Ten ways to improve test productivity and reduce testing costs.

With the new millennium, much progress has been made in the rubber industry to improve rubber test productivity, implement test rationalization and significantly reduce testing costs. These efforts have only been accelerated with the introduction of Six Sigma and Kaizen methodologies into the work place.

This article reviews ten ways that rubber test productivity can be improved by simply changing the way that rubber testing is done. By making any one or combination of these suggested changes, significant improvements in factory or R & D productivity can be achieved.

1. Replacing Mooney viscosity testers with a rubber process analyzer for productivity gains

The Mooney viscometer was invented by Dr. Melvin Mooney in the 1930s and has been used continuously, even today. This method has been standardized under ASTM D1646 and ISO 289 (ref. 1). Mooney viscosity is actually reported in arbitrary "Mooney" units; however, these Mooney units have become part of the language of rubber.

For many years, Mooney viscosity was the single, most commonly used, test for characterizing raw rubber and mixed stocks. It has been shown that Mooney viscosity crudely relates to the average molecular weight of a raw rubber. However, it is a very poor indicator of other quality changes such as molecular weight distribution (mwd) or long chain branching (lcb).

The rubber industry has been searching for a better method for testing rubber. The 70-year-old Mooney viscosity test is still used extensively, but is relatively insensitive to more subtle differences among different production lots of raw elastomers, and does not lend itself to automation. One problem is that the relatively large sample size (approximately 25 g) and the use of an unheated rotor cause this instrument to have a long temperature recovery time compared to alternate technologies used today (ref. 2). Also, at the end of the Mooney test, the rubber has to be physically pried off the rotor, and then two new pieces of pre-cut rubber must be placed in the Mooney for the next test. Because this changing time can vary from one operator to another (and other reasons), the test repeatability and reproducibility is poor (ref. 3) compared with other new technology test methods. In addition, the long temperature recovery time of the Mooney requires this instrument's test time to be a minimum of five minutes (ML 1+4). Also, if a Mooney stress relaxation test has been programmed to be performed immediately after the final Mooney viscosity measurement, then the total Mooney test time is a minimum of seven minutes. This seven minutes of test time, plus the rather lengthy sample preparation and change over time, results in the Mooney viscosity test being rather lengthy (ref. 4).

With the introduction of the RPA 2000 rubber process analyzer in 1992, it was found that the RPA complex dynamic viscosity ([[eta].sup.*]), complex modulus ([G.sup.*]) or complex torque ([S.sup.*]) at 7% strain, 0.1 Hz and 100[degrees]C correlated well to the traditional final Mooney viscosity values (ML 1+4) at 100[degrees]C for a wide variety of different raw rubbers (ref. 5). This correlation is shown in figure 1. Also, figure 2 shows an excellent correlation that was achieved between the RPA and different lots of the same grade of SBR. This RPA test condition was established as the first test condition performed in a three-point frequency sweep (for processability) that was accepted by the rubber industry in 1997 as ASTM Standard D6204 (ref. 6). It was also discovered that the RPA uncured tan [delta] correlated very well to the Mooney stress relaxation test as well (ref. 7). This is illustrated in figure 3. These two test conditions can be performed on the RPA in only three minutes (ref. 8). Also, the rubber sample can easily be placed into the RPA and removed quickly from the RPA at the end of the test. Therefore, the so-called change over time is much less with the RPA compared to the Mooney viscometer test.

The time required to run rubber on an RPA is normally only one-third the time required to get the same information from the Mooney viscometer. Therefore, the RPA 2000 is approximately three times as productive as the Mooney viscometer. In addition, the information from the RPA is of better quality and is more sensitive to real differences in quality among lots of raw elastomer (ref. 9). Figure 4 illustrates this point for different lots of SBR.

2. Automating the RPA for raw rubber testing

Obviously, having to pry the tested sample of rubber off the Mooney rotor makes it very difficult to automate this instrument. For many years, the rubber industry was looking for a rotorless Mooney viscometer that would be automatable. With the development of the RPA (a rotorless shear rheometer), automation is possible. As discussed earlier, the RPA can be used to substitute for the Mooney viscometer, providing three times the test productivity over the Mooney viscometer. However, an even greater gain in productivity can be achieved by automating the RPA.

Automating the RPA will provide a tray where 10 samples can be placed for automatic testing, or an assembly of 10 trays is provided for the automatic testing of 100 samples. Up to 100 samples can be tested in a single time period without anyone being present. The Auto-RPA uses a patented system with film and a pneumatic robotic arm to automatically feed samples into the instrument in a very effective way. This means that an operator does not have to constantly work with this instrument and is free to do other work assignments anywhere in the plant. Once set up, the Auto-RPA can do all of the testing. Allowing the laboratory operator to concentrate his or her time on other tasks provides for a significant test productivity gain.

3. Use MDR 2000 in place of ODR for productivity gain

Just as the rotorless design of the RPA provides very significant productivity advantages over the Mooney viscometer, so too does the special rotorless design of the MDR 2000 moving die rheometer provide very significant productivity advantages over the older technology ODR (oscillating disk rheometer). The original ODR technology was developed and introduced by Monsanto Instruments (now Alpha Technologies) in 1962 (ref. 10). The ODR was a very important invention because, for the first time, it allowed the rubber technologist to quickly measure processability, scorch, cure rate and ultimate state of cure in one test. The ODR was standardized as a rubber test method under ASTM D2084 and gained wide acceptance throughout the world (ref. 11).

Even though the ODR had a great effect on the rubber industry, there were some technical problems with its design. One of these problems was the fact that the unheated biconical rotor acts as a "heat sink," causing lengthy temperature recovery times (the time required for the sample temperature to reach the set temperature of the instrument). The MDR 2000 was introduced in 1987 (ref. 12). Not only is the MDR design much more sensitive to real quality differences and able to measure dynamic properties (which the ODR can not do) (ref. 13), but the MDR can provide cure information in half the time because of its much faster temperature recovery and the elimination of the rotor (ref. 14).

Figure 5 shows that typically the temperature recovery time for an ODR will be about four to five minutes compared to the MDR 2000 with a temperature recovery time of only 20 to 30 seconds. Also, figure 6 shows that the MDR 2000 reached tc90 (time to 90% state-of-cure) in only half the time compared to the ODR (R100) design. Therefore, the MDR can provide cure information in half the running time for the ODR. Also, with the rotorless design of the MDR, a sample can easily be removed from its dies after the test is finished. On the other hand, the sample has to be pried off the ODR rotor, which takes more time. In addition, film could not be used effectively with the ODR design because of the rotor. With the rotorless design of the MDR, the sample can be placed in the MDR as a "film sandwich," and then easily removed after the test with tweezers. The use of film prevents die fouling and minimizes cleaning time between tests, which again provides for a productivity gain.

Therefore, with the temperature-recovery advantage, the sample-change-over time advantage and the cleaning time advantage of the MDR, this moving die rheometer technology can at least double the test productivity compared to the older technology ODRs.

4. Use MDR technology for higher temperature cures

Because the torque transducer is mounted at the upper die in the MDR, a much cleaner torque signal is measured as the force passes through the rubber sample from the sinusoidal oscillation movement of the lower die. (By contrast, the ODR design tries to measure the torque through the mechanical friction of the rotor bearings.) Sometimes the MDR displays up to ten times the test sensitivity as compared to the ODR for measuring the same compound change (ref. 15). Also, the MDR can apply cure temperatures as high as 230[degrees]C and still provide useful information about the compound. By contrast, the ODR cannot be used above 200[degrees]C. Depending on the energy of activation ([E.sub.A]) of a given compound, the cure reaction increases significantly with a rise in cure temperature, and the corresponding cure times become shorter (ref. 16).

Testing with a higher cure temperature using an MDR can double the test productivity by providing the curemeter data in half the time required for a lower temperature test (see figure 6). Also, as shown in figure 7, the drop off in statistical test sensitivity (signal-to-noise ratio), as measured in accordance with ASTM D6600, generally does not drop off below the R 100 ODR value at a 30[degrees]C lower test temperature (ref. 17). In addition, there are certain curemeter parameters which maintain a relatively high level of test sensitivity, compared to other curemeter parameters, with a rise in cure test temperature, as shown in figure 8 (ref. 18).

[FIGURE 6 OMITTED]

From the data presented, substituting the MDR technology at a higher cure temperature in place of the ODR technology can triple or quadruple the testing productivity compared to the R100. By raising the cure test temperature by as much as 30[degrees]C, one can gain a significant productivity advantage while still maintaining the statistical test sensitivity originally achieved with the ODR. On the other hand, if the MDR is run at the same cure temperature as the R100 ODR, a smaller test productivity gain is achieved, but the MDR will display much better test sensitivity compared to the R100.

5. Automate the laboratory MDR

One very effective way to improve the productivity of an MDR in the laboratory is to automate it. Auto MDRs can be purchased OEM or MDRs already in service can be retrofitted (ref. 19).

The auto MDR uses a patented film system similar to that mentioned earlier for the RPA. A very special pneumatic robotic arm picks up pre-cut rubber specimens and places them between two layers of film in the MDR test cavity. Each tray can hold 10 specimens, up to 10 trays can be stacked, and the trays are robotically moved automatically in order to feed the instrument without an operator present. Film is used to carry the cured specimen away from the dies, without sticking, after completion of the cure test.

No operator needs to be present during the testing of many samples. Also, time between tests is reduced for better testing efficiency. Because no operator is needed during testing, variability due to the operator is eliminated. Lastly, the constant use of protective film greatly reduces any time needed for cleanup.

6. Replace the lab MDR with a production MDR

The production MDR instrument is in a hardened cabinet, possessing significant design improvements that allow it to work effectively under very harsh rubber factory conditions with proven reliability. This instrument is designed to provide a very easy means for the factory worker to load this instrument. All that is necessary is for a rubber sample to be placed on the staging platform. In fact, more than one sample can be "staged" or set up for testing. Several samples can be set up for sequential testing. The factory operator only needs to place the sample(s) on the staging platform and simply push the start button. The rest is automatic. After the test, the instrument is set up to give either a green light or a red light (go or no-go). This is in the SPC tradition of allowing the factory worker to take responsibility for the rubber process and to be responsible for needed corrective actions that might be necessary in order to keep a process in a state of statistical control (ref. 20).

Significant test productivity gains are achieved by using the production MDR in place of the laboratory MDR. By transferring the QA testing responsibility from the laboratory to the factory floor, no time is wasted in transporting samples from the factory floor to the laboratory. This significantly reduces (or eliminates) the plant laboratory operating costs. Also, any corrective action, if needed, is much more immediate and may prevent additional defective batches from being produced. This can also translate into significantly lower scrap and rework costs, which also translates into a production cost savings.

7. Replace both the Mooney viscometer and the curemeter functions with a single RPA or PPA test

Commonly, the 70-year-old Mooney viscometer test and the 40-year-old oscillating disk rheometer test are still being used together by many rubber fabricators. However, with the development of the new RPA technologies and the advent of the new ASTM D6204 Methods A, B and C, now a single, shorter and more sensitive test based on the RPA/PPA can substitute for the Mooney and ODR tests.

As already discussed earlier, the Mooney viscosity function can be replaced by a three-minute ASTM D6204 Part A or B procedure using RPA technology (ref. 21). An example of the output from this short test is shown in figure 9. In addition. without having to change rubber samples, the RPA can be programmed to also perform the new proposed ASTM D6204 Part C, a variable temperature cure which more than fulfills the ODR function and possesses superior test sensitivity to scorch (50% better) (ref. 22).

[FIGURE 9 OMITTED]

Just as replacing the lab MDR with the production MDR, a greater test productivity gain can also be achieved by replacing the laboratory Mooney viscometer and ODR with a PPA (production process analyzer) on the factory floor. The PPA utilizes the RPA technology in a very select way. The PPA is customized to carry out a specialized mission (such as preprogrammed test variants based on ASTM D6204 Parts A, B and C). This instrument possesses the same sample staging area and red light/green light functions discussed earlier for the production MDR. By substituting the PPA instead of the RPA for the Mooney and ODR curemeter functions, an even greater test productivity gain can be achieved by transferring the testing responsibility from the laboratory to the factory floor (ref. 23).

By combining the Mooney test function and the ODR curemeter function into one single, short RPA or PPA test, much less test time is required in the laboratory, much less time is required from laboratory personnel, there is a significantly lower maintenance requirement, and a test with far better sensitivity is created.

8. Replace curemeter and lengthy testing

Many manufacturers of rubber products used in dynamic applications have to utilize expensive dynamic testing resources to assure the proper performance of these products in their intended application. Figure 10 shows such testing being dynamically performed with an MTS with compression sinusoidal strain being applied to the cured rubber sample. The problem with this type of dynamic testing is that the rubber compound must be molded and cured into a test specimen. Then this test specimen must be skillfully positioned into the apparatus and carefully tested by a highly trained operator. Overall, it can take 24 to 48 hours to get dynamic property information, which prevents it from being a good on-line test for effective quality assurance.

[FIGURE 10 OMITTED]

A much faster and less expensive way to measure these dynamic properties is to cure the rubber compound in-situ in the RPA and automatically cool the temperature down to a predetermined lower temperature (such as 60 or 100[degrees]C) for an effective strain sweep to quickly measure after-cure dynamic properties. There is now an ASTM Standard (D6601) which discusses the most appropriate manner of testing with the RPA after-cure dynamic properties (ref. 24). Also, figure 11 shows the excellent correlation that is achieved between the RPA and MTS. By using the RPA instead of the MTS in factory testing, one can achieve rapid test results that can be used for on-line process control (ref. 25). Just as with the previous idea, once again still greater savings can be achieved by using a special customized PPA on the production floor. The RPA with in-situ curing will also replace the curemeter function as well.

[FIGURE 11 OMITTED]

By using RPA/PPA in-situ curing and after-cure dynamic property testing as a single test to replace both the traditional MTS testing and curemeter functions, better quality control can be achieved and more effective corrective action can be applied to significantly reduce scrap and rework. Also, the RPA/PPA testing costs are significantly lower than MTS testing costs. Typically, $50,000 to $80,000 per year in direct savings has been reported.

9. Replace traditional tests with shorter RPA tests

Many traditional rubber tests have been replaced by different RPA tests. For example, figure 12 shows the excellent correlation between the RPA and the so-called "300% modulus" (ref. 26). Figure 13 gives the correlation between the RPA and durometer A hardness (ref. 27). Figure 14 expresses the correlation found between the RPA and the Goodrich Flexometer heat buildup (ref. 28). Figure 15 shows the inverse correlation between the RPA tan 5 and % rebound (ref. 29). Figure 16 gives the excellent correlation between the capillary rheometer and the RPA (ref. 30). In fact, table 1 lists the 26 traditional rubber tests that the RPA can correlate and replace (refs. 31-35).

[FIGURES 12-16 OMITTED]

Therefore, through test rationalization and elimination of the older, more time consuming, inefficient and less sensitive test methods, a large test productivity gain can be achieved by using the newest test methods based on RPA technology.

10. Adopt new software or upgrade to better software

Everyone can identify with the proposition that computers and software developments have made us all personally more productive on the job, as well as in our own personal endeavors. Much of the productivity gains experienced in the United States in the late 1990s were because of better availability and improvements in computers and software. Therefore, it is not a surprise to anyone that using computers and upgrading software to run rubber tests will usually significantly improve test productivity.

Software upgrades can improve your personal productivity many times by significantly shortening the time you may spend on analyzing test data. New, modern software can also provide you with much taster and more flexible quality report writing capability. Sometimes, quality reports can be automatic. Also, with new on-line capabilities, test information can instantly be shared with select individuals all over the world. Quality decisions can even be made from home without having to come into the plant off-hours. Automatic batch validation decisions (pass/fail) can be obtained by using the integrated specification limit controls through modern software. The latest state-of-the-art software, especially designed for today's needs in the rubber industry, is available (ref. 36).

Test productivity can be increased through the use of state-of-the-art software to reduce the time required for data analysis, implementing quality decisions, creating quality reports, and selectively distributing information.

Table 1-26 traditional tests that the RPA can
correlate with by using different conditions of
frequency, strain and temperature

RPA data substitutes for or correlates with the following:

Mooney viscometer           MDR
Mooney stress relaxation    ODR
Mooney scorch               Cure simulators
DSR                         Tensile modulus
Defo                        Shear modulus
Wallace plastimeter         Compression modulus
Williams plastimeter        Durometer
Solution viscosity          Flexometer heat buildup
GPC (QC)                    Flexometer permanent set
MPT viscosity               Zwick rebound
MPT die swell               MTS
Garvey die                  Rheometrics
Garlock flow                DMA

References

(1.) J. Dick, "ASTM in the globalization of rubber standards and specifications," presented at a meeting of the Rubber Div., ACS, Fall, 1994, paper no. 98.

(2.) ASTM D1646, ASTM Book of Standards, Vol. 09.01.

(3.) J. Dick, Chapter 2, "General Test Methods," Basic Rubber Testing, Selecting Methods for a Rubber Test Program, pp. 20-30.

(4.) Ibid ref. 3.

(5.) H. Pawlowski and J. Dick, "Viscoelastic characterization of rubber with a new dynamic mechanical tester," Rubber World, June 1992.

(6.) J. Dick and T. Liotta, "New useful ASTM test methods and standards now available internationally for the rubber industry," Rubber World, January 2004.

(7.) Ibid ref. 5.

(8.) Ibid ref. 6.

(9.) W. Cousins and J. Dick, "Effective processability measurements of aco'lonitrile butadiene rubber using rubber process analyzer tests and Mooney stress relaxation," Rubber World, January 1998.

(10.) G.E. Decker, R.W. Wise and D. Guerry, "An oscillating disk rheometer for measuring dynamic properties during vulcanization," Rubber Chemistry and Technology, vol. 36, 451 (1963).

(11.) Ibid ref. 3, pp. 30-38.

(12.) J.A. Sezna, P.J. DiMauro and H.A. Pawlowski, "Improved cure testing," paper presented at a meeting of the Rubber Div. ACS, Fall, 1987.

(13.) J.A. Sezna and J.S. Dick, "The use of rheometers for process control," Rubber and Plastics News, Apr. 13 and 27, 1992.

(14.) J. Dick, "The optimal measurement and use of dynamic properties from the moving die rheometer for rubber compound analysis," Rubber World, January 1994, (and in Spanish in Revista Caucho, February, 1996).

(15.) Ibid ref. 14.

(16.) J. Dick and H. Pawlowski, "Application for the curemeter maximum cure rate in rubber compound development, process control and cure kinetic studies," Polymer Testing 15, (1996) 207-243.

(17.) Ibid ref. 14.

(18.) Ibid ref. 16.

(19.) J. Sezna, P. DiMauro and C. Shelley, "Automation of rotorless curemeters," paper no. 69, Rubber Division, ACS, Spring, 1992.

(20.) J. Dick, "Cost savings" through standard target values," Rubber World, June 1986, and Standardization News, September 1985.

(21.) Ibid ref. 6.

(22.) J. Dick, C.A. Sumpter and Brian Ward, "New effective methods for measuring processing and dynamic property performance of silicone compounds," KGK (Kautschuk Gummi Kunststoffe), September 1999 (52, pp. 600-607).

(23.) M. Ferraco and J. Dick, presentation given at the RMA, 2001.

(24.) J. Dick, chapter 2, General Test Methods, Basic Rubber Testing, ASTM, 2003, p. 49.

(25.) H. Pawlowski and J. Dick, "A new mechanical tester designed for testing rubber," paper no. 70, Rubber Division, ACS, Spring 1992.

(26.) Ibid ref. 5.

(27.) Ibid ref. 5.

(28.) J. Dick, H. Pawlowski and J. Moore, "Viscous heating and reinforcement effects of different fillers using the rubber process analyzer," Rubber World, January 2000.

(29.) J. Dick and H. Pawlowski, "Applications for the rubber process analyzer, part 2," Rubber and Plastics News, May 10, 1993.

(30.) J. Dick, "Comparison of shear thinning behavior using capillary and rotorless shear rheometry," Rubber World, January 2002, p. 23.

(31.) Ibid ref. 29.

(32.) Ibid ref. 30.

(33.) Ibid ref. 25.

(34.) Ibid ref. 5.

(35.) J. Dick, A. Vare and C. Harmon, "Quality assurance of natural rubber using the rubber process analyzer," Polymer Testing, 18, pp. 327-362 (1999).

(36.) "RPA with Pathfinder software, getting the most out of your production process," Alpha Technologies, 2004.