Introduction

The highly competitive optical fibre cable business forces cable machinery manufactures to develop highly efficient equipment with the highest production capacity. With an on-line proof tester we can combine two separated processes, the fibre draw and the proof test. By using this approach the set-up times can be reduced, and therefore remarkable economical benefits can be achieved. There are some theoretical as well as practical problems concerning the on-line proof tester. The greatest problem has been the constraints on proof testing fibre immediately after the draw. The main issue is to maintain the mechanical reliability at the same high level as with proof testing fibre with a conventional proof tester. This is a combination of well-maintained proof tension and dwell time. Recovering from a random fibre break is the biggest challenge for the machinery. The on-line proof tester has to be able to handle fibre breaks at full line speed without disturbing the draw process. The reel change has to also be done without disturbing production.

Theory

The effect of water on the strength
When drawing fibre, the preform has to be heated to approximately 2,000°C. The fibre cools rapidly during the draw and remains dry. If the fracture strain of the fibre is measured immediately after the draw[1], an obvious decrease in strain during the time after draw is seen. The value of the fibre strength decreases in a similar way.

Figure 1: Fracture strain for dynamic fatigue measurements
of fibre as a function of time t after drawing from the preform[1]

This phenomenon is caused by stress corrosion, the effect of which begins immediately after stress is applied on the fibre, and moisture is diffused through the coating material to the surface of the glass. The water entry into a glass can promote structural relaxation, and this leads to strength reduction. Figure 1 shows that fracture strain balances in 50 minutes after the draw. This value is dependent on the coating material’s permeability for water. The coating material’s main functions are to protect the fibre from mechanical abrasion and to reduce microbending; however, it also works as a diffusion barrier against water. Figure 2 shows the raw data for strength as a function of time after changing the ambient environment for the acrylate, polyimide, and silicone-coated fibres[2]. A two-point bend apparatus was used to measure the strength of fibre and it was operated with a constant faceplate velocity of 5,000µm/s. Twenty samples were broken at each humidity level. For all the coatings, moisture penetrates on a time scale of ~102 - 103s[2]. It can be seen in Figure 1 that fracture strain balances simultaneously.

Figure 2: Strength as a function of time after suddenly changing
the ambient humidity for the diffusion of moisture into (•) and out of
(o) single-acrylate-coated fibres, polyimide fibres, and silicone-coated fibres[2]

If we consider proof testing fibre immediately after the draw, we have to take into account the fact that fibre is considerably dry (as in an inert environment), and the strength balances only after several minutes from the draw. In that case the proof test tension has to be multiple compared to a conventional proof tester, and that would cause the weak spot to fracture. Therefore, the chances of damaging the warm and soft coating are higher. The preferred manner would be to wait for some time after the draw and execute the proof test later.

The effect of the coating material on mechanical properties
Earlier the theory determined the strength of fibre almost entirely based on the strength of the glass. Due to changes in dynamic fatigue test results, the behaviour of the coating material in different strain rates has been studied. It has to be noted that 75% of the volume of fibre is coating material and only 25% is glass. Standards (IEC 60793-1-30, TIA/EIA-455-31C) state that proof stress shall be applied uniformly through the cross-sectional area of the test sample[3, 4]. When the results of dynamic fatigue tests are examined, it is seen that contribution of the secondary coating can significantly affect the dynamic fatigue measurement of optical fibre[5]. Test was performed using six different stress rates: 0.005%; 0.025%; 0.25%; 2.5%; 25% and 100% per minute of gauge length. The problem has been the so-called S-shape existence within the log-stress rate/log-break stress curve of the dynamic fatigue test. It has been proven that the strength of the fibre coating is strain rate dependent, and that the contribution of the secondary coating can significantly affect the dynamic fatigue measurement of optical fibre. The thicker the coating, the more load is exerted on the coating. The effect is also similar when using high-strain speeds; the faster the strain rate, the more load the coating carries[5].

Optical fibre manufactures commonly use the proof test tension of 0.69GPa, which correlates to elongation of 1%. Due to increased production speeds, the loading rate of the proof stress can be extremely high, around 150-250GPa/s. In such a case the coating can carry a substantial part of the proof tension during the proof test and the risk of inadequate stress exertion to glass fibre could be possible. Therefore it is recommend that higher proof tension be used with an on-line proof tester than with a conventional proof tester. Another remarkable issue is the temperature dependence of the viscoelastic properties of the coating[6]. When glass fibre is coated, it goes to the UV-curing unit, where the coating material is cured. The temperature of the coating can be over the Tg-temperature, and therefore the microstructure is relatively soft. When proof testing fibre immediately after curing, it may be possible to damage the fibre unless it is properly cooled.

Dwell time effect on lifetime estimations
Proof testing standards require that a specified tension or proof stress be applied sequentially along the full length of the fibre. Because the starting of the drawing process is slow, all possible interruptions during the draw have to be minimized. The on-line proof tester requires an automatic break recovery system, and because of that, the proof test area has to be level. There is no possibility to use any weight wheel, and so the fibre line goes trough from the breaking capstan to the drive capstan, and the proof tension is measured directly from the capstan. Therefore, the proof test time td becomes short, approximately 0.01s to 0.04s depending on the line speed. Kapron shows[7] that increasing the proof stress level increases fibre reliability. In addition, increasing the dwell time has a similar effect. For example, if a crack of initial strength S before proof testing survives the proof test, it is reduced to the strength Sp after proof testing as given by [7]

Equation 1

Here B is the crack strength preservation parameter or B-value; fibres with a higher value of this parameter will experience relative less weakening. In the above equation the effective proof time is given by [7]:

Equation 2

where n is the stress corrosion susceptibility parameter or n-value. Note in Equation 2 that the dwell time is the biggest contributor to the proof time. By Equation 1, the weakening of a crack clearly increases with the dwell time[7]. If the dwell time is shortened, the lifetime estimations decrease. Reliability in service is characterised by a lifetime or by a failure rate. For survival probability P, the fibre lifetime to failure is:

Equation 3

Figure 3 shows results from lifetime calculations, which were made with 0.69GPa proof stress. It is easy to see that when the dwell time is reduced, the lifetime estimations drop. When proof stress is doubled and the dwell time is maintained the same, the lifetime rises to a similar level than with common dwell times (0.1s, 0.5s, 1s, 2s).

Figure 3: 1,000km Lifetime versus failure probability for
dwell times of 0.01s, 0.02s, 0.03s, 0.04s. Proof test stress 0.69GPa

Figure 4: 1,000km Lifetime versus failure probability for
dwell times of 0.01s, 0.02s, 0.03s, 0.04s. Proof test stress 1.38GPa

Current industry standards use the power law form for lifetime prediction. It has to be taken into account that the power law theory has no physical significance and it gives an overly optimistic lifetime estimation compared to other forms based on chemical kinetics theories[8].

Possibilities for on-line proof tests

Based on the theoretical information presented above, three different approaches for on-line proof tests are suggested:

1. The moistening of the fibre with specified equipment immediately after the fibre comes out of the furnace. After UV-curing the fibre must be effectively cooled to room temperature, and after that the proof test can be executed;
2. Artificial ageing before the proof test. This method utilises an equipment that adjusts the pressure, temperature and moisture, and therefore accelerates the penetration of water through the coating to the surface of the glass. In addition, some chemicals can be used to activate the process. Before executing the proof test the fibre must be cooled to room temperature;
3. Using a very sophisticated accumulator between the drawing tower and the proof tester. This machine holds the fibre in ambient air for approximately 20-30 minutes regardless of the line speed. During this relaxation time the humidity and temperature of the fibre balances, and the proof test can be executed immediately afterwards.

Random fibre break

When proof testing fibre, there are occasional breaks depending on the quality of the drawing process. These occur approximately two or three times per 100km (with good quality fibre)[9]. In a conventional proof tester fibre break causes a stop, and the operator has to thread the fibre again to the machine to continue the proof test. This interruption is not acceptable with on-line proof testing because ramping down the drawing process is a slow process. The properties and functions of the on-line proof tester are described in patent FI108754. There are also some patens concerning the so-called air-blown fibre, for example US4691896, US6022620, US5046815. Blown fibre is used when installing optical fibre to buildings or under ground piping. An airflow conveys the fibre inside the tube. On the surface of the fibre are small glass spheres, which help the transportation of the fibre. Some antistatic agents are also added to the airflow to minimize static electricity. In addition, the guiding duct can be made of an antistatic material (surface resistivity =106W).

Figure 5: Test equipment

In the tests, we focused on handling fibre breaks in the proof test region (test arrangement shown in Figure 5). A guiding tube was located between the braking capstan and the drive capstan, and the proof tension was measured from the moving capstan with a load cell. Five different pipe materials were tested; glass acrylic, aluminum, inside polished steel, vinyl with carbon fibre and Teflon. The line speed was set to values from 250m/min to 1,750m/min, and the break recovery probability was calculated from ten breaks per measured speed.

Figure 6: Break recovery result

As is clearly seen in Figure 6, Teflon is the best material; all the other materials induce reduction to the break recovery probability. This is caused by several different phenomena. First, Teflon is a quite effective insulator, and therefore it easily creates a strong surface electricity field around it. This field normally has the same polarity with the fibre. When we measured the surface voltage of Teflon, the values ranged from +5kV to +20kV and varied excessively depending on the point of measurement on the tube. This charge works as on opposite force and helps to convey the fibre inside the tube. Second, the friction coefficient between the fibre and Teflon is extremely low, and therefore friction force breaks least compared to other materials. Third, the so-called "rubber band effect" at the proof test region interferes with the threading event. This is probably the reason why we did not manage to receive a 100% probability.

The elimination of static electricity
When the fibre is led through the belt capstan, the surface voltage rises from zero to approximately +1.4kV to +1.6kV. This value was measured with a non-contacting voltmeter, and the result must be regarded only as trend-setting. The high-surface voltage is a result of frictional electricity, which is created when the rubber belt and the steely capstan wheel touch the fibre. When we calculated the effect of static electricity, it became clear that the electrical force can push the fibre against the inside of the tube much stronger than the gravity force. Elimination of static electricity is crucial in order to convey the fibre inside the tube, and therefore static electricity has to be removed as soon as possible after the fibre comes out of the capstan. This is resolved by using ionizing air nozzles which blow ionized air into the guiding tube, or an ionizer which has corona points inside the guiding tube. In such a case it should be noted that the tube cannot be made of any conductive material, or otherwise it reduces the effect of the ionizer.

Conclusions

The strength of fibre balances after several minutes from the draw process. The moisture of the ambient environment penetrates trough the coating, and because of that, the strength of fibre can be considerably higher immediately after the draw than approximately 20-30 minutes after the draw. In this work we have suggested three possibilities to overcome this phenomenon. Because the strength of fibre is based on the crack growth of the weak spot, the dwell time of the proof have a strong effect on lifetime estimations. High-production speed and the mechanical construction of the on-line proof tester forces the dwell time to become short. If the commonly used proof tension 0.69GPa (100kpsi) is used, it is possible that the weak spots of the fibre do not grow enough and the lifetime is low. If we increase proof tension and calculate the lifetime by the power law, lifetime estimations increase. Therefore, we suggest that higher proof tension be used with short dwell times. Another remarkable issue is that the coating material’s dynamic tensile modulus can change according to the stress rate. If high-stress rates are used, it is possible that the coating carries a substantial part of the tension and the glass fibre does not experience full test tension. This can be, however, compensated by using high test tension, but it might damage the coating material. The automatic recovery from random fibre breaks essentially belongs to the properties of an on-line proof tester. This means that if fibre breaks in the proof test region, it must be possible to thread the free end of the fibre to the take-up reel for spooling at the full line speed without down ramping the draw process. In the test we threaded the fibre inside a guiding tube with different line speeds and tried to optimize the break recovery system. The results show that static electricity is the main problem when conveying fibre in the tube. In addition, the friction between fibre and the tube has to be as low as possible, and the so-called "rubber band effect" in fibre breaks makes treading challenging.

References

[1] W. Griffioen, "Optical Fibre Mechanical Reliability," Leischendam, Nederland, p. 209 (1994).
[2] J. Mrotek, M. Matthewson and C. Kurkjian, "Diffusion of Moisture Through Optical Fibre Coatings," in: Journal Of Lightwave Technology, 19(7), 988-993 (July, 2001).
[3] TIA/EIA-455-31C, Proof Testing Optical Fibres by Tension, Telecommunications Industry Association (1999).
[4] IEC 60793-1-30, Measurement methods and test procedures - Fibre proof test, International Electrotechnical Commission (2001).
[5] B. Overton and G. Orcel, "The Effects of Coating Characteristics on the Determination of The Dynamic Fatigue Parameters for Optical Fibres," PITVII, 341-350 (1995).
[6] S. Apone, L. Chiaro and G. Grego, "Viscoelastic properties of optical fibres coating", PITVII, 331-339 (1995).
[7] F. Kapron, "The influence of prooftest dwelltime on fibre reliability," Proceedings of 48th IWCS, 55-60 (1999).
[8] J. L. Armstrong, M. J. Matthewson and C. R. Kurkjian, "Humidity Dependence of the Fatigue of High-Strength Fused Silica Optical Fibres," in: J. Am. Ceram. Soc. 83(12) 3100-108.
[9] M. Lipponen, "Development of an on-line proof tester of optical fibre," The University of Oulu, Faculty of Technology, 2001. (Department of Mechanical Engineering, Master's Thesis)

Nextrom (USA) Inc.
1650 Satellite Boulevard
Duluth, GA 30097 - USA
Fax: Int’l +1 678 957 3512
E-mail: [email protected]