Process optimization in a filament yarn plant

by Dr. Serap Dönmez Kretzschmar and Richard Furter,

Uster Technologies AG, Uster, Switzerland.
1. Introduction In this article we will concentrate on the melt spinning
process. Melt-spinning is described as the simplest method of fil-
Natural fibers have been used since thousands of years but ament yarn manufacturing because it does not involve problems
there are not enough of them to meet the demands of today’s associated with the use of solvents [2].
world population [1]. Therefore the invention of synthetic fibers Many scientists have worked on the analyses of melt spin-
was undoubtedly one of the most important discoveries of the ning, both theoretically and experimentally [2,4,5,6,7].The spinning
twentieth century. of synthetic fibers is sensitive to any variation of process parame-
Another point is the competition in the textile sector. In order ters, as for example, the temperature of the polymer melt or the
to satisfy new demands of the customer, various improvements in temperature of the quench air of the spinning unit [6]. Particularly,
the production technology of filament yarns and their character- the conditions in the quenching zone influence the formation of a
istics have been made and the filament yarn industry has become suitable filament yarn [5].
more and more important. As we can see in Fig. 1, in melt spinning there are mainly three
This article deals with the measurement of mass variations of stages; hopper, spinning and wind-up [3].
filament yarns and with the interpretation of test results. In spun
In the hopper stage, the raw material is stored, melted and
yarns, yarn evenness is inevitable because it arises from the fun-
then processed. The starting material for melt-spinning is polymer
damental nature of the fibers, their manufacturing methods and
granules or chips and they are first dried and then melted in the
the arrangement of fibers in the yarn.
extruder. Today in modern plants, polyester and polyamide are
In case of filament yarns, granulate heterogeneity, spinning produced in continuous polymerization units in which the melt is
process irregularity, faults in subsequent yarn cooling and winding directly transported from the final polymerizer to the melt-spin-
operations, together with machine defects and drafting faults ning unit. Polypropylene is different because polymerization leads
quickly lead to mass variations which can affect the mass of the to a solid product. For this reason it is separated from the spinning
yarn over thousands of meters of material, because the produc- process [2].
tion speed is very high. Such variations often cause difficulties in
In the spinning stage, every time the same amount of homog-
the subsequent processes like draw-twisting, texturing or dyeing.
enized and filtered melt, which is transferred from extruder to the
These mass variations reduce the quality of the yarn, and a com-
spinning pumps, is pressed through the orifices in the same
prehensive evenness testing is a must for a quality control system
[2] amount of time (Fig. 1). After the spinning heads, at a distance of
.
5 to 20 cm below the orifices, in the quench air duct, the fila-
Uster Technologies has been manufacturing textile testing ments spun from the melt are cooled by a jet of air and freeze.
systems for measuring filament yarns since 1955. The USTER® When using a multiple of orifices in the form of a spinneret, the
TESTER 5-C800 for filament yarns was introduced in the market bundle of filaments can be drawn off as an undrawn or partially
as the 5th generation in 2005. The quality characteristics of fila- drawn filament yarn [3].
ment yarns can be quickly assessed by means of this testing
In the wind-up stage, after leaving the quench air duct, the
system. The test results can also be used to judge the conse-
filament material is drawn over preparation rollers through an oil-
quences on subsequent processes.
fat-emulsion (spin finish). The filament bundle, which has passed
2. Frequent sources of faults in the melt spinning the spin finish application, is wound on spinning packages and
can be transported in this state to other processing machines [3].
process
The localization of faults during this spinning process will be
Filament yarns can be manufactured according to different illustrated by taking the example of polyamide yarn manufactur-
spinning methods. In Table 1 these methods are given: ing (Fig. 1) [3].

Table-1:Different spinning methods Raw material

hopper
of filament yarns [2]
Solution spin-
Melt-spinning Range of long-term variations
ning Hopper
or periodic faults with long
Wet-spinning: Wet-spinning: Extruder wave-lengths.
Dry-spinning
Direct solvent Derivative-based Hopper stage
Spinning
Cellulose
Polyamide 66 Acrylic Viscose rayon pump
diacetate Quench
air duct
Cellulose
Polyamide 6 Modacrylic
triacetate Spinning stage
Range of short-term variations
Acrylic Rayon Preparation or periodic faults with short
wave-lengths.
Spinning package
Polyester Polyurethane Polyurethane
Polyvinyl Polyvinyl Fig. 1:The localization of most
Polypropylene common faults during the melt
chloride alcohol
Aromatic Wind-up spinning process [3].
Polyethylene Chlorinated PVC
polyamide

44 PTJ September 2008

 In Fig. 1, the faults resulting from certain machine groups as Table-2: Numeric results of a filament yarn test
indicated with a “circle” and a “triangle” can be determined with
the help of the diagram, spectrogram, evenness value, variance- CVm CVm CVm CVm Rel. mMin mMax mMin mMax
U% CVm
length curve and relative count. The faults at machine groups Nr 1m 3m 10m 50m Cnt 1m 1m 10m 10m
which are indicated with a “triangle” can also be determined with % %
% % % % % % % % %
measurements undertaken throughout one full package with the
help of the diagram, spectrogram, evenness value, variance- 1/1 1.31 1.66 1.26 1.20 1.01 0.26 0.2 -3.5 3.8 -2.0 3.1
length curve and relative count [3]. 1/2 1.12 1.43 1.05 0.97 0.75 0.26 0.2 -3.3 3.4 -1.8 1.5
The maximum mass deviations from the nominal value within 1/3 1.16 1.49 1.09 1.03 0.84 0.24 0.0 -3.1 3.6 -1.8 2.1
the test length can also be measured [3].
1/4 1.10 1.42 0.99 0.93 0.77 0.28 0.0 -3.2 3.3 -1.9 1.9
3. Evenness testing of filament yarns and USTER® 1/5 1.13 1.45 1.04 0.98 0.83 0.33 0.2 -3.3 3.4 -1.9 2.2
STATISTICS for filament yarns 1/6 1.15 1.50 1.03 0.96 0.75 0.22 0.0 -3.1 3.4 -1.8 2.3
1/7 1.12 1.43 1.05 0.99 0.82 0.35 -0.1 -3.2 3.9 -2.1 2.6
As we have mentioned before, the quality characteristics of
filament yarns can be quickly determined by using USTER® 1/8 1.13 1.43 1.13 1.06 0.83 0.25 -0.3 -3.5 3.9 -2.1 1.8
TESTER 5-C800 for filament yarns. The test results which are 1/9 1.25 1.58 1.24 1.19 1.01 0.48 -0.1 -3.2 4.0 -2.2 2.5
obtained from this measuring system, can be evaluated both in 1/10 1.21 1.54 1.15 1.07 0.91 0.30 -0.2 -3.6 4.9 -2.5 2.7
graphical and numerical form.
Mean 1.17 1.49 1.10 1.04 0.85 0.30 -0.0 -3.3 3.8 -2.0 2.3
3.1. Diagram CV 5.7 5.4 8.3 9.1 11.4 25.5 0.2 5.4 12.2 11.8 20.8
s 0.07 0.08 0.09 0.09 0.10 0.08 0.2 0.2 0.5 0.2 0.5
The diagram is an important part of evenness test and pro- Q95 0.05 0.06 0.07 0.07 0.07 0.05 0.1 0.1 0.3 0.2 0.3
vides an enormous amount of details on the spinning process for
Max 1.31 1.66 1.26 1.20 1.01 0.48 0.2 -3.1 4.9 -1.8 3.1
a filament yarn specialist. Fig. 2 shows diagrams of two different
filament yarns. The lower filament yarn has a very high mass vari- Min 1.10 1.42 0.99 0.93 0.75 0.22 -0.3 -3.6 3.3 -2.5 1.5
ation compared to the upper yarn.
The column "Rel. Count" describes the relative fineness of
the yarn. The testing system calculates the mean of the yarn fine-
ness for the entire measuring series and always prints out zero as
a mean value. Afterwards, the system calculates the deviation of
each individual package relative to the mean. The basis for this
calculation is the capacitive measurement of the mass over the
entire test length. The columns mMin and mMax describe the
maximum deviation from the mean value during the tests.

Polyester, dtex 70f36 3.3. Spectrogram

The spectrogram as shown in Fig. 3 is a representation of
mass variation in the frequency domain, i.e. the measuring
system detects periodic mass variations. Fig. 3 shows the spectro-
gram of the polyester filament yarn, dtex 76f100, described in
Table 2.

Polyester, dtex 75f128

Fig.2: Diagram, filament yarns.

The evenness of the yarn of the upper diagram in Fig. 2 was

CVm = 1,15%, the evenness of the lower diagram was CVm =
2,60%. The source of the high mass variation of the lower dia-
gram was a significant problem with the air intensity and air guid- Fig. 3: Spectrogram,
ing in the quench air duct which causes intensive vibration during filament yarn.
the solidification of the filament yarn.
The spectrogram taken from the USTER® TESTER 5-C800
3.2. Numerical results of filament yarns (Fig. 3) shows a significant periodic fault with a wavelength of 1,2
m, which was caused by an eccentric spinning package during
Table 2 shows a selection of the result columns of a filament
wind-up. The second severe quality problem is shown in Fig. 3 as
yarn test. Yarn: Polyester, dtex 76f100. The test was carried out at
an increase of the spectrogram between 10 and 80 m. Such mass
10 POY packages, test length was 1000 m per package.
variations sometimes lead to misinterpretations if one only checks
The value U% is the evenness; the value CVm is the coeffi- the diagram because the variations look like strictly periodic faults
cient of variation of the yarn mass while the measuring system in the diagram. Only the spectrogram shows precisely what is
was set to "normal test". The values CVm 1 m, 3 m, 10 m and 50 happening. The origin is a non-optimized air stream in the quench
m represent the coefficient of variation of the yarn mass of vari- air duct (Fig. 4) as already mentioned above.
ous “cut lengths”.

PTJ September 2008 45

 The quench-
ing zone of the
spinning
machine is very
important (Fig.
4). A non-opti-
mized air system
in the quench air
duct is one of the
most frequent
sources of con-
siderable mass Fig. 6: USTER® STATISTICS, normal test.
variations of fila-
ment yarns. depends on the fineness (dtex) of the individual filaments in the
Since the take- filament bundle and on the amount of periodic and non-periodic
off of filament mass variations (Fig. 6). Fig. 6 can be used for partially oriented
yarns takes place yarn as well as for fully oriented yarn.
at very high
speed, the cool- 4. Conclusion
ing process in
the quench air As a result of continuous improvements in the filament yarn
duct has to be industry, the demand of reliable and reproducible test methods
efficient. If the for the filament yarn industry has also increased.
Fig. 4: Air stream in quench air duct (out-flow air stream is not Especially the yarn evenness is still a very important quality
quench) [2]. conducted prop-
parameter in the area of filament yarns since small mass variations
erly the individ-
can already have a considerable effect on the appearance of fab-
ual filaments start to vibrate. Because the filaments are not
rics, particularly after dyeing. As it is well-known, the evenness
solidified at this point of the manufacturing process, the vibra-
values of filament yarns can drop below CVm = 1%.
tions cause mass variations [3].
Fig. 5 is the recording of 10 spectrograms of the described fil- This means that even the smallest deviations can have an
ament yarn of 10 packages from the same spinning machine. All adverse effect on the product quality in the subsequent process-
the spectrograms show that the faults are common to all pack- ing. Uster Technologies began to test the evenness of filament
ages. yarns at a very early stage.
With the USTER TESTER 5-C800 for filament yarns, the qual-
ity characteristics of yarns can be assessed quickly and worldwide
compatibility of the same types of yarn can be guaranteed. This is
especially important because, in the high-performance produc-
tion of filament yarn spinning mills even a small reduction in qual-
ity can result in disastrous financial losses.

5. Literature
1- Schenek, A. “Chemical Fibers – the Replacement of cotton?
– Facts, Limits and Trends”, STF Anniversary Colloquium ‘99,
Swiss Textile, Clothing and Fashion College,
Fig. 5: Spectrograms of 10 spinning packages.
Switzerland,1999.
2- Gupta, V.B., Kothari V.K.,” Manufactured Fibre Technology”,
Spectrograms of filament yarns frequently have many peaks Chapman & Hall, 2-6 Boundary Row, London, First Edition,
which have to be interpreted correctly. Several peaks in the spec- 1997.
trogram do not necessarily mean that there are several manufac- 3- Uster Technologies AG, Application Manual, Testing of
turing problems. The correct interpretation of the peaks, however, Filament Yarns,V1.0, no.410 107-04020, August 2006.
can provide detailed information where manufacturing problems 4- Fourné, F., “Synthetische Fasern: Herstellung, Maschinen
exist. In order to find the correct origin of the manufacturing und Apparate, Eigenschaften; Handbuch für
problem, the USTER® TESTER 5-C800 also has a Knowledge Anlagenplanung, Maschinenkonstruktion und Betrieb”, Carl
Based System which simplifies the interpretation of the spectro- Hanser Verlag, München, Wien, 1995.
grams. 5- Harder, C. “Finite element analysis of melt spun yarn”,
Journal of Materials Processing Technology, 2001,118, 454-
3.4. Benchmarking for polyester and polyamide filament
459.
yarns 6- IDESAKI, A. et al., “Fine SiC fiber synthesized from organosil-
The USTER® TESTER 5-C800 also supports the user with icon polymers: relationship between spinning temperature
experience values which can be used for benchmarking and eval- and melt viscosity of precursor polymers, Journal of Materials
uation. Fig. 6 shows the USTER® STATISTICS of polyester and Science, 36, 5565 – 5569, 2001.
polyamide filament yarn tests. It represents the evaluation of 7- TAE HWAN OH et. al, “Numerical Simulation of the Melt
mass variations of various packages. The coefficient of variation Spinning of Hollow Fibers” , Textile Research Journal, 68, 6,
p449-56, 1998.

46 PTJ September 2008