-
Nuclear Instruments and Methods in Physics Research A 667 (2012)
5–10
Contents lists available at SciVerse ScienceDirect
Nuclear Instruments and Methods inPhysics Research A
0168-90
doi:10.1
n Corr
E-m
journal homepage: www.elsevier.com/locate/nima
On-machine non-contact dimension-measurement system with
laserdisplacement sensor for vane-tip machining of RFQs
Y. Kondo a,n, K. Hasegawa a, H. Kawamata b, T. Morishita a, F.
Naito b
a J-PARC Center, JAEA, Japan Atomic Energy Agency, 2-4
Shirakata-Shirane, Naka, Ibaraki 319-1195, Japanb KEK, High Energy
Accelerator Research Organization, 1-1 Oho, Tsukuba, Ibaraki
305-0801, Japan
a r t i c l e i n f o
Article history:
Received 29 August 2011
Received in revised form
17 October 2011
Accepted 22 November 2011Available online 1 December 2011
Keywords:
Proton accelerator
RFQ linac
Machining
Vane-tip profile
On-machine measurement
02/$ - see front matter & 2011 Elsevier B.V. A
016/j.nima.2011.11.065
esponding author. Tel.: þ81 29 284 3133; faxail address:
[email protected] (Y. Kon
a b s t r a c t
We have developed a new on-machine non-contact
dimension-measurement system using a laser
displacement sensor (LDS) for machining the vane tips of RFQs.
The LDS was attached to a milling
machine, and the dimensions of a test piece (TP) with a
modulated vane tip were measured to evaluate
the performance of the system. A longitudinal profile was
measured with the LDS and it was compared
with the profile measured with a coordinate measuring machine
(CMM). A transverse profile was also
measured. The required accuracy of dimension measurement was
achieved using our system.
& 2011 Elsevier B.V. All rights reserved.
1. Introduction
For the fabrication of radio-frequency quadrupole (RFQ)
cavities, itis very important to machine the vane tips accurately.
Fig. 1 shows asemi-assembled (before assembling the upper vane) RFQ
cavity; thevane tips with modulation are clearly seen in this
figure. The accuracyof the dimensions and geometries of the vane
tips are very importantto realize the design performance of the
RFQ. To ensure the machin-ing accuracy after the final machining
operation, the dimensionsshould be measured on a
computer-numerical-control (CNC) millingmachine or measured using a
three-dimensional coordinate-measur-ing machine (CMM). However, the
vanes of RFQs in modern protonaccelerators are typically made of
annealed oxygen-free copper (OFC),which is a very soft metal.
Therefore, contact-type probes causeindentations at the measurement
points, as shown in Fig. 2. Thediameter of the indentation is 0.3
mm. The surface electric field of theRFQ vane tip is high
(typically 1.8 Kilpatrick); therefore, such damageto the vane tips
should be avoided. Even if the damage by the contact-type probe is
regarded as acceptable, it is difficult to performmeasurement at
sufficient number of points to confirm the modula-tion profiles
within a realistic time frame. The measurementof modulation
profiles using a CMM with a contact-type probe isalso
time-consuming. Thus, a non-contact-type measuring device is
ll rights reserved.
: þ81 29 284 3719.do).
desired that can perform measurement at many points in a
shortperiod of time.
To this end, we have developed a new method for
measuringvane-tip profiles on the CNC milling machine. In this
on-machinenon-contact measurement system, a laser displacement
sensor (LDS)is used to measure the vane-tip profiles, thus
eliminating the need forany contact-type probes. It is difficult to
measure the vane-tip profileaccurately using optical devices
because the vane tip has a complexshape and a glossy metallic
surface. However, an LDS can be used forthis type of measurement.
We have developed this system for thefabrication of an RFQ to be
used in the beam-current upgrade of theJ-PARC (Japan proton
accelerator research complex) [1] linac. Similarto many modern
proton linacs, the J-PARC linac uses a four-vane RFQ.The resonant
frequency of the J-PARC RFQ is 324 MHz, and itaccelerates H�
particles from 50 keV to 3 MeV.
The machining accuracy of the vane tips of the J-PARC RFQ
isrequired to be within 720 mm. According to the SUPERFISH
[2]calculation, the frequency shift corresponding to a
displacementof 20 mm of one vane tip is 0.2 MHz. This accuracy
requirementis typical for � 300-MHz RFQs; however, careful
machining isnecessary to achieve this accuracy. During the final
machiningoperation of RFQ II for the J-PARC linac [3–5], we
measured thedimensions of the vane tips on the milling machine
after each cutexcept for the final cut. The measured values are fed
back to offsetthe final cut depths and achieve the required
machining accuracy.The dimensions were not measured on the milling
machine afterthe final cut to avoid the damage described above.
Details of thison-machine measurement are given in Section 2.
www.elsevier.com/locate/nimawww.elsevier.com/locate/nimadx.doi.org/10.1016/j.nima.2011.11.065mailto:[email protected]/10.1016/j.nima.2011.11.065
-
Fig. 1. Photograph of a semi-assembled RFQ cavity.
Fig. 2. Photograph of the indentation caused by the contact-type
probe on theannealed OFC TP. The diameter of the indentation is 0.3
mm. The scale-like
patterns are marks made by the ball nose cutter.
Y. Kondo et al. / Nuclear Instruments and Methods in Physics
Research A 667 (2012) 5–106
The target of our new on-machine non-contact measurementsystem
is to measure the ridge profiles of vane tips with
sufficientaccuracy; i.e., the mean value of deviation from a
profile mea-sured with highly accurate devices such as the CMM1
should bewithin 710 mm and the dispersion should be within 720
mm.Because the parameters of the J-PARC RFQs are typical for�
300-MHz RFQs, our system is applicable to dimension mea-surement of
the vane tips of many RFQs in this frequency range.
1 Accuracy of CMMs is typically within a few mm.
In Section 2, the procedure of final machining of the vane tips
ofRFQ II is briefly summarized. The details of the on-machine
non-contact measurement system are presented in Section 3.
Theresults of the test measurement are discussed in Section 4.
Finally,in Section 5, a summary is presented.
2. Final machining of vane tips of J-PARC RFQ II
In this section, the procedure of final machining of the RFQ
IIvane tips is described, because it is necessary to know the
procedureto understand the need for the on-machine measurement
system.
The RFQ II cavity consists of three longitudinally dividedunits,
called units 1, 2, and 3. The vane lengths of each unit are1057.2
mm, 1053.6 mm, and 1061.3 mm. Each unit consists of twomajor vanes
(upper and lower parts) and two minor vanes (leftand right parts),
and the four vanes are brazed together. Ball nosecutters were used
for machining the vane tips of RFQ II [3,4].
The final machining process of the RFQ II vane tips was
carriedout as follows:
On-machine measurement before final machining
(residualdimension: 0.2 mm)- first cut (residual dimension: 0.14
mm) - on-machinemeasurement
Fig. 3. Photograph of the vane-tip machining (upper figure) and
the on-machinemeasurement with the contact-type probe (lower
figure).
-
number of cuts3210
resi
dual
dim
ensi
on (m
m)
0
0.05
0.1
0.15
0.2
unit 1
number of cuts3210
resi
dual
dim
ensi
on (m
m)
0
0.05
0.1
0.15
0.2
unit 2
number of cuts3210
resi
dual
dim
ensi
on (m
m)
0
0.05
0.1
0.15
0.2
unit 3
Fig. 4. Target and measured residual dimensions of the vane-tip
in the finalmachining process. The horizontal axes represent the
number of cuts, and the
vertical axes denote the differences between the target or
measured dimensions
and the design ones. The vane-tip dimensions were measured on
the machine
immediately after each cut, and the target values of the final
cuts were
compensated for by the measured dimensions. The dashed lines
represent the
expected values of the final dimensions.
Fig. 5. Schematic drawing of the LDS measurement system.
Fig. 6. Photograph of the measurement setup.
Y. Kondo et al. / Nuclear Instruments and Methods in Physics
Research A 667 (2012) 5–10 7
- second cut (residual dimension: 0.08 mm) -
on-machinemeasurement- final cut - no measurement
where residual dimension indicates the difference between
thedimension after each cut and the design dimension. Fig. 3
showsphotographs of the vane-tip machining (upper figure) and
on-machine measurement (lower figure) processes. To compensate
forthe expansion of the spindle of the milling machine, the
dimensionsof the vane tips were measured on the machine. After each
cut wasfinished, the tool of the milling machine was changed from
the ballnose cutter to a contact-type probe. The probe was moved to
themeasurement point and touch the vane tip, as shown the lower
partof Fig. 3. The dimensions can be derived from the position of
the mainspindle. The accuracy of this measurement is within 10 mm.
Theresults were fed back to adjust the depths of the final cuts.
The on-machine measurement of RFQ II was performed at the
followingpoints: For unit 1, two points on the radial-matching
section, fourpoints at the top, and four points at the bottom of
the modulationwere sampled. For units 2 and 3, ten points at the
top of themodulation were sampled. One residual dimension was
derived foreach vane by averaging the measured residual dimensions
of thesampled points. The dimensions were manually read out and it
tookabout 5 min to measure ten points. More points should be
measured
to confirm the overall shape of the vane, for example, to
determinewhether it is warped or not. However, an increase in the
number ofmeasurement points increases the measurement time. The
operationof the machine cannot be stopped for the long duration of
a finalmachining process, because as the machine cools down, the
expandedspindle shrinks back to its original dimensions, thus
rendering thecompensation ineffectual. Acceptable interruption is
about 1 h andtypical duration of interruption during a procedure of
the finalmachining operation of RFQ II was less than 30 min.
Fig. 4 shows the target and the measured residual dimensionswith
respect to the design dimensions. The target residual
dimensionsdenote the offset value inputed to the milling machine,
and themachine cuts the surface according to this residual
dimension. At thefinal cuts, the target residual dimensions are
compensated for bythe measured dimensions. The dashed lines in the
figures representthe expected final dimensions. However, the final
dimensions werenot measured on the machine to prevent damage to the
vane-tipsurface due to the contact-type probe.
3. Apparatus for on-machine non-contactmeasurement system
In the previous section, the procedure of final machining ofthe
RFQ II vane tips has been described. We will now describe
theon-machine non-contact measurement system.
-
yCNC (mm)89
devi
atio
n (m
m)
-0.04
-0.03
-0.02
-0.01
0
0.01 Top 1
yCNC (mm)
devi
atio
n (m
m)
-0.04
-0.03
-0.02
-0.01
0
0.01 Slope 1
devi
atio
n (m
m)
-0.04
-0.03
-0.02
-0.01
0
0.01Bottom
yCNC (mm)yCNC (mm)
yCNC (mm) yCNC (mm)
devi
atio
n (m
m)
-0.04
-0.03
-0.02
-0.01
0
0.01 Slope 2
devi
atio
n (m
m)
-0.04
-0.03
-0.02
-0.01
0
0.01 Top 2
-4
devi
atio
n (m
m)
-0.04
-0.03
-0.02
-0.01
0
0.01 Flat
-3 -2 -1 0 1 2 3 4
90 91 92 93 94 95 96 97 98 89 90 91 92 93 94 95 96 97 98
89 90 91 92 93 94 95 96 97 9889 90 91 92 93 94 95 96 97 98
89 90 91 92 93 94 95 96 97 98
Fig. 7. Linearity of the LDS measured at each measurement point
on the TP. The ‘‘Top 1’’, ‘‘Slope 1’’, etc., correspond to the
positions indicated in Fig. 5. The open squaresare the deviations
between the displacements measured with the LDS and the movements
of the LDS inputed to the CNC machine. The hatched areas represent
the
required ranges for the measurement of the RFQ II vane-tip
dimensions. The ‘‘Flat’’ shows the result of linearity measurement
on the flat surface on a OFC block as
reference.
Y. Kondo et al. / Nuclear Instruments and Methods in Physics
Research A 667 (2012) 5–108
A laser displacement sensor KEYENCE LK-G50002 is used forthis
system. The sensor head is LK-H022, and the controller isLK-H500V.
This sensor head uses a red laser diode with awavelength of 650 nm
and a power of 0.95 mW. The spot size is25 mm on the object to be
measured, and the laser spot on theobject is viewed with a CMOS
sensor. If the distance from the laserdiode to the object is
varied, the position of the image of the laserspot on the CMOS
sensor is also varied and this displacement canbe translated into
distance. This sensor head was used because ithas a measurable
range of 2073 mm, and with this range, we canmeasure the dimensions
of the vane tips of many � 300-MHzRFQs. For example, the minimum
bore radius of RFQ II is 2.11 mmand the maximum bore radius is 4.85
mm. Therefore, the width tobe measured is 2.74 mm (71.37 mm), and
this range is muchnarrower than the measurable range of LK-H022.
The maximumsampling frequency of LK-H022 is 392 kHz, but in the
followingmeasurement, a default setting of 5 kHz was used. The
nominalvalue of the linearity is 70.02% of the full range (6 mm),
i.e.,71:2 mm, and the nominal value of the repeatability is 0:02
mm.Both values are obtained by measuring white
diffuse-reflectivework pieces.
2 KEYENCE Co., 1-3-14, Higashi-nakajima, Higashi-Yodogawa, Osaka
533-8555,
Japan, http://www.keyence.com.
Fig. 5 shows a schematic drawing of the measurement
setup.LK-H022 was attached to a holder and installed in the tool
chuckof the main spindle of a machining-center FNC-1063
(MakinoMilling Machine Co. Ltd.4). Fig. 6 shows a photograph of
themeasurement setup. A TP made of OFC shown in this figure
wasmeasured to test the system. The length of the TP was 500
mm.
4. Results of test measurement
4.1. Linearity of LDS
To evaluate the basic performance of our on-machine mea-surement
system, test measurement was performed. First, thelinearity of the
LDS was confirmed. In the following measure-ment, the longitudinal
direction of the TP was defined as thez-axis, the vertical
direction was y-axis, and the horizontaldirection was x-axis, as
shown in Fig. 5. For the linearitymeasurement, the x- and
z-positions were fixed at each measur-ing point and the y-position
of the LDS was scanned within themeasurable range of the LDS. Then,
the distance of motion of the
3 This machining center was used for the test measurement; RFQ
II was not
machined with this machining center.4 Makino Milling Machine Co.
Ltd., 3–19 Nakane 2-chome, Meguro-ku, Tokyo
152-8578, Japan
http://www.keyence.com
-
z (mm)410
y mes
(mm
)
90.5
91
91.5
92
92.5
93
93.5
devi
atio
n (m
m)
-0.02
-0.01
0
0.01
0.02
0.03
0.04yLDSyCMMdeviation
deviation (mm)-0.05
coun
ts/0
.005
mm
0
2
4
6
8
10 µ = 0.006 mm
-0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05
420 430 440 450
Fig. 8. (a) Comparison between the LDS and CMM measurement
results of thetransverse profile of the modulation of the TP. The
open circles are the dimensions
measured with the LDS, and the open squares represent those
measured with the
CMM (left axis). The closed circles indicate the deviations
between the LDS and
CMM measurement (right axis). (b) Histogram of the deviations.
The mean value
of the deviation is 6 mm.
x (mm)-3
y mes
(mm
)
89
89.5
90
90.5
91
91.5
92
92.5
93
93.5
devi
atio
n (m
m)
-0.04
-0.03
-0.02
-0.01
0
0.01
0.02
0.03
0.04
0.05
yLDSyCMMdeviation
-2 -1 0 1 2 3
Fig. 9. Measured transverse profiles of the vane tip. The open
circles are thedimensions measured with the LDS and the open
squares are those measured with
the CMM (left axis). The solid circles are the differences
between the LDS and
CMM measurement (right axis). The deviations are within 715 mm
in the regionof �2 mm to 2 mm from the ridge of the vane.
Y. Kondo et al. / Nuclear Instruments and Methods in Physics
Research A 667 (2012) 5–10 9
LDS inputed to the CNC machine and the displacement measuredwith
the LDS were compared. The measured points were twopeaks, two
centers of the slopes, and one bottom of the modula-tion, they are
denoted in Fig. 5 as ‘‘Top 1’’ to ‘‘Top 2’’.
Fig. 7 shows the results of the linearity measurement.
Thehorizontal axis of each figure is the y-position of the LDS, and
thevertical axes are the deviations. The deviation is calculated
bysubtracting the movement of the LDS inputed to the CNC
machinefrom the measured displacement obtained by the LDS. The
scanswere started from the largest y-values, and therefore, the
devia-tions are zero at the maximum y-values. The ‘‘Flat’’
indicates theresult of linearity measurement on the flat surface of
an OFC blockmeasured as reference. The hatched areas represent the
requiredranges of the vane tips of RFQ II to be measured. In these
ranges,the deviations were within 710 mm at ‘‘Slope 1’’, ‘‘Slope
2’’, and‘‘Bottom’’; this is sufficiently high to achieve the target
measure-ment accuracy. However, the deviations were from �35 mm
toþ0 mm at ‘‘Top 1’’ and ‘‘Top 2’’. The measured linearities
wereworse than the nominal value because the vane tip has a
complexshape and metallic luster. Especially, the linearity was not
good atthe top of the modulation; this is because the vane behaves
as aconvex lens at the top of the modulation, that is, if the axis
of thelaser is not exactly perpendicular to the vane, the position
of theimage of the laser spot on the CMOS sensor changes
drasticallyowing to a small change of the displacement. Precise
tuning of thelaser axis is difficult, and therefore, the
performance of the systemwas evaluated including this
imperfection.
4.2. Longitudinal profile measurement
Next, a longitudinal profile of the two periods of modulationwas
measured with the LDS, and the measured profile wascompared with
that measured with the CMM. In this measure-ment, the x- and
y-positions of the LDS were fixed on the ridge ofthe modulation and
the z-position was varied with a 1-mm pitch.The pitch of the CMM
measurement was also set as 1 mm. Theorigin of the y-axis was set
at the bottom surface of the TP. InFig. 8a, the profiles measured
with the LDS and CMM are shown.The open circles are the dimensions
measured with the LDS, andthe open squares are those measured with
the CMM; they appearalmost overlapped. The solid circles represent
the deviations,which were calculated by subtracting the dimensions
measuredwith the CMM from those measured with the LDS. Fig. 8b
shows ahistogram of the deviations. The mean value of the
deviations is6 mm. If the values measured at the top of the
modulation aretruncated, where the linearity was found to be
insufficient asdiscussed in Section 4.1, the dispersion of the
deviations from themean value is within 720 mm.
4.3. Transverse profile measurement
A transverse profile of the vane tip was also measured. Fig.
9shows transverse profiles measured with the LDS and the CMM.The
open circles represent the dimensions measured with theLDS, and the
open squares are those measured with the CMM. Thesolid circles
represent the deviations between the LDS and CMMmeasurement. In the
region from �2 mm to 2 mm from the ridgeof the vane, the deviations
are within 715 mm. This accuracy issufficiently high to confirm the
transverse profile of the vane.
4.4. Three-dimensional scanner
Fig. 10 shows the demonstration of an LDS as a
three-dimen-sional scanner. The LDS was scanned in the longitudinal
(z)direction, and the scans were performed along seven
scanninglines. The measured range of x-direction was from 0.0 mm
to
3.0 mm, and the pitch was 0.5 mm. The read-out rate of theLDS
was set to 10 Hz; the data is processed at a rate of 5 kHz inthe
LK-H500V but read out every 500 samples. The LDS wascontinuously
moved along the z-direction. The number of points
-
Fig. 10. Demonstration of the LDS as a three-dimensional
scanner. The LDS was scanned along the z-direction, and the
scanning lines are distributed from x¼0.0 to 3.0 mmwith a pitch of
0.5 mm.
Y. Kondo et al. / Nuclear Instruments and Methods in Physics
Research A 667 (2012) 5–1010
along one scanning line was 1.33 points/mm, i.e., the
movingspeed of the LDS along the z-axis was 7.5 mm/s. The data
wereaveraged over 256 samples, and thus, the y-dimensions are
thevalues averaged over the z range of 0.38 mm. The number ofpoints
in one scanning line was 435 points. With this setting, thetime
required for scanning one line was 43.5 s. If a 1-m vane ismeasured
with a 1-mm pitch, the number of points is 1000 andthe data for one
line can be taken in 100 s. The read-out rate canbe increased up to
the sampling frequency. However, if themoving speed is increased,
the position of the LDS should be keptstopping while reading the
data. The setting of the LDS and themilling machine should be
carefully tuned to achieve the bestperformance.
5. Summary
An on-machine non-contact dimension-measurement systemwith an
LDS was developed to measure the dimensions of RFQvane tips. The
longitudinal profile of a modulated test piece wasmeasured and
compared with the profile measured with a CMM.The absolute value of
the deviation was 6 mm, and the dispersionwas 720 mm; these values
meet the requirement of this system.
A transverse profile was measured with an accuracy of 715 mmin
the region of 72 mm from the ridge. The speed to measure
thelongitudinal profile of one vane tip with a pitch of 1 mm is0.01
m/s. This speed and number of points are much higher thanthose of
the contact on-machine measurement of J-PARC RFQ II;they were
typically 0.003 m/s and 10 points/m, respectively.Improving the
linkage between the data-acquisition system ofthe LDS and the
motion of the milling machine remains an issue.We will use this
system for the on-machine measurement of ournext RFQ to be used in
the beam-current upgrade of theJ-PARC linac.
References
[1] T. Koseki, et al., Challenges and solutions for J-PARC
commissioning and earlyoperation, in: Proceedings of IPAC’10, 2010,
pp. 1304–1308.
[2] J.H. Billen, L.M. Young, POISSON SUPERFISH, Technical
Report, LA-UR-96-1834,1996.
[3] T. Morishita, et al., Fabrication of the new RFQ for the
J-PARC linac, in:Proceedings of IPAC’10, 2010, pp. 783–785.
[4] T. Morishita, et al., Vane machining by the ball-end-mill
for the new RFQ in theJ-PARC linac, in: Proceedings of LINAC2010,
2010, pp. 518–520.
[5] T. Morishita, et al., Vacuum brazing of the new RFQ for the
J-PARC linac, in:Proceedings of LINAC2010, 2010, pp. 521–523.
-
本文献由“学霸图书馆-文献云下载”收集自网络,仅供学习交流使用。
学霸图书馆(www.xuebalib.com)是一个“整合众多图书馆数据库资源,
提供一站式文献检索和下载服务”的24 小时在线不限IP
图书馆。
图书馆致力于便利、促进学习与科研,提供最强文献下载服务。
图书馆导航:
图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具
http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/vip.htmlhttp://www.xuebalib.com/db.phphttp://www.xuebalib.com/zixun/2014-08-15/44.htmlhttp://www.xuebalib.com/
On-machine non-contact dimension-measurement system with laser
displacement sensor for vane-tip machining of RFQsIntroductionFinal
machining of vane tips of J-PARC RFQ IIApparatus for on-machine
non-contact measurement systemResults of test measurementLinearity
of LDSLongitudinal profile measurementTransverse profile
measurementThree-dimensional scanner
SummaryReferences
学霸图书馆link:学霸图书馆