|
Thermographic cameras detect an infrared
radiation image of object surfaces measured in a certain wavelength band. The
required surface temperature can be precisely evaluated only if the following
process parameters are known: emissivity of the surface measured, reflection
radiation from the ambient and transparency of the environment between the
object and the camera [1]. The lack of knowledge on the mentioned parameters
values can results in a measurement error in tens or even hundreds percent of
the true temperature especially in the case of glossy surfaces.
Usually the object emissivity (temperature variation and spatial homogeneity)
has the greatest influence on the non-contact measurement accuracy. There are
commonly used two approaches to overcome the emissivity problem. The first
solution is represented by a high emissivity paint applied on the measured
surface to attain the known uniform radiation from all over the surface.
However this can not be used if the paint injures the object or if the paint
due to the different radiation compared to the original surface changes the
surface temperature being measured [1],[2]. The second approach stands for a
direct emissivity measurement. In this case the surface temperature can be
measured locally by a contact method as well. The emisivity is evaluated on
the radiation detecting device in order to the non-contactly measured
temperature reaches the temperature measured contactly [3]. This method
however has its limitation in cases where the emissivity spatially
considerably changes as on the electronic printed circuit boards for instance.
A printed circuit board consists of different components made of materials
with strongly varying radiation properties. Therefore almost each of the point
in the thermal image has to be considered with individual emissivity value
[1]. Additional difficulties are brought about the soldering process that has
to be measured. The soldering paste changes its emissivity during melting and
spreading over the joints soldered.
The work presented deals with the thermographic measurement of the hot-air
soldering process. The objective is to identify temperatures on soldered
joints and on insulated parts of the printed circuit board. The goal is to
optimize the technological process by comparing alternatives to find the best
from the points of the resulting solder joint and the process symmetry [7].
As the solution method, the industrial thermographic system making the direct
transient temperature field measurement possible was used. The thermography
application in this particular case has to resolve two basic tasks. Firstly,
it is the sample small dimensions requiring the additional optical system of
the camera. Secondly, it is the printed board emissivity spatial variation
that was additionally measured by the known uniform temperature method.
| 2. Experiment and solving procedures |
2.1. Hot-air soldering technology
The Fig.1 shows the technological set-up for the hot-air soldering of joints
on the sensor printed circuit board [7]. The hot air flows with varying
temperature from the nozzle. The soldering paste is applied on the three
square joints before the sensor is inserted into the holder. During the
hot-air affection the paste melts and spreads over the joint and establishes
the joint. The other electronic elements on the board are insulates by the
shutter preventing the hot-air to flow over this part of the sensor.
 |
 |
| (a) |
(b) |
Fig.1: Technology for the hot-air
soldering - the soldering workplace detail
involving the hot-air nozzle and the sensor in the holder (with the shutter
in the open state), - sensor detail
Fig.2: Thermographic measurement of
the hot-air soldering technology
2.2. Thermography measurement of the soldering process
Fig.2 shows the schematic arrangement of the thermographic measurement of the
soldering process [7]. The thermographic camera ThermaCAM SC2000 [3] is placed
in the holder above to maintain a stable view during the whole process. The
temperature on the joints soldered is visible by a narrow slit between the
hot-air nozzle vent and the shutter. The camera is equipped with the 64x48
mm/150 mm close-up lens with the resolution 0.2 mm/px. Dynamic temperature
field is sampled directly to the computer attached at 5 Hz rate.
2.3. Printed circuit board emissivity measurement
The emissivity of the printed circuit board (Fig.3) is measured by the known
uniform temperature method [6]. All the object is heated up to a certain
temperature. The thermo-graphic system measures the object radiation image.
The variations in temperature observed are caused by differences in the
emissivity value.
The emissivity measurement workplace [5], [6] is shown in Fig.3. The plane
temperature source is represented by the titanium hot plate PZ 28-3 TPD (Detlef
Gestigkeit GmbH) with the maximum temperature 650 ºC which is precisely set by
the PR5 3T (Detlef Gestigkeit GmbH) controller. The plate surface and ambient
temperatures are additionaly measured by thermocouples. The printed circuit
board emissivity is evaluated in the temperature range from 100 °C to 200 °C.
|
 |
 |
| (b) |
(c) |
Fig.3: Emissivity measurement
workplace -(a) schematic arrangement, - (b) photo,
-(c) printed circuit board measured
A computer controlled ThermaCAM SC2000 [3] with the close-up
lens attached is used for the IR radiation measurement. The emissivity of the
circuit board individual parts is evaluated by the ThermaCAM Researcher
software [4]. Emissivity variations in the selected area can be analyzed by
the software functions as a deviation of max/min temperature from the average
area temperature or as a histogram of temperatures in the area.
2.4 Dynamic surface temperature field evaluation
Temperatures are evaluated and analyzed in the Flir ThermaCAM Researcher [4]
software environment. This software is capable to consider different
emissivity values for selected points or areas of the thermogram but not as an
emissivity map of the object. Additionaly the transient thermograms are viewed
and exported as infrared video sequences.
3.1 Emissivity
The measurement reveals the emissivity variation not only among the circuit
board parts made of different materials (epoxy substrate, metalic joints),
however even in the area of the same material. The average emissivity value of
the three soldered joint square areas is e = 0.20, instead of the emissivity e
= 0.80 for the board. The effect of the emissivity variation at the true
temperature 200 °C is shown in Fig.4(a) for e = 0.80. Differences for the
circuit board ranges in a relatively narrow interval about ± 5 K. Fig.5(b)
displays temperatures for e = 0.20. The error in the square joint area
temperature is about ± 25 K [6].
As for the soldering process evaluation, the unpleasant fact is brought by the
soldering paste emisivity changes. The paste initial emissivity is of the same
value as the printed circuit board (e = 0.80). During the soldering process
the emissivity non-uniformly changes to e = 0.20 and the paste is not visible
on the joint area in the measured infrared wavelength band.
 |
 |
| (a) |
(b) |
Fig.4: The effect of emissivity
variation on the circuit board temperature - true temperature 200 °C,
emissivity (a) e = 0.80, (b) e = 0.20
3.2 Joints Temperature Evolution
Individual alternatives of the technology are compared in temperature of
the sensor circuit board area near the joints soldered. It is the area AR2
shown in Fig.5 just behind the tin square of the middle joint. This area is
the only visible with the temporaly stable emissivity value for all the
samples investigated.
The effect of the soldering paste emissivity changes and the joint
emissivity variations on the temperature evolutions are shown in Fig.5 to
demonstate potential errors if the constant emissivity value is considered.
The Fig.5 shows the max, min and average temperature for the middle joint AR1
evaluated for e = 0.20. The joint temperature is compared with the board
temperature of the area AR2 used for the alternatives comparison as mentioned
above.
Fig.5: Max, min and average
temperature evaluated for line analysis AR1 on the middle joint (e = 0.20) and
AR2 on the circuit board (e = 0.80) - (open shutter alternative).
3.3 Electronic Circuit Temperature
Loading
Circuit board temperature behind the insulated titanium or ceramic shutter is
evaluated as well. Fig.6 shows the comparison of the average temperature just
behind the shutter (Fig.2). The maximum temperature reached for the titanium
shutter is bellow 200 °C, for the ceramic shutter is above 300 °C. This result
disqualifies the second one for the technological use.
Fig.6: Temperature on the circuit
board behind the titanium and ceramic
insulation shutter 3.4
Soldering process comparison
The emissivity changes of the soldering paste produce difficulties for the
temperature evaluation. On contrary they are very useful to investigate the
course of the melting process. The contrast between the paste and the joint
surface in the infrared images compared to the visual ones makes the changes
in the soldering process highly visible.
Three characteristic times of the process can be recognized on the
thermography video-sequences: the first visible motion of the paste on the
joint surface, the considerable change in the paste melting accompanied by the
emissivity rapid changes and the end of the process with no further changes on
the paste melting process.
Fig.7: Soldering paste melting process
between the 33.2 s and 42.0 s in the infrared projection (open shutter
alternative) Fig.7 shows infrared
images of the circuit board during the paste melting for the open shutter
configuration as an example. One can see the first change on the middle joint,
later on the top one and finally on the bottom one.
Thermographic measurement of the hot-air
soldering process has been done on the printed circuit joints. Prior to the
transient temperature fields evaluation, the emissivity distribution has been
measured by the known uniform temperature method. The average emissivity value
of tin parts including soldered joints is 0.20, the emissivity of the circuit
board is 0.80. The soldering paste changes its emissivity non-uniformly from
0.80 to 0.20 during melting and spreading over the joint. This is the main
reason why the temperature evolutions from individual points of the thermogram
in the joint areas are accompanied by a massive error if the spatial or
temporal constant emissivity is considered. However the emissivity changes
allow us to reliably identify characteristic times of the soldering process -
the beginning, considerable change and the end of the paste melting. This
three times evaluated for the three joints on the circuit board and infrared
video sequences of the process have been used to specify the best
technological alternative from the set being measured.
This paper is based upon work sponsored by the
Ministry of Education of the Czech Republic under research and development
project LN00B084.
[1] Nondestructive testing
handbook, Volume 3. American society for nondestructive testing, Arlingate
Lane, Columbus, USA, 2001. 718 p. ISBN 1-57117-044-8
[2] Holst G.H.. Common sense approach to thermal imaging. SPIE
Optical engineering press, Bellingham, Washington USA, 2000. 377 p. ISBN:
0-8194-3722-0
[3] ThermaCAM SC2000 operator’s manual, FLIR Systems, Danderyd,
Sweden, 1999. 58 p. (http://www.flirthermography.com/cameras/camera/1009/)
[4] ThermaCAM Researcher operator’s manual, FLIR Systems,
Danderyd, Sweden, 2000. 132 p. (http://www.flirthermography.com/software/SWF004.asp)
[5] Kunes J., Honner M., Litos P. Infrared thermography
measurement in thermomechanics of technological processes. In Proceedings of
NTC. 1st edition. Pilsen: UWB NTC, 2002. pp. 51-58. ISBN 80-7082-891-9.
[6] Litos P.: Thermography and infradetectors in physics of
technological processes. MSc. Thesis. University of West Bohemia, Pilsen,
2002. 87 p. (in Czech)
| 
