Using an optical micrometer for 
mechanical thermostat membrane 
expansion measurements
can be constructed in many ways
and may use a variety of principles
to measure and regulate the tempe-
rature. The output of the thermostat
then controls the heating or cooling
gadgets. Common sensors include:
bi-metal mechanical sensors, expan-
sion medium sensor, electronic ther-
mistors, electrical thermocouples
etc. This article describes the pre-
liminary tests for controlling of the
expand medium filling in production
of expansion medium sensor based
ther-mostats, such can be seen in fi -
gure 1. The whole temperature sen-
sor consists of probe, capillary tube,
diaphragm with ceramic button and
expansion medium. When the sen-
sor is heated, the medium heats up
and expands. The expansion of the
medium increases the pressure in
the closed-circuit system. The pres-
sure increase is converted into a di-
splacement in the diaphragm. This
displacement, also called travel,
actuates a snap-action switch which
opens or closes the contacts in the
electric circuit. The reference varia-
ble is set via the thermostats adju-
sting spindle.
Mag. Jure Rejc, univ. dipl. inž.,
prof. dr. Marko Munih, univ.
dipl. inž., University of Ljublja-
na, Faculty of Electrical enginee-
ring, Laboratory of Robotics and
Biomedical Engineering, Ljublja-
na, Slovenia
Jure REJC, Marko MUNIH
Abstract: This paper presents a series of preliminary tests for measuring expansion of the mechanical thermostat
diaphragm during filling with appropriate expansion medium. Presented is a solution with a high-precision
optical micrometer thrubeam laser sensor and a reference crevice, enabling compensation of undesired rotations.
With this system is possible to measure expansions down to a few micrometers accurate. All measurements were
compared to a mechanical, calibrated micrometer.
Keywords: mechanical thermostat, expansion medium, optical micrometer, precise measurements, mechanical
micrometer,
 1 Introduction
A proper regulation of temperature
is in present time very important, be-
cause it can be found everywhere: at
home, at work, in industry, etc. All
devices that enables regulation of
temperature have a name: a thermo-
stat.
The thermostat is a device for regu-
lation of the temperature in a system
in a way that the system’s tempera-
ture is maintained near a de-sired
temperature (Snajder, 1982). The
thermostat does this by controlling
the flow of heat energy in or out of
the system. That is, the thermostat
switches heating or cooling devices
on or off as needed to maintain the
correct temperature. Thermostats
Figure 1. Description of a mechani-
cal thermostat 
Figure 2. Filling sequence of  fi lling 
machine 
MERITVE
318
Ventil 13 /2007/ 5

For thermostat that switches at set
temperature is important that the pro-
cess of expansion medium filling is
controlled, so that diaphragm is pro-
perlyexpanded for desired thermostat
working range.
This must be done at proper tempe-
rature and fill pressure. In the cur-
rent production line, the temperatu-
re of the expansion medium can not
be controlled to full-fill the predefi-
ned temperature and this temperatu-
re difference must be compensated
with higher or lower filling pressu-
res, which is set manually.
Before filling with expansion me-
dium, the temperature sensor is
opened at the probe side, enabling
expansion medium to be filled. Fil-
ling procedure is done on the ro-
tary machines, where many sensors
can be filled at once. Installing of
the sensors into the filling machine
special heads is done manually. The
four stages of filling that follow are
evident in fi gure 2 and in fi gure 3.
First stage is vacuuming of the sen-
sor, where it is tested for airtightness.
Follows a filling procedure with the
expansion of the diaphragm depen-
dent on the medium filling pressure.
After a few cycles, expansion is stop-
ped, some types of the sensors stay
in filling heads and some are pu-
shed out, but all are squeezed with
a tongs to be closed. After this stage,
electrowelding is performed to relia-
bly close the sensor. At the end each
sensor is automatically removed from
the filling machine. The maximum
expansion of the diaphragm during
filling, for all types of thermostats, is
from 0.046 mm to 0.46 mm.
Our research team worked on a
task to find a proper technical so-
lution for measuring expansion of
the diaphragm in all four the most
important stages of filling procedu-
res of the thermostat sensor on the
current filling machines. From upper
information the measuring system
must be able to measure expansion
to less than 0.01 mm and without
major remaking of the current filling
machines.
In the work described in this article,
we have made a series of comparison
measurements, utilized high-accuracy
sensors and some simple mathema-
tical calculations. To check the mea-
suring method, some mechanical mi-
crometer comparison measurements
were performed and analyzed. The fi-
nal mathematical analysis is also pre-
sented at the end, but only the main
idea with a basic
equation.
 2 Diaphragm 
expansion 
measurement 
approach
Due to a fact that
current filling ma-
chines should not
be changed or
modified consi-
derably, an ap-
proach presented
in fi gure 4 for
performing mea-
surements was
tested. The who-
le measurement system consist of a
fixed curtain laser micrometer, expan-
ding diaphragm, the fixation object
and the reference object. To measure
expansion in four points at the machi-
ne, also four measuring systems would
be needed, as can be seen on figure 2,
marked with a letter L. These sensors
are fixed and are not part of the rotary
machine, only sensor diaphragms are
rotated repeatedly into the mea-suring
position. At the vacuuming phase, a
hole or crevice between the diaphragm
ceramic button and a reference object
would be measured and set to 0.000
mm initial value. In all other positions,
the same distance would be measured.
Since only first and last position is nee-
ded for appropriate pressure regula-
tion, the rest two positions are needed
for additional filling information. For
this kind of measurements it is impor-
tant that the machine places all mea-
sured heads in the same position. This
is done completely mechanically and
very accurately.
 3 Hardware
The following equipment was used
(Figure 5):
– laser optical micrometer sensor
Keyence,
– reference mechanical microm-
eter Mitutuyo,
– Epson robot.
3.1 Laser optical micrometer 
sensor Keyence
For testing purpose a local represen-
tative for Keyence Corporation lent
us the optical mi-crometer sensor
LS-5041 with LS-5501 controller.
(Keyence, 2006). Measurement sy-
stem consists from two parts, tran-
smitter and receiver, joined in a metal
rod setting, both parts being mounted
160 ± 40 mm from each other. Light
source is a visible red semiconductor
laser with 670 nm wavelength, Class
2 by IEC specifications. Measuring
range is from 0.2 mm to 40 mm, with
measurement accuracy of ± 2 μm
and repeatability of ± 0.3 μm.
3.2 Reference micrometer
Every measurement should be com-
pared to a higher class calibrated
measurement device. We used a
mechanical micrometer manufactu-
red by Mitutuyo Corporation (Mitu-
toyo, 2006). It has a mark 293-666,
measuring range of 30 mm, linear
scale with a resolution of 1 μm, error
limit of 2 μm, flatness less than 0.3
μm and with parallelism less than 2
μm. It has a digital data output, non
rotating spindle and measuring faces
are made out of a carbide. To imitate
real conditions a ceramic button was
attached to one of the tips.
Figure 3. Filling diagram of the thermostat sensor 
MERITVE
319
Ventil 13 /2007/ 5

3.3 Epson robot
The SCARA Epson E2S651 robot with
4 DOF (RRTR - Rotation, Rotation,
Translation, Rota-tion) is frequently
used for automation of assembly in in-
dustrial processes. It has a cylindrical
working range with radius 280 mm
to 650 mm. The velocity can be up
to 6300 mm/s for the first and second
axes, 1100 mm/s for the Z-axis and
1870°/s for the rotational U-axis. The
repeatability specifications are very
good: 15 μm for the first and second
axes, 10 μm for the Z-axis and 0.02°
for the U-axis. The Epson robot was
used as an carrier for the laser measu-
rement sensor.
 4 Methodology
4.1 Variable crevice width
Measurements of a variable crevice
width were performed to determine
if the selected laser micrometer is ap-
propriate for desired measurements.
On the end-effector of the robot a
laser measuring system was attached.
Robot use was necessary for positio-
ning of the laser system and also for
performing some additional tests.
In front of the robot a special table
was set and with another laser system
levelled to 0.1 mm. On the table a
reference micrometer was attached
with a few cramps and additional
metallic parts. One of the microme-
ter measuring faces was equipped
with a ceramic button to simulate
real conditions of the diaphragm
(Figure 5). Surrounding temperature
was stable at 23 °C.
Both, micrometer and laser control-
ler, were connected to the personal
computer via serial port for data sam-
pling with frequency of 2 Hz. The ini-
tial crevice width was set to 0.3 mm.
At this width the whole system was
positioned and rotated to laser sensor
returning as close as 0.3 mm. This
procedure involved laser system po-
sitioning and rotating as well as mi-
crometer levelling to set micrometer
faces as parallel to the laser curtain as
possible. During measurements, the
crevice width was manually changed
by hand with a step of 0.100 mm to
the final width of 2.000 mm. At every
step of 0.100 mm a small pause was
made to sample a few measurements.
Four series of measurements were
performed, two while enlarging the
crevice width and two while redu-
cing the width. This was necessary to
check possible laser micrometer hy-
steresis existence.
4.2 Stability of fi xed width of 
micrometer crevice
The measurement conditions and
equipment in this set of tests did not
differ from conditions in previous
test. Measurements were performed
to determine if laser measurements
of fixed width of mechanical mi-
crometer crevice is stable in longer
period. Four series of measurements
were performed, where the width of
the micrometer crevice was 0.500
mm, 1.000 mm, 1.500 mm and
2.000 mm. The appropriate width
was set and left for approximately
15 minutes with sample frequency
of micrometer and curtain laser sen-
sor set at 2 Hz.
4.3 Stability of moving laser 
system when width of mi-
crometer crevice was fi xed
These measurements are totaly equal
to the previous, except the laser mi-
crometer was moved by the Epson ro-
bot. These tests should verify if the la-
ser micrometer system shows different
measurement values if the mechanical
micrometer crevice lying in different
positions of a laser curtain. Again, four
fixed width of micrometer crevice was
set: 0.500 mm, 1.000 mm, 1.500 mm
and 2.000 mm. The laser system cur-
tain width was large enough to move
laser system for 20 mm in 1 mm step.
 5 Results
5.1 Variable crevice width
Figures 6 and 7 show measurement
results for conditions when reference
micrometer crevice width was chan-
ging - enlarging or reducing. Both fi-
gures show on the horizontal axis the
value measured by reference micro-
meter and on vertical axis the diffe-
rence between the micrometer value
and value measured by laser curtain
sensor. All measured laser values at
certain reference width were avera-
ged and for this reason also values in
range of 1 μm arise. As can be seen,
the diference on figure 6 is between
34 μm and 38 μm and has a positive
tendency, probably originating from
a temperature influence or even from
the human manually rotating micro-
meter spindle. On figure 7 difference
is between 38 μm and 40 μm.
It can be seen that the error value is
quite constant in a range from 0.3
mm to 2 mm, between 36 μm and
39 μm. This is a systematic error,
probably originating from parallel
misalignment of the mechanical mi-
crometer faces, influencing the laser
micrometer measurements. When
the relative measurement values are
observed the differences are below
the desired 0.01 mm, having values
between 2 μm and 4 μm.
Figure 4. Tested measuring principle
Figure 5. Measurement equipment
MERITVE
320
Ventil 13 /2007/ 5

5.2 Stability of fi xed width of 
micrometer crevice
Figures 8 and 9 show the stability of
the fixed width of micrometer crevi-
ce over longer period of time. Hori-
zontal axis represents samples with
frequency of 2 Hz while the vertical
axis shows difference between refe-
rence micrometer value and a value
measured by laser system.
Figure 8 shows results for crevice
width of 1 mm and figure 9 for width
of 2 mm. Again systematic error of
cca. 35 μm is present. Observing the
value over the whole time period,
very stable measurement value can
be observed not changing more than
1 μm.
5.3 Stability of 
moving laser 
system when 
width of 
micrometer 
crevice was 
fi xed
Figure 10 shows
results of measure-
ments where laser
system was moved
by the robot at
some fixed width
of micrometer cre-
vice. Horizontal
axis represents the
position of the la-
ser system relative
to the start posi-
tion, while the vertical axis shows
difference between reference micro-
meter value and a value measured
by laser system.
Figure 6. Difference between mechanical and laser mi-
crometer when crevice width was enlarged 
Figure 7. Difference between mechanical and laser mi-
crometer when crevice width was reduced 
Figure 8. Difference between mechanical and laser mi-
crometer with crevice fi xed width at 1 mm 
Figure 9. Difference between mechanical and laser mi-
crometer with crevice fi xed width at 2 mm
Figure 10. Difference between mechanical and laser 
micrometer with crevice fi xed width and laser system 
being moved
MERITVE
321
Ventil 13 /2007/ 5

Legend shows which bar column be-
longs to a certain crevice width.
Again systematic error exists having
value from 35 μmt o3 9μm. From
the measured data can be concluded
that moving the laser do not influen-
ce measurement values compared to
the case when it was fixed.
 6 Compensation of 
parallel misalignment 
between laser micrometer 
curtain and measuring object
6.1 Theory
Tests and results above demonstrated
that is possible, by using tested laser
micrometer curtain sensor, to mea-
sure the expansion of the diaphragm 
in the process of thermostat sensor
filling. These results also showed that
perfect parallelism between ceramic
button attached on the diaphragm 
and the reference object can ba a
problem if various sensors are used.
This problem does not fade away if
some reference crevice is used for 
calibration and then all four sensors
are levelled in a way that the error,
due parallel misalignment, is on all
stages the same. It might happen that 
all diaphragms are not completely
the same in dimension, the fixation
rod is not bend by the same angle,
etc. For this reason it is necessary to 
be able to measure or calculate and
compensate the diferences in rota-
tion that can occur during manufac-
turing.
6.2 Possible solution
In the figure 4 one reference object
can be seen. In order to compensate
for diaphragm errors then only one
reference object is not enough, an
additional reference crevice must be
used.
A situation is sketched in fi gure 11,
where diaphragm fixation object,
middle cube and right reference ce-
ramic are fixed together.
In the upper right-hand part of the
figure is a ceramic button, part of the
expanding diaphragm. The radius of
this round button is marked with h/2.
In the upper middle part is a referen-
ce cube with a half of dimension of
L/2 and D. The dimension dX
1
is a
measured crevice dimension, when
no rotation error is present. In the
upper righthand part of the figure is
another ceramic rounded button, but
this one is fixed. Dimension dX
2
is
fixed and well known or premeasu-
red with some high accurate device.
At the bottom of the figure is de-
picted a situation, where the whole
system is rotated over the rotation
point. In the real situation both,
diaphragm and laser system, can be
rotated, but here only the diaphragm
system with appropriate mechanical
parts are virtually rotated. If some
rotation is applied then dimension
dX
1
is changed, to the dX’
1
. Also the
measured dimension dX
2
changes to
dX’
2
.
6.3 Mathematical analysis
In figure 11 we can see that four ed-
ges of present objects are marked
with a dot or a point. These points
have a coordinates dependant to the
rota-tional point coordinate system:
P1=(0, -h/2)
P2=(dX
1
, L/2)
P3=(dX
1
+ D, -L/2)
P4=(dX
1
+ D + dX
2
, h/2)
When the dimension of the second
or reference crevice is premeasured
and then the same crevice is measu-
red at some rotation, then is possi-
ble to calculate the rotational angle
of the whole system for one dimen-
sion. After some simple mathema-
tical calculations, mainly by using
homogeneous transformation matri-
ces (Lenarcic and Bajd, 2003), the
rotational angle can be calculated,
by getting two solutions, only the
smaller angle is the right one. When
this angle is calculated it is possible
to calculate the crevice width dX
1
(Equation 1), even if width dX’
1
was
measured by curtain laser system.
Figure 11. Difference between micrometer and laser system when crevice 
width was reduced 
MERITVE
322
Ventil 13 /2007/ 5
(1)

 7 Conclusion
The manufacturing of expanding me-
dium sensor thermostats is a very dif-
fcult process. As pre- sented here, the
process of filling of the sensor system
has to be accurate to a few hundre-
ds of a millimeter. That is why is most
necessary to equip current filling ma-
chines with appropriate high-tech and
precise measurement systems in order
to monitor the production quality and
be able to respond and control the ne-
cessary process parameters.
The presented idea was fully tested in
the laboratory. The results show that
Keyence laser cur- tain micrometer
system enables measurement of de-
sired parameters with an accuracy
of a few micrometers. This system is
also very stable over longer period
of time. The system is also inde-pen-
dent to position of the crevice inside
the laser curtain.
In the tests also demonstrated that
the reference system should be pa-
rallel to the diaphragm ce- ramic
button for accurate measurements.
This is impossible in production line
and for this reason a solution with
one variable and one reference, di-
mensionally known crevice is also
presented, where simple mathema-
tical analysis is needed.
References
[1] Keyence (2006). http://www.
keyence.com. Internet link.
[2] Lenarcic, J. and T. Bajd (2003).
Robotski mehanizmi. Fakulteta
za elektrotehniko. Ljubljana,
Slovenia.
[3] Mitutoyo (2006). http://www.mi-
tutoyo.com. Technical report.
[4] Snajder, J. (1982). Avtomatiza-
cija umerjanja termostatov. RSS.
Ljubljana, Slovenia.
Uporaba optiÏnega mikrometra za merjenje raztezka membrane mehanskega termostata
Razširjeni povzetek
îlanek opisuje testiranje idejne zasnove sistema za nadzor oziroma merjenje raztezanja membrane me-
hanskega termostata v procesu polnjenja sistema diastata z ustreznim oljem. Sistem diastata je sestavljen iz
membrane in temperaturnega ïutila, ki ju povezuje tanka kovinska cevka, imenovana kapilara. Diastat se v
neki proizvodni fazi napolni s posebnim oljem, ki služi kot razteznostni medij. Za pravilno delovanje oziro-
ma preklapljanje stikala mehanskega termostata pri nastavljeni temperaturi je najbolj pomembno, da je pro-
ces polnjenja diastata s posebnim oljem nadzorovan in se membrana v tem postopku raztegne za doloïeno
razdaljo. To razdaljo je potrebno doseïi na nekaj stotink milimetra natanïno.
V ïlanku predstavljamo rešitev, za katero smo kot merilnik raztezka membrane uporabili zelo natanïen laser-
ski zavesni mikrometer podjetja Keyence LS-5041 s krmilnikom LS-5501, raztezek membrane pa smo simuli-
rali z referenïnim kljunastim mikrometrom Mitutoyo 293-666. Zavesni merilnik Keyence ima definirano na-
tanïnost ±2 μm in ponovljivost ±0,3 μm. Referenïni merilnik Mitutoyo ima merilno napako 2 μm, podatek o
ponovljivosti pa ni bil podan. Opisani merilni sistem bi namestili poleg obstojeïe rotirajoïe polnilne naprave,
saj zaradi vrtenja in izredno malo prostora namestitev merilnikov na samo napravo ni mogoïa. Testiranja so
pokazala, da uporabljeni merilnik Keyence ni samo natanïen, ampak so meritve tudi zelo stabilne v daljšem
ïasovnem obdobju, položaj merjene špranje v merilnem obmoïju laserja pa tudi ne vpliva na natanïnost
meritev, saj se vse merjene vrednosti nahajajo znotraj podanega obmoïja ±2 μm.
Žal pa so poskusi pokazali, da merjenje širine ene špranje ni dovolj, da bi lahko kompenzirali nepravokot-
nost laserske zavese merilnika Keyence in špranje, ki smo jo ustvarili z referenïnim kljunastim mikrometrom
Mitutoyo. Zato je za kompenzacijo potrebno uporabiti še dodatno špranjo. Njena velikost pa mora biti znana
vnaprej. Z meritvijo te špranje in primerjavo z njeno dejansko širino pa lahko izraïunamo naklon oziroma
nepravokotnost v vseh potrebnih oseh. V ïlanku podajamo tudi grafiïno ponazoritev in osnove matematiïne
analize za kompenzacijo nepravokotnosti med lasersko zaveso in špranjo.
IzvleÏek: V ïlanku predstavljamo nekaj zaïetnih meritev raztezanja membrane mehanskega termostata v
procesu polnjenja s posebnim oljem. Prikazujemo rešitev z zelo natanïnim laserskim zavesnim mikrometrom
in referenïnim mehanskim kljunastim mikrometrom. S tem sistemom je mogoïe meriti raztezke membrane
na nekaj mikrometrov natanïno. Vse meritve smo primerjali z referenïnim mehanskim mikrometrom.
KljuÏne besede: mehanski termostat, polnilno olje, laserski mikrometer, natanïne meritve, mehanski mikro-
meter,
MERITVE
323
Ventil 13 /2007/ 5