CALIBRATION OF CRYOGENIC THERMOMETERS FOR THE LHC

Ch. Balle, J. Casas-Cubillos, N. Vauthier, and J. P. Thermeau

Citation: AIP Conference Proceedings 985, 965 (2008); doi: 10.1063/1.2908696

View online: http://dx.doi.org/10.1063/1.2908696
View Table of Contents: http://aip.scitation.org/toc/apc/985/1
Published by the American Institute of Physics
CALIBRATION OF CRYOGENIC THERMOMETERS
FOR THE LHC

Ch. Balle1, J. Casas-Cubillos1, N. Vauthier1 and J.P. Thermeau2

1
CERN
Geneva, CH-1211, Switzerland
2
IPN, Technologie des acclrateurs et dtecteurs
Orsay, F-91406, France

ABSTRACT

6000 cryogenic temperature sensors of resistive type covering the range from room
temperature down to 1.6 K are installed on the LHC machine. In order to meet the stringent
requirements on temperature control of the superconducting magnets, each single sensor
needs to be calibrated individually. In the framework of a special contribution, IPN
(Institut de Physique Nuclaire) in Orsay, France built and operated a calibration facility
with a throughput of 80 thermometers per week.
After reception from the manufacturer, the thermometer is first assembled onto a support
specific to the measurement environment, and then thermally cycled ten times and
calibrated at least once from 1.6 to 300 K. The procedure for each of these interventions
includes various measurements and the acquired data is recorded in an ORACLE-
database. Furthermore random calibrations on some samples are executed at CERN to
crosscheck the coherence between the approximation data obtained by both IPN and
CERN. In the range of 1.5 K to 30 K, the calibration apparatuses at IPN and CERN are
traceable to standards maintained in a national metrological laboratory by using a set of
rhodium-iron temperature sensors of metrological quality.
This paper presents the calibration procedure, the quality assurance applied, the
results of the calibration campaigns and the return of experience.

KEYWORDS: cryogenic thermometer, calibration, LHC.

ATTACHMENT I
CREDIT LINE (BELOW) TO BE INSERTED ON THE FIRST PAGE OF EACH
PAPER EXCEPT FOR ARTICLES ON pp. 152 - 159, 225 - 234, 277 - 283, 284 -
290, 685 - 692, 1083 - 1090, 1359 -1366, 1375 - 1382, 1383 - 1391, 1467 - 1474,
1491 - 1498, 1499 - 1506, 1507 - 1514, and 1629 - 1639

CP985, Advances in Cryogenic Engineering: Transactions of the

Cryogenic Engineering ConferenceCEC, Vol. 53, edited by J. G. Weisend II
2008 American Institute of Physics 978-0-7354-0504-2/08/$23.00

965
INTRODUCTION

The 27-km LHC accelerator requires the world largest cryogenic system. To control
and monitor the various LHC machine components during the different operation phases
such as cool-down, beam-run, quench and more, 6000 low-temperature sensors are
necessary. FIGURE 1 shows the thermometers requested and deployed for prototypes, test
facilities and experiments in- and outside CERN.

CALIBRATION REQUIREMENT

The main parameters defining the stringent requirements are environment, industrial
installation, small uncertainty band and reliability. To fulfill these requirements, firstly a
dedicated support carrying the sensor has to be developed and secondly the best fitting
sensor available on the market has to be selected.
The temperature sensors were installed either in the insulation vacuum of the LHC-
machine or directly immersed in the cryogenic fluid. Accordingly three different types of
sensor supports were developed at CERN [1] and produced by Anco in Greece and Air
Liquide in France. Once the sensor is mounted on its support the assembly is called long,
short or finger thermometer according to the type of supports. The design of the different
assemblies is shown in TABLE 1.
The selection of the sensor was restricted to the family of resistive temperature
detectors (RTD), which are well known and used since decades in all kind of technical
applications [2]. After investigation on calibration, irradiation and thermal cycling at
CERN and IPN [3-4] two types of sensors were selected for the LHC-machine: a thin film
zirconium oxynitride sensor with the brand name Cernox, manufactured by Lake Shore
Cryotronics and an industrial platinum RTD of type Pt100. The latter covers a range from
50 to 300 K and follows a standard response curve within a defined class of tolerance
according to the European norm IEC 751. Therefore it is interchangeable and no individual
calibration is needed. The chosen Cernox -model CX-1050 shows a monotonic R-T
behavior from 1.6 K to 300 K, has a sensitivity of at least 1000 /K below 4 K, and
requires an individual calibration for every single sensor.

ACS Superconducting Cavities DFB Electrical Feed Box

DSL El. Superconducting Link LB Dipole Magnet
LE Connection Cryostat LQ Quadrupole Magnet
QRL Cryogenic Distribution Line QUI Cryogenic Interconnection Box
FIGURE 1. Cryogenic thermometer quantities requested for the LHC main components, experiments and
tests.

966
TABLE 1. Thermometer requirements as a function of the LHC main components.
Interconnect, Magnet QRL QRL
DFB, DSL,
QRL, ACS,
Requirements & Solutions QUI, DSL, LE
Temperature Range [K] 1.6..300 1.6..300 1.6..300 50..300
Installation in vacuum He He vacuum
Interchangeability
Quench-resistant
Lifetime 20 years 20 years 20 years 20 years
Thermal Cycle 300..1.8 K 1/year 1/year 1/year 1/year
Radiation 5 x 1011 neutrons / cm2 over 10 years
Non-expert installation Yes Yes Yes Yes
Uncertainty below 4 K 5 mK 5 mK 5 mK 5 mK
Uncertainty above 4 K 1% 1% 1% 1%
Secure wire connection

Sensor selected Cernox Cernox Cernox Pt100

Support type

From a logistic point of view, the platinum RTD is a standard supply. Since the
standard calibration satisfies the requirements where it is installed, the platinum RTD will
be ignored in the following sections. The choice of the Cernox RTD with its singularity
had a deep impact regarding logistics, procedures and tools to be used along the lifetime of
the thermometer. Consequently a dedicated cryogenic thermometer database and a
calibration facility was build-up.
The Oracle 8i thermometer database, called ThermBase, stores the curriculum vitae
of every thermometer including the calibration and test data. Also the LHC control
database and CERN equipment database are pointing to the thermometer database for
enabling downloads of calibration tables to the Front-End Controller (FEC) [5].
A cryogenic thermometer calibration facility was designed and constructed at CERN
for the sensor selection campaign, for validating the supports and for small series
calibrations [6]. Compared with test facilities in industry and other laboratories, the CERN
insert is designed to fix the LHC cryogenic thermometer directly on it. Thus calibration
under LHC installation conditions is ensured.
The gained knowledge and experience was transferred to IPN (Institut de Physique
Nuclaire) in Orsay, France within the framework of a special contribution. A system of
two automatic calibration benches was built to calibrate the thermometers for the LHC
machine [3]. All the calibrations were made by comparison with working standard

967
thermometers of rhodium-iron type. Those standards are maintained at CERN by means of
secondary standards and a calibration apparatus from the Institute for Physical Research
VNIIFTRI, Moscow, Russia. The secondary standards are maintained once per year by the
Italian national metrological institute INRIM (Istituto Nazionale di Ricerca Metrologica) in
Torino, Italy. They are traceable in the range of 1.6 K to 30 K and have an uncertainty of
2 mK to 5 mK. For temperatures above 30 K, the calibration of the working standards
is followed by a platinum standard thermometer coming from NPL (National Physical
Laboratory) in Teddington, England.
A copy of the CERN thermometer database was installed at IPN, where several
upgrades and modifications of the structure and the Oracle Forms user-interface were
made. A fast transfer of IPN calibration data to CERN was possible thanks to a direct
connection between both laboratories.

PROCEDURE

A network composed of several suppliers and laboratories was created to take

advantage of their competencies and experience (See TABLE 2).

Assembling

Over a period of 2 years Lake Shore Cryotronics supplied to CERN a total of

7000 Cernox sensors of type CX-1050-SD with the characteristic R(1.7 K) = 22500
5000 . An electronic file, containing the resistive values at room, liquid-nitrogen and
liquid-helium temperature of the batch of sensors supplied was attached to every shipment.
The sensor was carefully soldered by hand to a support, specific to the measurement
environment and supplied by Anco. Like for the sensors, the support deliveries were
accompanied with a test sheet listing the resistive values of the conductors. The serial
number of the assembled thermometer is graved on it to avoid any mistake. An automatic
electrical continuity test and a 4-wire resistance measurement at room temperature
compared with the values supplied by Lake Shore Cryotronics could validate the
thermometer conformity. The data acquired during these tests and the type of support are
added to the file coming from Lake Shore Cryotronics.
The above-mentioned tests were repeated every time the thermometer undergoes an
intervention such as installation and shipping, or in case the thermometer shows an
unexpected behavior. This set of data is called thermometer traveler.

TABLE 2. List of suppliers and laboratories and their contribution.

Collaborators Contribution

Lake Shore Cryotronics , Westerville, USA Cernox-RTD manufacturing
Anco, Athen, Greece Support manufacturing
INRIM, Torino, Italy Secondary standard calibration facility
Data analyzing
Calibration plateau definition
IPN, Orsay, France Thermal cycling facility
Mass calibration facility
Data pre-analyzing
Oracle 8i database

968
FIGURE 2. The main steps of the procedure from an off-shelf sensor to a calibrated LHC thermometer.

Thermal Cycling

After assembly, each thermometer was packed in a foam-cushioned box. In parallel,

an electronic file containing its data was send to IPN. At IPN this file is imported to
ThermBase. Then a fully automatic reception test was executed on the thermometers; the
data compared and recorded. An alarm was sent out if the difference was out of tolerance.
In order to improve the long-term stability of the sensors, each thermometer was
thermally cycled ten times between 4.2 K and 300 K with an automatic cycling facility [4].
No thermometer data were acquired during cycling.

Calibration

IPN was running two facilities: one for calibrations in liquid helium and the other
under vacuum conditions. The long thermometers were calibrated at least once under
vacuum over the full temperature range from 300 K to 1.6 K. This corresponds to step 4 in
the FIGURE 2. The short and finger thermometers underwent step 4 and 5 of the
FIGURE 2. They were calibrated in liquid helium in the range from 1.6 to 4.2 K, their
future working conditions, and under vacuum up to 300 K so that they can be used for
monitoring during cool-down and warm-up operation. Both calibrations were automatically
driven by a Programmable Logic Controller (PLC).

7000
Number of calibrated thermometers

6000

5000

4000

3000

2000

1000

0
2001 2002 2003 2004 2005
Year

FIGURE 3. The bold curve represents the sum of calibrated long, short and finger thermometers over 5
years. Up from May 2002 the calibration rate is higher, because the second additional vacuum insert is under
operation.

969
 A personal computer (PC) running LabVIEW from National Instruments acquired
the measurement and process data and wrote them directly to the database. The calibration
time was short compared to the time of the operator for installing and connecting up to 90
thermometers on the insert, entering the configuration set-up (i.e. channel, position and
serial-number) and bringing the system into operation. Therefore a second insert for the
calibration under vacuum was constructed and put in place in May 2002. The increase of
the calibration rate due to this second insert is shown in FIGURE 3. An optimized set of
calibration temperature setpoints was computed by INRIM after having analyzed the
calibration curve progression of 300 thermometers.

Data Analysis

After calibration the IPN operator pre-analyzed the acquired data with a software tool
developed at IPN and written in LabVIEW. Every thermometer had at least one
approximation function of its temperature-resistance (T-R) characteristic, which is
typically a polynomial function of several degrees. No calibration data were removed from
the database. Non-conforming measurements were simply flagged-out. Before shipping the
thermometers back to CERN a final electrical test was executed and recorded at IPN.
Finally all the newly acquired data are transferred to ThermBase at CERN.
At CERN, a linear interpolation function for the FECs was deduced from the
polynomials. INRIM has developed a method to generate cubic spline functions for better
approximating the thermometer T-R points without the risk of oscillations, which are
characteristic for polynomials. However this method was abandoned owing to the lack of
robustness when one or more anchor points were not performed. The present polynomial
interpolation obtained is satisfactory.

RESULTS

Crosscheck Calibration

CERN executed random calibrations on some calibrated thermometers coming from

IPN in order to crosscheck the coherence between the approximation data obtained by IPN
and CERN. FIGURE 4 shows the difference (T = TCERN(R) - TIPN(R)) of 6 thermometers
20

15

10

5
T [mK]

-5

-10

-15

-20
1 10 100
Temperature [K]

FIGURE 4. Crosscheck between IPN and CERN calibration. (T = TCERN(R) - TIPN(R))

970
 6

Cell 5
Cryogenic Cell
4
Frequency Cryogenic

2
Frequency

0 5 10 15 20 25
|T| within a LHC Cryogenic Cell [mK]
|T| within a LHC Cryogenic Cell [mK]
FIGURE 5. Absolute temperature spread of thermometers within a LHC cryogenic cell at 1.95 K, assuming
FIGURE 5. Absolute
that the gradient temperature
of temperature spread
of a cell of thermometers within a LHC cryogenic cell at 1.95 K, assuming
is zero.
that the gradient of temperature of a cell is zero.
at a given temperature. Below 10 K a systematic deviation of - 5 mK can be observed This
atanomaly
a givenwas
temperature. Below
constant over the 10
fullKproduction
a systematic deviation
period. Furtherofinvestigations
- 5 mK can be observed
have to be This
anomaly was constant
made to understand over the full production period. Further investigations have to be
the reason.
made to understand the reason.
Thermometer in LHC Operation
Thermometer in LHC Operation
During cool-down of the first sector of the LHC-machine the performance of the
thermometers was within of
During cool-down thethe
expectations.
first sectorAsof an
theexample FIGUREthe
LHC-machine 5 shows that theof the
performance
temperature spread within 24 isothermal cryogenic cells is, with one exception, inside the
thermometers was within the expectations. As an example FIGURE 5 shows that the
tolerance band.
temperature spread within 24 isothermal cryogenic cells is, with one exception, inside the
tolerance band.
CONCLUSION

CONCLUSION
From the years 2001 to 2005 more than 6500 cryogenic thermometers equipped with
Cernox-RTD have been calibrated for the LHC-machine and for other applications. The
From
feedback theusers
from yearsand
2001
the to
first2005 moreofthan
readings 6500 cryogenic
the cooled thermometers
sector of the equipped with
machine are positive.
Cernox-RTD have been
The high number calibrated
of sensors forassembled,
to be the LHC-machine andindividually
cycled and for other applications.
calibrated, The
together with
feedback fromsevere
users environmental conditions
and the first readings and cooled
of the tight uncertainties
sector of thespecification,
machine aremadepositive.
this thermometer project of
The high number a great
sensorschallenge. Well-definedcycled
to be assembled, procedures were necessary
and individually to
calibrated,
guarantee the requirements.
together with severe environmental conditions and tight uncertainties specification, made
this thermometer project a great challenge. Well-defined procedures were necessary to
guarantee the requirements.
ACKNOWLEDGEMENTS

We thank all the people inside and outside CERN who participated to this project.
ACKNOWLEDGEMENTS
We appreciated the constructive feedbacks from the end-users in the field, as well as the
encouragement of our colleagues and the availability and precise work of the technicians
We thank all the people inside and outside CERN who participated to this project.
involved.
We appreciated the constructive feedbacks from the end-users in the field, as well as the
encouragement of our colleagues and the availability and precise work of the technicians
involved.

971

971
REFERENCES

1. Balle, Ch. and Casas-Cubillos, J., Industrial-type Cryogenic Thermometer with Built-in Heat
Interception in Advances in Cryogenic Engineering 41B, Plenum, New York, 1996, pp. 1715.
2. Rubin, L. G., Cryogenic Thermometry: A review of Progress since 1982 in Cryogenics 37, 1997, pp.
341-356
3. Junquera, T., Amand J.F., Thermeau J.P., and Casas-Cubillos, J., Neutron Irradiation Tests of
Calibrated Cryogenic Sensors at Low temperatures in Advances in Cryogenic Engineering 43A,
Plenum, New York, 1997, pp. 765.
4. Balle, Ch., Casas-Cubillos, J. et al, Influence of Thermal Cycling on Cryogenics Thermometers in
Advances in Cryogenic Engineering 45B, Plenum, New York, 1999, pp. 1817.
5. Ainslie, Ch., Cryogenic Thermometer Database System User Requirements and Implementation in
International Workshop SoftTools-MetroNet held at Ljubljana, 2004.
6. Balle, Ch., Casas-Cubillos, J. and Thermeau, J.P., Cryogenic Thermometer Calibration Facility at
CERN in Advances in Cryogenic Engineering 43A, Plenum, New York, 1997, pp. 741.
7. Thermeau, J.P. et al, Cryogenic Thermometer Calibration Facility for the LHC in Advances in
Cryogenic Engineering 45B, Plenum, New York, 1999, pp. 1825.
8. Ichim, D., Data Computing for LHC Cryogenic Thermometer in CERN Engineering Specification
LHC-QIT-ES-0004, 2003

972