16 M e a s u r i n g 16.1 S e n s o r s in in I n j e c t i o n M o l d s Molds Crucial to the quality of injection moldings is the state of the melt after plastication and how the mold-filling process went, which is characterized by the change in pressure and temperature It is therefore advisable to incorporate appropriate sensors into the molds for monitoring and perhaps control purposes On the other hand, installing them incurs additional costs, could even leave marks on the moldings to be made, and could possibly make mold changing difficult Moreover, the cooling channels and the ejector units take up considerable space, with the result that there is limited choice as to the best place to install them For these reasons, it is not standard practice to measure melt temperature in production molds However, pressure measurements in molds are being used more and more to monitor production (for documentation and scrap monitoring) and are a prerequisite for pressure-dependent switching and for controlling the pressure in the mold [16.1] 16.2 Temperature Measurement 16.2.1 M e a s u r i n g M e l t T e m p e r a t u r e s in M o l d s U s i n g IR Sensors The thermocouples and resistance thermometers used to measure and control mold wall temperatures are unsuitable for measuring melt temperatures (these are discussed in Chapter 18: Temperature Controllers for Injection and Compression Molds) Melt temperatures are measured by methods that detect the inherent radiation emitted by the melt in a defined wavelength range To this end, a bundle of optical fibers is installed inside the mold, flush with the wall The inherent radiation emitted by the melt is fed to a radiation-sensitive detector The electrical output signal from the detector then correlates with the melt temperature The radiation emitted to the environment by a body as a result of its temperature depends on its emission power Ideal blackbody emitters emit all their heat, i.e., they have an emission power of Real bodies have a lower emission power, the amount of radiation emitted being reduced by the amounts lost through reflection and transmission Polymer melts inside a closed mold behave almost like ideal blackbodies since the reflected and transmitted radiation emitted from the cavity surfaces impinge on the IR sensor This is why plastics that have an emission coefficient of 0.9 in air have emission coefficients of 0.99 in molds [16.2-16.4] Melt temperatures inside molds can therefore be measured with high precision (±1% of the display value) and speed (response time < 10 milliseconds) and with high reproducibility using optical fibers connected to IR measuring devices The measurement ranges depend on the wavelength of the optical fiber, lying between 70 and 260 C, and between 177 and 427 0C [16.2] 16.3 Pressure Measurement 16.3.1 Purpose of Pressure M e a s u r e m e n t Figure 16.1 illustrates the change in pressure within a cavity and the properties of the finished part that are affected by it The shape of the filling phase has a major influence on the texture of molding, the degree of orientation, the level of crystallinity in the layers close to the surface, and thermal and mechanical stress on the melt Complete filling of the contours of a mold cavity, along with the occurrence of flash and possible resultant mold damage depend on how the pressure changes as it rises in the compression phase During the holding-pressure phase, the weight and shrinkage of their molding is determined Just as in the filling phase, the degree of crystallization and orientation of the macromolecules in the inner regions of the molds can be affected here 16.3.2 S e n s o r s for M e a s u r i n g M e l t Pressures in M o l d s Filling Completeness, shaping of contours, flash formation, mold damage Composition of surface layer Appearance, orientation, crystallinity Compression longitudinal flow path Vod i s sink marks, shrinkage, distortion, degree of crystallization, orientation in the interior, demolding problems Stress on the melt (thermal, mechanical) Temperature distribution A Filing phase Figure 16.1 B C Compresso i n phase Holding pressure phase TIME How change in pressure affects quality of finished parts Pressure in the mold The informativeness of the change in pressure inside the mold depends primarily on the position of the installed detector The pressure sensor measures the pressure during the injection phase, as soon as the flow front passes over the point of installation and throughout the holding-pressure phase until the molding shrinks so much at the point of installation that it loses contact with the cavity For this reason, pressure sensors should Compression at the site of D measurement be installed close to the gate as the change in pressure measured there best describes molding formation for the longest possible period A position close to the gate is also recommended for any intended measures to control the change of pressure inside the mold To measure the degree of filling, by contrast, installation away from the gate is best Therefore, it would be better to position the pressure sensor away from the gate in the case of molds that have very long flow paths or that tend to overpack away from the gate A range of different designs based on piezoelectric and strain gage detectors are available for direct and indirect measurement of internal mold pressures [16.2, 16.5] 16.3.2.1 Direct Pressure Measurement Adjust In the direct method of measuring internal mold pressure (Figure 16.2), the hole for the detector is bored right into the cavity such that the head of the detector is in contact with the melt The head is shaped as a piston or membrane It transmits the pressure to the measuring element behind it The head of pressure detectors fitted with a piston can be machined and shaped within certain limits to the contour and texture of the cavity Piezoelectric pressure detectors return a charge signal proportional to the mechanical load that is transformed in a downstream amplifier into the corresponding voltage Even in the most unfavorable measurement points, system drift by the analyte signal is better than 0.5%/min Moisture and contamination of the feed lines can exacerbate the drift and it is absolutely vital to keep them clean or clean them as necessary [16.5] Figure 16.2 Direct measurement of cavity pressure with the aid of a piezoelectric pressure sensor [16.2] Piezoelectric measuring systems are only suitable for making dynamic and quasistatic measurements and they must therefore be brought into a defined state before every measurement This can be effected automatically by a signal from the machine control (e.g., "Move forward") prior to the start of injection Piezoelectric pressure detectors feature high accuracy and linearity (< ± 1% of measuring range) and a low temperature drift (< ± 0.01%/°C); common measuring ranges extend up to 200 MPa at 200 to 350 0C (depending on model) [16.5] Strain gage detectors (working on the principle of a change in resistance) cannot attain the same high threshold frequencies as piezoelectric detectors but their dynamic behavior is totally adequate for measuring very rapid changes in internal pressure during the mold filling process On account of their measuring principle and construction, they are suitable for both static and dynamic measurements Resistance changes brought about by temperature changes in the strain gage elements are compensated because the entire measuring chain, consisting of pressure detector and amplifier, is automatically calibrated before each measurement Again, this is effected by means of a signal from the machine control prior to the start of injection (see above) The overall error arising from measurements with strain gage detectors is < ± 1% of the measuring range, and temperature drift varies according to model from ± 0.02 to 0.04%/°C; common measuring ranges extend up to 150 MPa at 200 0C [16.2] 16.3.2.2 Indirect Pressure Measurement Additional marks in the molded part can be avoided by using sensors for indirect measurement of the melt pressure placed under ejector pins that are already present anyway The analyte parameter is the force acting on the ejector; the diameter of the ejector is used for calculating the cavity pressure (Figure 16.3) During mold construction and installation, a good fit of the ejector pins is necessary to ensure that the measurement is not falsified by frictional forces There are two types of design; one is bolt-shaped with piezoelectric or strain gage elements If the pressure is monitored only during mold proving and not continuously, the sensors can be removed and replaced by "blank sensors" [16.2, 16.5] The data of the sensor for indirect pressure measurement correspond to the data of the sensor for direct pressure measurement Figure 16.3 Sensor with piezoelectric or DMS element [16.2] The data quoted for measuring accuracy, temperature dependency and response time, just like those concerning measuring ranges and service temperatures are guide values; the manufacturer's instructions must be observed each time 16.4 U s e of S e n s o r - T r a n s d u c e r Probes The sensors described can be used for various processes associated with production These can be divided into process-optimizing and quality-monitoring mechanisms 16.5 Process Optimization Optimum machine setting parameters are crucial to the quality of the molded parts Optimum process control entails minimizing cycle times and increasing the service lives of machines and molds It is often not enough to just consider process parameters on the machine side as a means of optimizing the injection molding process This is because this approach fails to take account of the filling behavior of a mold, for instance, viscosity fluctuations of the raw material, and changes in ambient conditions such as shop temperature and humidity The more sensors which are fitted to the mold, the more information that can be gained from the production process The most useful information about the process of molded-part formation can be obtained from measuring the mold cavity pressure Optimizing the shape of this process curve can greatly influence molded-part properties such as dimensions, weight, and internal properties, as well as cycle times and mold wear The measurements also help to speed set-up when molds are used on different machines Figure 16.4 illustrates various cavity-pressure curves Of most interest is the linear rise in the mold cavity pressure, synonymous with virtually constant flow front speed [16.11] during injection (e), the switch-over phase from the speed-controlled injection phase to the pressure-controlled holding-pressure phase (n) and the duration and value of the holding pressure Figure 16.4 shows pressure curves that are obtained when the switch-over point in the mold is altered Switching too late from the injection to the holding-pressure phase (1) causes the cavity pressure to peak and may lead to overpacking of the cavity, demolding difficulties and damage to the mold, while switching too early may cause the cavity to be filled under holding pressure (2) Incompletely filled molded parts and a high level of orientations are the result The optimum cavity-pressure curve has neither peaks nor pressure breaks (3) With the aid of the cavity pressure sensor, the optimum point can be determined rapidly Similarly, the gate-sealing point, the point when the gate is frozen and no more material can be transported into the mold, can also be determined and the holding-pressure time adapted The residual cooling time is the time elapsed until the molded part detaches from the mold and is equivalent to the time when the mold cavity pressure has reached the zero limit The whole process can be recorded by placing the mold-cavity pressure sensor close to the gate The farther the sensor is located away from the optimum position, the less information is obtained about the process Aside from manual corrections to the molded-part formation processes performed with the aid of cavity-pressure curves, modern injection-molding control units offer the possibility of using the cavity pressure to switch over from the injection to holding phase Once a threshold value is exceeded, a corresponding signal is transmitted This switching process makes for greater reproducibility Above all, it allows for fluctuations Switching too late Pressure peaks - Demolding problems - Flash - Orientation Cavity pressure pwi Zeitt (2) Cavity pressure pwi Cavity pressure pwj (1) Switching too early Filing under holding pressure - Sink marks - Surface flaws - Orientation Timet (3) Idea! switching (e) Timet Figure 16.4 Change in pressure at different switching points in metering The threshold value must be determined iteratively in trials or with the aid of simulations Automatic determinations can be made by measuring the change in pressure on-line [16.5, 16.12] The significant rise in the pressure change serves as the criterion for the determination and the switch-over point from the injection phase to the holding phase is made during the cycle This process allows the system to respond automatically to fluctuations in materials or the process, without the need for external intervention After the mold-cavity pressure, which is the most relevant process parameter during injection molding, temperature measurements may be used to obtain further important information for optimizing and stabilizing injection-molding production Infrared sensors for measuring the temperature in front of the screw provide information about the actual melt temperatures and uncover possible processing errors due to specification limits that have not been observed The temperature that has been preset for cooling systems or measured and depicted describes the temperature of the cooling medium as it returns from the mold to the machine The actual mold-wall temperature, however, can only be measured with the aid of thermocouples mounted just beneath the mold wall in the mold Recording these temperatures allows not only non-observed specification limits to be identified and compensated but also dynamic process changes, such as improperly set controllers and ambient influences 16.6 Monitoring Quality High quality demands imposed on injection-molded parts require intensive testing by the molded-parts manufacturer Statistical process control (SPC) is standard in companies and uses random samples to make inferences about the process level or quality level and to identify trends in production The constantly rising demands on injection-molded products and the random controls of SPC frequently make continuous quality control monitoring (CQCTM) unavoidable Sensors in the mold and machine are used to effect on-line monitoring The goal is to calculate quality properties direct from the measured process parameters, such as cavity pressure and temperature This provides the Process knowledge Design of experiments Conducting of experiments Changes in process curves Molded part quality Charact parameters Statistical analysis Process optimization Poor Good On-line process monitoring Figure 16.5 Sequence of steps for setting up process models for monitoring quality possibility of performing a 100% quality control without the need to examine each and every molded part The key to the quality calculation is a model that describes the dependencies of molded-part properties and process parameters The approach for determining such a model is illustrated in Figure 16.5 Starting from existing knowledge of the process, such as that of a machine setter, a statistical experimental plan is drawn up in which alterations to machine settings are set out during trials to be conducted Statistical experimental design makes it possible to design trials efficiently and to gain a great deal of information from technical processes with a low number of experiments In injection molding, process parameters are recorded during the trial phase on the machine and the molded parts produced are removed and labeled so that clear classification of trial parts and measured process parameters is ensured during the modeling phase When the trials are over, the quality characteristics of the produced molded parts are determined Modeling of the interdependencies of the quality characteristics and process parameters is performed off-line Mostly, statistical methods are used as models The choice of regression to use depends on the type of quality characteristic to be evaluated Continuous characteristics, such as weights and dimensions, are evaluated with multiple regression calculations Logistic regression is used for describing the dependencies of attributive characteristics This group of quality characteristics includes those which can only be subdivided into classes, such as "Characteristic present" and "Characteristic absent" Burns and sink marks numbers are among these Increasingly, statistical methods are being replaced by algorithms from research areas of artificial intelligence, such as neural networks These have the advantage of being easier to apply since the relationships between process parameters and quality characteristics can be learned independently [16.13, 16.14] With the aid of the models created, the quality of molded parts can be calculated online during injection molding from the process parameters measured during the production cycle Thus, as soon as the cycle is over, information is available on the product just made This can be used to aim for 100% quality documentation and establish a "reject route" for molded parts that fail to meet specifications References [16.1] [16.2] [16.3] [16.4] [16.5] [16.6] [16.7] [16.8] [16.9] Recker, H.; Spix, L.: MeB-, Steuer- und Regeleinrichtungen In: Johannaber, F (ed.): Kunststoffmaschinenfiihrer, 3rd Ed., Carl Hanser Verlag, Munich, Vienna, 1992, pp 1123-1180 Company brochure, Dynisco Gerate GmbH, Heilbronn Wittemeier, R.: Temperaturmessung im Schneckenvorraum einer SpritzgieBplastifiziereinheit mit Infrarotsensoren, Study report, Univ GH, Paderborn, 1990 Bluhm, R.: Methoden zur Bestimmung der Massetemperatur - Infrarotsensoren und miniaturisierte Thermoelemente, Paper presented at the 17th Tooling Conference, IKV, Aachen, 1994 Company brochure, Kistler Instrumente AG, Winterthur, Schweiz Sarholz, R.; Beese, U.; Hengesbach, H A.: ProzeBfiihrung beim SpritzgieBen, IKV, 1977 Stitz, S.: Konstante Formteilqualitat beim SpritzgieBen durch Regeln des WerkzeugDruckverlaufs, Kunststoffe, 63 (1973), 11, pp 777-783 Rorick, W.: Zur Praxis der ProzeBregelung im Thermoplast-SpritzguBbetrieb, Dissertation, RWTH, Aachen, 1979 Sarholz, R.: Rechnerische Abschatzung des SpritzgieBprozesses als Hilfsmittel zur Maschineneinstellung, Dissertation, RWTH, Aachen, 1980 [16.10] Hellmeyer, H 0.: Ein Beitrag zur Automatisierung des SpritzgieBprozesses, Dissertation, RWTH, Aachen, 1977 [16.11] Lauterbach, M.: Ein Steuerungskonzept zur Flexibilisierung des ThermoplastspritzguBprozesses, Dissertation, RWTH, Aachen, 1989 [16.12] Bader, C : Die Bedeutung der WerkzeuginnendruckmeBtechnik beim SpritzgieBen Reprint at the conference on: Qualitatssicherung und -optimierung beim SpritzgieBen, IKV, Aachen, 1997 [16.13] Gierth, M M.: Methoden und Hilfsmittel zur prozeBnahen Qualitatssicherung beim SpritzgieBen von Thermoplasten, Dissertation, RWTH, Aachen, 1992 [16.14] Schnerr, O.; Vaculik, R.: Qualitatssicherung mit Neuronalen Netzen Paper presented at the 18th Tooling Conference, IKV, Aachen, 1996