centimetres. The application on short distances (010 cm) has
entailed the improvement of performance parameters such as single shot precision, average precision
and walk error up to mm-level and to mm-level respectively.
The single-shot precision (s-value) has resulted to range from 43 to 63 ps (910 mm), having
considered the measurement range (6.510 mm) corresponding to 12 V signal; this result agrees well
with estimates made from simulations. The average precision has resulted to be dependent on the
number of measurements and can reach a value equal to 7 25 mm, whenever the measurements
frequency is sufciently high. For example, if the required scanning speed is 1000 points/s and the
required average precision is 725 mm, then a pulses frequency of 3050 MHz is needed, considering
signal amplitude varying between 12 V.
On the whole, the performance of this new device, based on PTOF has proven to be adequate to its
employment in the eld of restorative dentistry.
& 2013 Elsevier Ltd. All rights reserved.

Keywords:
Intraoral scanner
Pulsed time of ight
Laser range-nding techniques
Restorative dentistry
CAD/CAM

1. Introduction
Three-dimensional scanning of the mouth is required in a large
number of procedures in dentistry such as restorative dentistry
and orthodontics. The aim of the 3D mapping of the oral cavity is
to create digital impressions. Restorative dentistry is of course the
main eld which requires the application of very accurate 3D
intraoral scanners. In fact, the realization of any dental prosthesis

Abbreviations: APD, avalanche photodiode; B, bandwidth; CAD/CAM, computer


aided design/computer aided manufacturing; CFD, constant fraction discriminator;
F(M), excess noise factor of APD; G, laser gain; GBW, gain-bandwidth product;
, wavelength; LVPECL, low voltage positive emitter-coupled logic; LVTTL, low
voltage transistor-transistor logic; M, multiplication factor of APD; NEP, noise
equivalent power of APD; PD, photo-diode; PECL, positive emitter-coupled logic;
PTOF, pulsed time of ight; q, charge of electron (1.6  1019 C); R(), spectral
responsivity of APD; rms, root-mean-square; RX, receiver; S/N or SNR, signal-tonoise ratio; SR, slew rate; SSP, single shot precision; TDC, time-to-digital converter;
TOF, time-of-ight; tr, rise time (1090% of full pulse amplitude); tf, fall time
(9010% of full pulse amplitude); TX, transmitter; Vn  amp, noise of electronics
(ampliers); Vout, voltage of the pulse in the timing point; Z, gain of a transimpedance amplier
n
Correspondence to: Via Curtatone e Montanara, 2-06129 Perugia, Italy.
Mobile: 39 348 7142939.
E-mail addresses: sililog@hotmail.com (S. Logozzo), arik@ee.oulu. (A. Kilpel),
anssi.makynen@ee.oulu. (A. Mkynen), elisabetta.zanetti@unipg.it (E.M. Zanetti),
giordano.franceschini@unipg.it (G. Franceschini).
0143-8166/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.optlaseng.2013.07.024

requires building the three-dimensional mathematical model of


dentition, performing a reverse engineering procedure; the prosthesis can be designed on this model and it can be realized by
means of CAD/CAM systems [1].
This work is focused on a new intraoral PTOF laser scanner [2];
previous articles [3, 4] have reported an extensive review of existing
techniques for reverse engineering; PTOF can represent an interesting alternative because these systems are capable of covering a large
volume of the scene in a single image or few images and accuracy
does not depend by target distance, unlike other sensors.
However the application of PTOF to dentistry implies a measurement range of zero to a few centimetres; therefore the device
employed for industrial measurements needs to be redesigned in
order to achieve improved single-shot precision and accuracy.
This work introduces this new system and describes its main
components; the performances of this new system are rst calculated on the basis of analytical formula, and then are experimentally
tested on a prototype. The possible impact of this new device is
nally discussed.

2. Materials and methods


This paragraph summarizes main steps followed for the design
of the 3D PTOF laser scanner.

188

S. Logozzo et al. / Optics and Lasers in Engineering 54 (2014) 187196

Table 1
Target performance parameters of the new PTOF
system.
Performance parameters
Minimum scanning speed
AP
Walk error
Measurement range
Mean measurement distance

1000 pts/s
7 25 mm
7 25 mm
110 cm
5 cm

2.1. Specications
The required performance parameters (Table 1), which could
make the new PTOF laser rangender suitable for dental applications, have been established on the basis of the performances
exhibited by the other intraoral scanning devices [1,5] and on the
basis of the requirements, in scientic literature.

The main specications of Thorlabs APD110A2/M are disclosed


in Table 2.
The main performance parameters of an APD are the multiplication factor M and the excess noise factor F(M), and are dened
by the following equations:
M
1

1
 r
V
V br



1
1kr
FM kr M 2
M

The basic structure of the PTOF laser range nder is presented


in Fig. 1 and the components are listed below:
1. Oscillator or signal generator (STANDFORD RESEARCH SYSTEM,
model DG535);
2. noise generator (own designed circuit);
3. oscillator or signal generator (AVTECH AVPP-1-C);
4. laser driver or pulse generator circuit (own designed circuit);
5. START timing discriminator (own designed circuit);
6. pigtailed laser diode (Thorlabs LP660-SF40, 660 nm and Thorlabs
LP405-SF10, 405 nm);
7. pigtailed aspheric collimator (Thorlabs CFS2-532-FC);
8. tube;
9. 50/50 glass beam splitter;
10. Visible Achromatic Doublet Pairs (Thorlabs MAP107575-A);
11. target;
12. Receiver Avalanche photodiode unit (Thorlabs APD110A2/M),
which embeds a transimpedance preamplier and a current
feedback postamplier;
13. STOP timing discriminator (Analog Devices EVAL-ADCMP 573
and own designed circuit);
14. TDC circuit with evaluation kit (ACAM ATMD-GPX Evkit);
15. IBM-compatible PC.
This is a two channel PTOF system with a single-axis optic.
2.3. Electronics
The aim of the jitter circuit (or noise generator circuit) is
avoiding the synchronization of the oscillators of the transmitter
and of the time measuring unit, by adding random noise to the
oscillator of the transmitter. If the single-shot measurements are
repeated n times and the oscillators, used in the transmitter and in
the time measuring unit, are not synchronous (the single-shot
measurements are independent of each other), the precision of the
p
averaged result is improved by a factor equal to n, compared to
the single-shot precision.
Two laser diodes have been chosen to be tested in this work:
Thorlabs LP405-SF10 and Thorlabs LP660-SF40. The laser driver
circuit has been home-made. The circuit has been fed with the
input signal generated by the oscillator.
The receiver channel is composed of the avalanche photodetector Thorlabs APD110A2/M which embeds a silicon APD and a
ultra-low noise transimpedance preamplier and a current feedback
post-amplier.

where V is the bias voltage, Vbr is the breakdown voltage, r is a


coefcient depending on the material [6], and kr is the ionization
ratio, depending on the material [7].
The APD delivers an output voltage, which is a function of
incident light power Popt, detector responsivity R(), multiplication
factor M and transimpedance gain Z, given by
V out P opt RMZ V

2.2. Main components

The timing discriminator bears the following functions:


it changes the analog pulses to logic level pulses, which are fed
to the time measuring unit,
it separates the timing pulses from the noise pulses.
The most important property of the timing discriminator is to
keep the timing event at the same point independently from the
amplitude of the incoming pulse.
The pulse shape in laser range nders usually remains stable
and only the amplitude changes as a function of the distance and
of the reectivity of the target [2]. The pulse shape may change if
the target surface is almost parallel to the measurement beam;
the measurement beam hits two targets located at different
distances at the same time; and
the amplier or the attenuator has a nonlinear behavior.
The performance of a timing discriminator can be expressed in
terms of its walk error, drift and precision.
The most critical is usually the walk error, produced by changes of
the averaged amplitude or of the shape of the incoming pulses. If the
walk error is small enough and the dynamic range of the input pulse
amplitude does not overcome the linear range of the ampliers, an
adjustable electrical or optical attenuator is not needed.
The STOP CFD is composed of the comparator Analog Devices
ADCMP573, its evaluation board (Analog Devices EVAL-ADCMP
573), working as a timing comparator, and of a home-made noise
comparator [8].
The time to digital converter circuit is ACAM AM-GPX with the
evaluation kit ACAM ATMD-GPX. The ATMD-GPX evaluation system consists of a motherboard together with the AM-GPX plug-in
module, mounted in a metal case. It is connected to the ATMD-PCI
interface card (mounted in the PC) by a SCSI-type cable [9].
The TDC AM-GPX has four operating modes: I, G, R and
M-mode. The laboratory tests have been performed using R-mode
and sometimes M-mode when the desired minimum frequency per
channel was 1 MHz and the best resolution was to be achieved.
According to [10], the features of each mode are the following:
R-mode
2 channels with 27 ps resolution.
Differential LVPECL inputs, optional LVTTL.
Measurement range 0 s up to 40 ms.
5.5 ns pulse-pair resolution with 32-fold multi-hit capability 182 MHz peak rate.

S. Logozzo et al. / Optics and Lasers in Engineering 54 (2014) 187196

189

Fig. 1. Schematics of the new PTOF rangender for dental applications.

Trigger to rising or falling edge.


40 MHz continuous rate per channel.
40 MHz continuous rate per chip.
M-mode
2 channels with 10 ps resolution (70 ps peakpeak).
Differential LVPECL inputs.
Measurement range 0 ns up to 10 ms.
Trigger to rising or falling edge.
Max. 500 kHz continuous rate per channel.
Max. 1 MHz continuous rate per chip.

2.4. Optics
The optics is single axis and it is composed of the following
parts:

Pigtailed aspheric collimator (Thorlabs CFS2-532-FC);


50/50 glass beam splitter;
Visible Achromatic Doublet Pairs (Thorlabs MAP107575-A); and
Tubes, in order to change the focused distances of the transmitter and the receiver.

190

S. Logozzo et al. / Optics and Lasers in Engineering 54 (2014) 187196

A single-axis optic has been chosen for the laser range nder
because of its high linearity and its wide measurement range.
Moreover, the single-axis optics, when properly adjusted, exhibits
a very limited variation of the optical power.
The optics schematic is shown in Fig. 2. The laser beam is
transmitted by means of the pigtailed aspheric collimator. Two
tubes are used as a reference, in order to change the focused
distance of both the transmitter and the receiver.
The performance parameters of the optics have been calculated
and also simulated by means of software, implemented at the
Electronics Laboratory of the University of Oulu, on the basis of the
theory reported in [11]. The name of this software is SNR Calculation Program for Laser Radar (2008) by K. Mtt.
The calculations and simulations have been performed considering two kinds of targets: a paper screen and a tooth. The
reectance index was set at 0.17 at 405 nm and at 0.45 at 660 nm
for the teeth enamel and to 1 for the paper. Fig. 3 shows the
average reection spectrum of the teeth enamel at several wavelengths [12].
The used beam splitter is 50/50, so the transmittance index is
0.25. The index for the rest of the losses has been set at 0.15, on the
basis of the measured value of the optical power. So the total
transmittance index of the optics is 0.21. The optical amplitude has
been measured by an oscilloscope connected to an optical converter. The optical power of the laser pulse has been measured
using a power meter. The measured peak optical power has
resulted to be equal to about 30 mW for the blue laser, and to
40 mW for the red laser.

The following diagrams are the results of the simulation of the


optics, by means of the software SNR Calculation Program for Laser
Radar (2008).
Basically, the aim of the simulations has been to forecast the
amplitude variation ratio, when the distance of the target changes
in a range of about 10 mm (5 mm backward and 5 mm forward
from the in-focus position).
As Fig. 2 shows, some parameters can be tuned, in order to
improve the performance of the optics.
p2 is the distance from the transmitter to the rst lens of
the pair;
p1 is the distance from the receiver optics to the rst lens of
the pair;
q2 is the focused distance of the transmitter;
q1 is the focused distance of the receiver;
d is the distance between the lenses of the pair;
TX is the diameter of the optic ber at the transmitter; and
RX is the diameter of the optic ber at the receiver.
The system has been simulated considering the Visible Achromatic Doublet Pairs Thorlabs MAP107575-A as a single lens, with a
focal length equal to the combination of the focal lengths of the
single lenses. The efciency of the lenses has been estimated to be
equal to the experimentally measured transmission index of the
optics. The diameter of sensor's active area has been calculated

Table 2
Specications Avalanche Photo-detector Thorlabs APD110A2/m.
Specication

Value

Multiplication factor, M
Detector active area diameter
Operating wavelength
Responsivity at 405 nm and M 1
Responsivity at 660 nm and M 1
Max responsivity at 800 nm and M 50
Max input power
Transimpedance gain Z
Max output voltage
Power supply

50
1 mm
2001000 nm
0.17 A/W
0.45 A/W
25 A/W
1 mW
50,000 V/A
3.6 V
7 12 V at 200 mA

Fig. 3. Average reectance factor of enamel. Bars indicate standard error at a few
wavelengths.

Fig. 2. Schematic of optics.

S. Logozzo et al. / Optics and Lasers in Engineering 54 (2014) 187196

taking in account the optic ber's diameter (0.0625 mm), the


numerical aperture of the ber and the sizes of the FC connector.
The results of these simulations have been checked against
experimental tests, as reported in the Results section.
The rst simulation has been performed considering the blue
laser as a source and the paper screen as a target. The used
parameters are listed in Fig. 4. The same gure shows that the
maximum received optical power reaches about 7200 nW, when
q1 q2 7.6 cm. The correspondent maximum voltage amplitude
is 3.08 V (see Eq. (3)).
If the position of the target changes in a range of 7 5 mm from
the focused distance, the minimum received power is about
400 nW and the correspondent minimum voltage amplitude is

191

0.17 V, which is too low, while the amplitude variation ratio is too
high (3.08/0.17). According to this simulation, the distance variation range cannot exceed 72 mm from the focused distance.
In this situation, the minimum optical power is about 2600 nW
and the corresponding amplitude is 1.11 V; the amplitude variation
ratio is about 2.8 (3.08/1.11).
The second simulation has been performed using the red laser
and the paper screen as a target. The used parameters are reported
in Fig. 5; the same gure shows that the maximum received
optical power reaches about 1250 nW, when q1 is 7.4 cm and q2 is
8 cm. The correspondent maximum amplitude is 1.41 V see Eq. (3).
If the position of the target changes in a range of 75 mm from the
focused distance, then the minimum received power is about

Fig. 4. First simulation's parameters and results.

Fig. 5. Second simulation's parameters and results.

192

S. Logozzo et al. / Optics and Lasers in Engineering 54 (2014) 187196

250 nW and the correspondent minimum amplitude is 0.23 V, that


is too low, while the amplitude variation ratio is too high (1.41/
0.23). According to this simulation, the distance variation range
should be limited to about 72 mm from the focused distance.
In this situation, the minimum optical power is about 550 nW and
the corresponding amplitude is 0.62 V. So the amplitude variation
ratio is about 2.3 (1.41/0.62)

The excess noise factor F(M) of the APD has been calculated
according to Eq. (2) and other experimental equations reported in
[13], considering the ionization ratio kr 0.02; in these conditions,
F(M) is about 3.
The slew rate (SR) of the signal can be calculated as
SR

V outampl
V =s
tr

where
2.5. Evaluation of single shot precision, pulses frequency and average
precision
The total single shot precision tTOT can be calculated as the
square root of the sum of squared single shot precisions tptof,
due to the PTOF system and tTDC, due to the TDC circuit:
q
t tot t 2ptof t 2TDC s
4
tTDC is at least about 40 ps using R-mode, as experimentally
measured, whereas tptof must be calculated; considering the case
of APD, the following equation holds:
s s
v2nampl 2qV outCFD MFMbZ
v2nrms

s
5
t ptof
SR2
SR2
where:

tr can be calculated as
tr

0:35
s
B

vn  amp is the root mean square value of the electronics noise and
can be calculated as follows:
p
vnamp NEP MRZ B V
8
where:

vn  rms rms value of the total noise in the timing moment [V];
vn  amp rms value of the electronics noise [V];
Vout  CFD pulse voltage in the timing point [V];
M multiplication factor of the APD;
F(M)excess noise factor of the APD;
B signal bandwidth [Hz];
Ztransimpedance gain of the amplier, in this system it is
equal to 50 k (Table 2);
q electron charge [C]; and
SR slew rate of the signal [V/s].

Vout CFD can be set equal to the product of the threshold voltage
level at CFD and the received amplitude voltage. The threshold
voltage level at CFD is 0.4 V (Tables 3), that is the 50% of the
Vout  CFD, according to the LVPECL logic levels ((2.41.6 V)/2).
The multiplication factor M of the APD is 50 (Table 2), the signal
bandwidth B is 50 MHz (Table 3).

Table 3
Input and calculated performance parameters of the new PTOF system.
Performance parameters
M multiplication factor of APD
F(M) noise excess factor of APD
R(6 6 0) APD responsitivity at 660 nm
q electron charge
Vn  amp
Vout  ampl
Vthreshold  CFD
Vout  CFD
B bandwidth
Z
SR
tTDC
tptof
tTOT
Target AP in mm
Target AP in time
n (number of measurements needed)
Scanning speed
Minimum pulses frequency

tr is the pulse rise time [s],


Vout  ampl is the output voltage at the receiver's amplier. As an
approximation, it can be estimated to be equal to the 80% of the
received amplitude voltage; a more precise estimation can be
reached a posteriori measuring Vout  ampl with an oscilloscope,
after having built a physical prototype.

50 (Table 2)
3 Eq. (2)
0.45 A/W (Table 2)
1.6  10  19C
1.87 mV (measured)
2V
0.4 V
0.8 V
50 MHz
50 k (Table 2)
228.57 MV/s (measured)
40 ps (measured)
43.6 ps Eq. (5)
59.2 ps Eq. (4)
50 m (Table 1)
0.33 ps (50  10  6  2/3  108)
31,540 Eq. (9)
1000 points/s (Table 1)
31.5 MHz

R() is the APD responsivity at a certain wavelength


[A/W];
p
NEP is the noise equivalent power of the APD [W= Hz].
The measured and calculated values of vn  amp were almost
coincident and they were equal to 1.87 mV.
As already mentioned, if the single-shot measurements are
repeated n times and the oscillators used in the transmitter and in
the time measuring unit are not synchronous (the single-shot
measurements are independent of each other), the precision of the
p
averaged result is improved by a factor n, compared to the
single-shot precision. Thus, if the value of the desired average
precision (AP) is known in terms of distance, it can be converted to
time multiplying by 2 and dividing by the speed of light
(3  108 m/s); nally, the value of n can be calculated using Eq. (9):
t tot
AP p s
n

where n is the number of the needed measurement replications.


When the value of n and the required scanning speed (points/s)
are known, also the value of the sampling rate can be calculated,
which corresponds to pulses frequency; this last is limited by the
peak TDC frequency per channel.
In the present case, the maximum frequency per channel is
9 MHz.
Table 3 shows the calculated SSP, AP and pulses frequency and
the respective input parameters.
These calculations show that the required performance parameters (Table 1) can be achieved with a pulses frequency of about
30 MHz, when a single receiver channel is used. A multichannel
receiver needs to be used in order to increase the scanning speed
even more, and to decrease the pulse frequency.
2.6. Walk error evaluation
In this paragraph the walk error curve is calculated, referring to
2 V as the maximum received amplitude (Vampl). Assuming the
trailing and leading edges of the input signal are linear ramps, and
their rise and fall times are tr and tf, respectively, the time tcross,
required to reach a voltage level V0 in the output, can be so

S. Logozzo et al. / Optics and Lasers in Engineering 54 (2014) 187196

calculated [14]
r
V of f
20 V 0 1V ampl
1 1 s
t cross D
SR
t t
r

10

where
0 time constant corresponding to the corner frequency of
unit gain (1/GBW);
D delay time constant (the comparator has a propagation
delay after which the input signal is transmitted to the output);
tcross response time required for reaching a voltage level V0 in
the output;
Voff offset voltage;
Vampl peak amplitude of the input voltage;
SR combined slew rate of the leading and trailing edges of the
input pulses in the timing point. It can be calculated from
Eq. (6), as the 80% of Vampl divided by tr.
The rst two terms of the equation represent the contribution
of the propagation delay (and thus of the GBW product of the
comparator) and the third term takes in account the effect of the
offset voltage.
Eq. (10) allows calculating the walk error as the difference
between tcross for a given Vampl and tcross for Vampl 2 V; input values
are summarised in Table 4. The respective curves shown in Fig. 6. It
can be noticed that
fast comparators give a smaller walk error than slow comparators;
faster pulses give smaller walk error than slow pulses;
if a fast comparator and slow pulses are used, a small offset
voltage is sufcient to correct the walk error [2,14]; and
if the pulses are fast, a larger offset voltage is needed.

193

According to the walk error calculated for Voff 3 mV (Fig. 6),


the optimum ratio between the minimum and the maximum
received amplitude is 0.6/2, that is 1/3.33. It means that, if the
received amplitude varies in this range, the walk error can be
expected to be less than 1.8 ps that is less than 0.27 mm.
According to walk error calculated for Voff 3.5 mV (Fig. 6), the
optimum ratio between the minimum and the maximum received
amplitude is 1/2. It means that, if the received amplitude varies in
this range, we can expect that the walk error will be less than
0.5 ps that is less than 0.075 mm.
The value of the output voltage at the receiver channel is
strongly inuenced by the optics, which must be designed to stay
in the optimum variation ratio.

3. Experimental tests
The experimental tests have been aimed to the evaluation of
the performances of the optics, of the TDC circuit and of the built
PTOF system, in terms of single shot precision (SSP), average
precision (AP), walk error and pulses frequency.
3.1. Tests on optics
Basically, all optics layouts have been built according to Fig. 2
and optimizing the performance in terms of low amplitude
variation ratio. Some of the parameters used in the simulation
have been directly measured as, for example
the actual transmission index of the optics 0.21;
Table 4
Input and calculated parameters used to evaluate walk error.
Performance parameters

The expected walk error of the designed PTOF device can be


derived from these curves, when the maximum variation of the
output amplitude voltage at the receiver channel is known.
According to Voff 2.5 mV curve (Fig. 6), the optimum ratio
between the minimum and the maximum received amplitude is
0.36/2, that is 1/5.56. It means that, if the received amplitude
varies in this range, we can expect the walk error to be less than
4.1 ps that is less than 0.615 mm.

ADCMP573 G
ADCMP573 BW
ADCMP573
D
tr (rise time), tf (fall time)
0
V0
Voff
Max Vampl (received amplitude)

Fig. 6. Walk error curves calculated with Voff 2.53,03,5 mV.

500
8 GHz
4 THz
150 ps
7  10  9 s Eq. (7)
2.5  10  13 s (1/GBW)
0.4V
2.533.5 mV
2V

194

S. Logozzo et al. / Optics and Lasers in Engineering 54 (2014) 187196

the actual emitted optical power of the blue laser 30 mW;


the actual emitted optical power of the red laser 40 mW.

The experimental set-up is shown in Fig. 7.


The rst test has been aimed to the validation of the rst
numerical simulation described in Section 2.4. The laser used in
this test is Thorlabs LP405-SF10. The measured optical peak power
of the laser is approximately 30 mW. All parameters have been set
according to the rst simulation's parameters (Fig. 4).

this kind of optics is suitable for applications where the


distance range is 72 mm;
the simulation software SNR Calculation Program for Laser Radar
(2008) is able to predict the actual behavior of the optics; and
when the distance range is wider than 72 mm, the optics
needs further improvements.

Fig. 7. Experimental set-up for optical tests.

Table 5
Results of R-mode measurements without noise
(true delay value 5 ns)
Max measured delay
Min measured delay
Number of measurements
Mean value
Single shot precision in time (std)
Single shot precision in mm (std)
APD in time
APD in mm

3.2. Tests on ACAM TDC circuit

5.083 ns
4.794 ns
1023
4.934 ns
40.5 ps
6.078 mm
1.267 ps
0.19 mm

The ACAM TDC-GPX circuit has been tested before interfacing it


to the other components, in order to evaluate its own performance
parameters. The circuit has been tested using R operating mode.
The circuit has been connected to a pulse generator and it has
been fed with one start TTL input and one stop TTL input. The
pulse generator has been connected to the start and stop input of
the TDC circuit by means of two cables whose lengths differed by
1 m; therefore the true value of the time delay is about 5 ns, as
experimentally evaluated. Performance parameters have been
evaluated with and without the noise generator circuit. As already
mentioned, the noise generator is needed in order to avoid the
synchronization of the pulse generator and the TDC circuit.

Table 6
Results of R-mode measurements with noise (true
delay value  5 ns).
Max measured delay
Min measured delay
Number of measurements
Mean value
SSP in time
SSP in mm
AP in time
AP in mm

The maximum and the minimum measured amplitude are


2.55 V and 1.00 V, when the target moves backward and forward
of about 2 mm from the focused distance; however, the rst value
corresponds to channel saturation, therefore it cannot be considered. The amplitude's minimum value is very close to the simulated one (1.00 V against 1.11 V, Fig. 4), so the numerical calculus
has been proved to be a good estimate.
The second test has been performed in order to validate the
results of the second numerical simulation. The laser used in this
test is Thorlabs LP660-SF40, and the measured optical peak power
of the laser was about 40 mW. All the parameters have been set
according to simulation's parameters of Section 2.4 (Fig. 5).
The maximum and the minimum measured amplitude have
been resulted to be equal to 1.40 V and 0.60 V, respectively, when
the target moves backward and forward of about 2 mm from the
focused distances. These values are consistent with the simulated
ones (1.41 V and 0.62 V, Fig. 5).
These tests have experimentally conrmed that

5.3 ns
5.035 ns
1023
5.172 ns
36.2 ps
5.427 mm
1.131 ps
0.17 mm

Table 8
Performance parameters of the new PTOF system.
7 0.025 mm
1000 points/s

AP
Scanning speed

Table 7
Results of R-mode experimental measurements and numerical evaluations.
Amplitude [V]
1.6
Exp
# of Measurements
Max delay [ns]
Min delay [ns]
Mean value [ns]
SSP in time [ps]
SSP as a distance [mm]
AP in time [ps]
AP as a distance [mm]
a
b

1.5
a

20,440
59.36
58.83
59.04
64.2
9.63
0.449
0.067

Experimentally measured values.


Values calculated from Section 2.5.

(Section 2.5)

63.3
9.49
0.447
0.067

1.4
a

Exp

20,440
59.28
58.80
59.01
64.9
9.73
0.454
0.0681

(Section 2.5)

64.6
9.70
0.453
0.068

Exp

1.3
a

20,440
59.31
58.80
59.02
66.1
9.92
0.463
0.0694

(Section 2.5)

Expa

(Section 2.5)b

66.05
9.91
.462
0.0694

20,440
59.28
58.75
59.98
67.3
10.09
0.471
0.071

67.7
10.15
0.473
0.071

S. Logozzo et al. / Optics and Lasers in Engineering 54 (2014) 187196

195

Fig. 8. Images of a 3D tooth model.

All results of R-mode measurements without noise are reported


in Table 5.
Results of the R-mode measurements with noise generator are
reported in Table 6.
As expected, the precision is improved when the noise generator is included in the system.
3.3. Single shot precision and average precision measurements
Several tests have been executed using ACAM TDC R-Mode,
when the amplitude of the received signal was varying from 1.3 V
to 1.6 V. More than twenty thousand measurements have been
collected for each amplitude. The measured performance parameters are consistent with the calculated ones (Section 2.5), as
reported in Table 7.
In the calculations, the single shot precision of the TDC has
been considered to be about 40 ps, as evaluated from tests shown
in Section 3.2.
In all conditions, 20 MHz pulses frequency has proven to be
sufcient to achieve the measured values of precision, providing a
scanning frequency equal to 1000 points/s.

4. Discussion and conclusions


The results of experimental tests have demonstrated that the
designed device is suitable for applications in restorative dentistry
eld. The calculation model is able to predict its actual performance, in terms of single shot precision, average precision and
required pulses frequency. Therefore, operative parameters can be
chosen according to the desired performances (Table 1): 7 25 mm
average precision can be obtained with a scanning speed of at least
1000 points/s, if the laser pulsing frequency varies in the range
3050 MHz, whenever the received peak optical amplitude ranges
from 1 V to 2 V. The single-shot precision (s-value), in turn, is
equal to about 4363 ps (i.e. 910 mm) in the measurement range
(6.510 mm), when the signal amplitude varies in a range of 12 V
(see Table 8). Reported experimental results have been obtained

with a 50 MHz channel bandwidth; but higher values of precision


can be achieved with faster receivers and larger channel bandwidth, as resulting from Eq. (5). It is also important to note that
one single 50 MHz bandwidth channel is not even adequate to
3050 MHz pulses frequency, and making use of a faster channel
or recurring to multichannel receivers is strongly recommended.
Regarding walk error, it should be emphasized that the numerical model is likely to produce an optimistic estimate, and further
experimental tests are going to be performed. According to the
numerical model, the maximum error is less than 737 mm when
the offset voltage at the STOP CFD is 3.5 mV and the received peak
optical amplitude varies from 2 V to 1 V or according to the
optimum amplitude variation ratio which is 2:1 (Fig. 6).
The nal objective is the obtainment of 3D teeth model (Fig. 8)
to this aim, the here designed PTOF needs to be integrated with a
multichannel receiver (such as a 2D APD array), or with a 2 axes
scanning system (for instance by using MEMS mirrors).
The most innovative aspects of this work are
the applicability of PTOF technology to the dental eld has been
inquired for the rst time;
the current maximum values of precision and accuracy of the
existing devices and also of the PTOF laser rangenders have
been improved on short distances;
an absolute distance result value is achieved in short time;
the single axis optics is very indicated to be applied on small
size objects and to avoid shading;
the number of parallel measurement channels (receivers) can
be increased easily; and
the values of precision and accuracy can be improved even
more, by using higher channel bandwidth and faster receivers.

Acknowledgements
This paper is based on research work carried out at the
Electronics Laboratory of the Department of Electrical Engineering

196

S. Logozzo et al. / Optics and Lasers in Engineering 54 (2014) 187196

of the University of Oulu thanks to an ERASMUS grant awarded for


7 months. We express our gratitude to Prof. Kari Mtt for his
help in solving many electronics and optics issues and for providing software and scientic material and Lauri Hallman for his help
in some laboratory's 'issues'.

References
[1] Logozzo S, Franceschini G, Kilpel A, Caponi M, Governi L, Blois L. A
comparative analysis of intraoral 3D digital scanners for restorative dentistry.
Internet J Med Technol 2011;5.
[2] Kilpel A. Pulsed time-of-ight laser range nder techniques for fast, high
precision measurement applications.. Oulu: Acta Universitatis Ouluensis;
2004.
[3] Pieraccini M, Guidi G, Atzeni C. 3D digitizing of cultural heritage. J Cultural
Heritage 2001;2:6370.
[4] Ramos Barbero B, Santos Ureta E. Comparative study of different digitization
techniques and their accuracy. Comput Aided Des 2011;43:188206.

[5] Logozzo S, Zanetti EM, Franceschini G, Kilpel A, Mkynen A. Recent advances


in dental opticsPart I: 3D intraoral scanner for restorative dentistry. Opt
Lasers Eng 2013. http://dx.doi.org/10.1016/j.optlaseng.2013.07.017.
[6] Dragoman D, Dragoman M. Advanced optoelectronic devices. Springer-Verlag;
1999.
[7] Sze SM. Semiconductor devices, physics and technology. John Wiley & Sons,
Inc.; 1985.
[8] Analog Devices, Evaluation Board for ADCMP572/ADCMP573/ADCMP580/
ADCMP58/ADCMP582, 2006.
[9] ACAM, ATMD-GPX. TDC-GPX Evaluation system-datasheet, ACAM Mess
Electronic, 2005.
[10] ACAM, TDC-GPX. Ultra-high performance 8 channel time-to-digital converter
datasheet, ACAM mess electronic, 2007.
[11] Wang J, Kostamovaara J. Radiometric analysis and simulation of signal power
function in a short-range laser radar. Appl Opt 1994;33:406976.
[12] ten Bosch JJ, Coops JC. Tooth color and reectance as related to light scattering
and enamel hardness. J Dent Res 1995;74:37480.
[13] PerkinElmer, Avalanche photodiodea user guide, vol. Application NoteHigh
performance sensors, PerkinElmer precisely, 2010.
[14] Kilpel A, Ylitalo J, Mtt K, Kostamovaara J. Timing discriminator for pulsed
time-of-ight laser rangending measurements. Rev Sci Instrum 1998;69:
197884.