Resin patterns obtained in vacuum-sealed sand moulds

4. Initial properties of the resin patterns

The initial properties of the resin patterns, which are formed in the negative mould - vacuum-sealed or plastic, are necessary
- for the explanation for some results of further wear tests (forming of surface layer and initial surface roughness),
- and for the including some of them (the surface roughness and strength properties) in the information basis of the software for the
  exploitation prognoses of the patterns, where they serve as a restriction in the choice of one or another compound for
  patternmaking.


4.1. Forming of surface layer

A foil from the walls of the vacuum-sealed mould changes the conditions for forming of surface layer of the hardened resin pattern. For the determination of these conditions, the wetting of the vacuum-sealed mould by the epoxy resin is investigated here. For that purpose, the macroscopic angle of wetting Q [14] is used. Because of some difficulties for its direct measurement on the foil of the vacuum-sealed mould, it was done on the resin plates-replicas (Fig.13). These replicas are made in a mould (Fig.14) which has sectors formed from different silica sand fractions including 0.25, 0.32 and 0.40 mm. The foil thickness is 0.10 mm and the vacuum in the mould is 0.4.105 Pa.

     
          Fig.13. Resin plate-replica                                        Fig.14. Mould for making of resin replicas                                            
The geometrical change of the resin drops is registered through the taking of a photograph at once after their dropping and continues during definite periods of time until 24 hours. The last stage of the research is the determination of Q directly on assembled set of photos. These operations are well known and here is enclosed only set for fraction 0.25 mm (Fig.15). All other sets are analogous.

The found results for Q are compared to the values related to the resin plate made by the conventional technology (Fig.16).

               
   Fig.15. Photos for fraction 0.25 mm              Fig.16. Influence of the mould over the macroscopic angle of wetting Q

As a result of regression analysis the following equations for Q = f (t) depending on the mould are worked out:
plastic mould,                                     y = 102.43 x - 0.133 ,
vacuum-sealed mould, fraction 0.40,     y =   83.86 x - 0.118 ,
vacuum-sealed mould, fraction 0.32,     y =   68.64 x - 0.105 ,
vacuum-sealed mould, fraction 0.25,     y =   61.81 x - 0.095 .

The results show that Q for plastic mould is greater than the values referred to the vacuum-sealed moulds. In this instance, the known dependence "Q - surface relief" is confirmed. When the surface roughness is bigger, therefore with bigger real areas from micro-unevenness, the wetting is hard by the increase of the adhesion interaction.

As it really is, the liquid compound wets and flows the foil well. That reduces the possibility for closing of the microscopic air volumes. The gas bubbles from the mixing of the components are pushed easier by the front of the flowed down compound and the gases do not stay in the contact boundary compound - mould. The surface compounds contain thixotropic fillers decreasing the flowing, but in that condition even, the bubbles are minimum and that predestines to obtain of dense surface layers of the resin patterns made in vacuum-sealed sand moulds.


4.2. Surface roughness

The compound copies precisely the relief of the vacuum-sealed mould and this circumstance forms one or another surface roughness of the pattern. The combined influence of the foil, sand and vacuum on the roughness of the resin patterns is shown in Fig.17.

Fig.17. The influence of the foil, sand and vacuum on the mould (pattern) roughness
1 - foil; 2 - sand; 3 - gaps between the sand grains.
By analogy with the previous point, a few resin plates-replicas are used. They are manufactured with the help of an experimental master-pattern which has a metal part with a mean value deviation of the profile Ra=1.08 µm. The replicas are obtained in vacuum-sealed moulds which contain sectors (as it is shown in Fig.14) formed from different silica sands according to the Bulgarian standard [15] - 1PK 010-70, 1PK 016-70 (AFS GFN = 56.2), 1PK 025-70 (AFS GFN = 50.8), 1PK 032-70 (AFS GFN = 47.8) and 1PK 040-70, and different silica sand fractions 0.10, 0.16, 0.25, 0.32 and 0.40 mm. The vacuum in the mould during the separate runs is maintained at a value of 0.4, 0.6 and 0.75.105 Pa. The foil thickness in all runs is 0.10 mm, and the vibration regime is constant - frequency 2790 min-1, amplitude 0.8-1 mm and duration 10-15 s. The temperature of the sand is one and the same,
20-25 ºC.

With the help of the same master-pattern a plate is manufactured in plastic mould too. In Fig.18 is shown a selected set of plates, including a replica from sand fraction 0.63 mm which indicates the strong influence of the used sand grains over the pattern roughness.


Fig.18. A set of resin plates-replicas
1-made in plastic mould;
2-6 - made in vacuum-sealed mould, at vacuum 0.75.105 Pa;
2 - fraction 0.16 mm; 3 - 0.25 mm; 4 - 0.32 mm; 5 - 0.40 mm; 6 - 0.63 mm.

The test findings are given in Fig.19. The roughness of the plate by the plastic mould is Rà=3,5 µm and it is marked in Fig.19 with horizontal blue lines. The roughness of the master-pattern is marked with horizontal red lines. All mean values are obtained from five measurements for each sample.


       a)                                                                              b)
    Fig.19. The influence of standard silica sands (a) and their fractions (b) on the pattern roughness

As a result of regression analysis the following equations for Rà = f (Vacuum) depending on the used sand are worked out:
1PK040,    y = 2.6408 x - 0.5026 ,             fraction 0.40,    y = 2.8432 x - 0.5810 ,
1PK032,    y = 2.1590 x - 0,2377 ,             fraction 0.32,    y = 1.8938 x - 0.4452 ,
1PK025,    y = 0.9473 x - 0.5335 ,             fraction 0.25,    y = 0.9348 x - 0.5638 ,
1PK016,    y = 0.9255 x - 0.3692 ,             fraction 0.16,    y = 0.9190 x - 0.4531 ,
1PK010,    y = 0.9268 x - 0.3370 ,             fraction 0.10,    y = 0.9174 x - 0.4336 .

In Fig.20 are presented the profilograph recordings for plates obtained in both kinds of moulds.


 a)                                                                                    b)
Fig.20. Profilograph recordings for vacuum-sealed mould - sand 1PK025-70, vacuum 0.75.105 Pa (a) and plastic mould (b)

The results show that the surface roughness of the resin patterns obtained in vacuum-sealed moulds grows with the increase of the vacuum and the grain size of the sand, and it is a matter of course. In most cases, the roughness of the plates (patterns) manufactured in vacuum-sealed moulds is less than the roughness of the plates (patterns) made by the conventional technology. At vacuum 0.4.105 Pa and in use of silica sands / fractions 0.10-0.25 mm, the roughness of the plates from vacuum-sealed moulds is less even than that of the master-pattern. But, the dependence of the resin patterns quality on the conditions of their manufacture is a weak side of the technology, especially in non-compliance with the technical instructions for work.


4.3. Accuracy of the patterns

The shrinkage of the filled compounds is small: 0.1-0.2% [7,16], therefore here is turned main attention to the dimensions of the vacuum-sealed mould, which are not change in whole time of the process of resin patternmaking. Since the formulations defined by known authors [17-21] are confirmed, and the regimes of mould manufacturing are not comment widely here.

The measurement of the dimensions is done with the help of master-pattern (Fig.21-a), which has some calibrated steel belts. After its covering with a foil, in appropriate places to the metal belts some small permanent magnets are put (Fig.21-b). The field of the magnets is weak, so they are not an obstacle for the recession of the foil from the master-pattern to the sand when the flask is connected to the vacuum system. The built in the sand magnets save the vacuum-sealed mould from a local deformation when the measurements are made and that insures their high accuracy. The vacuum is gauged with a sonde.

a)









b)





 Fig.21. Test set for the accuracy research
a) master-pattern; b) vacuum-sealed mould
1-steel belts; 2-permanent magnets;
3-silica sand; 4-sonde.

The experiments are carried out with a flask with a height of 60 mm. The master-pattern position towards the flask walls is constant. The quantity of the silica sand in the mould is one and the same. The homogeneity grade of the used sand 1PK025-70 is 70%, It is known [6,19,20], in this kind of sands, the fine-grained particles put in the empties of coarse-grained particles and that increases the bulk density of the sand and the contact area between the particles, which ensures the raising of the strength and the hardness of the vacuum-sealed mould. The sand grains are with a round form. The foil, the frequency and the amplitude of vibration are the same as in the roughness test. So, the change of the dimensions A, B and C (Fig.22) is defined for different regimes of mould making - value of vacuum, without vibration and vibration with different duration, and with use of an additional weight (plate) on the sand in vibration.



Fig.22. A change of the dimensions A, B and C depending on the regimes of mould making
1-without vibration; 2-vibration 10 s; 3-vibration 20 s; 4-vibration 20 s+10 kg weight;
4'-dimension A in 0.1% shrinkage of the pattern and vibration 20 s+10 kg weight.

According to expectation, the dimensions A and C increase, and B decreases in consequence of the shrinkage of mould, respectively the expansion of its working empty. In all cases, the vertical dimensions are changed less, regardless of their sizes. It is typical that the variation of vertical dimensions depends on the distance to the unworkable surface (side) of the mould. In this case, dimension C depends on the distance D, especially when does not vibrate. In C/D=1 (it is not shown in Fig.22) the increase of C is maximum, because the conditions permit to regroup of bigger quantity of sand and bigger shrinkage of that part of the mould. In that way the visible roundings R are appeared. They often coincide with moulding (pattern) roundings, but they are unfavorable in the core prints.

The possibility for mould making without vibration is acceptable in the manufacture of resin patterns, when the metal casting is used as a master-pattern. It is known, in all other cases the vibration is compulsory. A dotted line in graph for dimension A in Fig.22 shows the shrinkage of 0.1% at regime 4, i.e. the dimension A of some of real made resin patterns.

As a result of the experiments, the following regression equations are worked out for the change of the dimensions A, B and C depending on the vacuum x105 Pa, (õ1)  and  the duration of vibration, s, (x2):

yA = 300.485 + 0.796x1  - 0.027x2 ,     yB = 249.712  - 0.495õ1 + 0.016x2 ,     yC =   40.014 + 0.109õ1  - 0.003x2 .

The equations are worked out again with the help of StatGraphics Centurion. It is known, this kind of computer programs gives full statistical analysis and comments in conjunction with the equations. Generally, the chosen linear regression has an adequate relationship with the experimental data with 95% confidence level.

With the help of master-pattern from Fig.21-a, 30 experimental resin patterns (Fig.23) are made using different regimes with vibration compulsory. A part of routine statistical processing of the results for the dimension A is presented in Fig.24. The analysis shows a normal distribution with 90% or higher confidence. The standardized skewness and kurtosis test fix the symmetry and the flatness of the distributional curve in the permissible limits.

      Fig.23. One of the experimental resin patterns

      Fig.24. A frequency histogram and a part of corresponding statistical data

The results show that the accuracy of the patterns depends on the regimes of the mould making and any work out of the optimum parameters of the process would lead to preconditions for inadmissible variations in the dimensions. Naturally, that puts the technology at a disadvantage compared to the traditional scheme, wherein the plastic moulds, and respectively the resin patterns, have high accuracy. In order to achieve a maximum dimensional accuracy of resin patterns obtained in vacuum-sealed sand moulds, it is necessary to keep hardly the parameters of the process constant. In reality, the checks done for the accuracy show that the manufactured in vacuum-sealed moulds more than 1000 resin patterns meet the specific requirements of customers without exception.


4.4. Strength properties

In the most cases the resin characteristics given from their producers can not serve for the determination of strength properties of the compounds because they consist of various components in different proportions, and they may polymerize in different conditions. That is why, here are presented some results for the bending and impact strength of samples made of different fillers to the resin. The choice of both strength properties is done owing to a cause-and-effect analysis for the destruction of the resin patterns in exploitation. This analysis shows that the destruction happens accidentally or as a result of gross errors in the construction of the patterns and in their production technology.

The routine tests are done according to the standards BDS EN ISO 178:2002 [22] and BDS EN ISO 179-1:2003 [23].

The experimental compounds contain Bulgarian epoxy resin AP1 [24] made with plasticizer, hardener 10% to the resin, and the following fillers:
- Fe-dust WPL, product of Mannesmann - Germany, content of 0.01 %C and ~99.4 %Fe, fraction 0.04-0.06 mm, and weight ratio to   the epoxy resin 1:1, 1.5:1, 2:1, 2.5:1 and 3:1;
- corundum, fractions 0.16, 0.25 and 0.32 mm, and weight ratio to the epoxy resin 1:1, 1.5:1 and 2:1;
- silica sand, fractions 0.10, 0.16, 0.25, 0.315 and 0.40 mm, and weight ratio to the epoxy resin 1:1, 1.5:1, 2:1 and 2.5:1.

The selection of the type, grain size and the quantity of the filler is in conformity with the wear test.

Some samples with Fe-dust contain fiberglass threads in 2% by weight to the quantity of the whole compound. Some compounds are put to vacuum treatment which includes:
- a conventional treatment before pouring for about five minutes in vacuum chamber at vacuum 0.6.105 Pa, regime V-r1. Here is
  impossible a long vacuum treatment because the polymerization begins and goes. When the treatment is made without hardener,
  the effect is low, because it follows the mixing wherefore the air inserts again;
- or vacuum treatment in the period of the polymerization in the vacuum mould, following the scheme from Fig.3, 60-90 minutes after
  pouring at vacuum 0.2.105 Pa, regime V-r2.

The mould type for the resin patterns making does not influence over their strength properties, because of that all results are related to the both examined technologies.

The test findings are given in Fig.25. Only for comparison, the epoxy resin AP1 has a bending strength of 63 MPa and an impact strength of 7.8 kJ/m2. Naturally, a non-filled resin has the best bending strength. The use of non-filled systems, however, is not profitably on account of the decrease of the wear-resistance and greater shrinkage. The economic side is important too, when bigger quantities of filler are inserted in large resin patterns for value reduction. The value of exothermic reaction resin-hardener is decreased too. These things define the use of filled systems.

a)                                                                                               b)
Fig.25. Strength properties of the hardened compounds
a) bending strength; b) impact strength; Fe-ferrous dust; C-corundum; S-silica sand; T-fiberglass threads;
V-r1 - vacuum treatment before pouring; V-r2 - vacuum treatment in the polymerization.

In these awaited results, the samples which contain Fe-dust or silica sand show a bigger strength than these with corundum in all cases. Under equal conditions, this is only due to their different grain form. Whereas the round ferrous and sand grains further the good strength properties, the fractured particles of corundum which have sharp edges make the destruction easier.

It is obvious as well a tendency to the strength decreasing in use of more coarse grain fractions of corundum and silica sand. The quantity of the filler influences still more over the strength properties, at that the best strength have the samples which contain a filler in equal proportion 1:1 to the epoxy resin. Of course, it is known and it is confirmed here again. The combined influence of the grain size and the quantity of the filler on the strength explains with the different structural distribution of the components. Some possibilities for the destruction are presented in Fig.26, where the distribution of the components is made by way of example.

Fig.26. Some possibilities for
    the destruction of resin samples

When the filler quantity is small, its grains are at a maximum distance each to another and the resin "bridges" between the grains have a maximum lengthwise section. At load these "bridges" bend considerably up to the moment of destruction. The destruction is predominantly in cohesive way. In the weight ratio 1:1 and 1.5:1 to the resin, the microscopic examination (x40) shows that the samples have a smooth and shining fracture formed by the epoxy resin. Under the resin the grains of the filler are seen. In this way of destruction the strength properties are the best. In growth of the filler quantity and in use of coarse-grained fractions, the area of the resin connections between the apart grains decreases. It changes the deformation status - the bending of the "bridges" is weakly expressed or it is missing. Then the destruction is predominantly in adhesive way. The fractures are situated in the boundary filler-resin and they are granular and mat. The bright zones of cohesion destruction are seen too, but they have small area. The strength decreases as a whole.

When a threadlike materials are included in the compounds, the maximum strength is achieved. That is an expected result, but must mark that the using of threads - fiberglass, glass cloth, polymer cloth, etc., is justified in judicious limits - in weight ratio of the filler up to 2.5:1 to the resin, and no more than 1.8-2% by weight to the quantity of the whole compound. In opposite case, many air is closed when the compounds are mixed and when the samples are made. That reduces the section of the samples, decreases the strength and creates problems for other service properties. It is known, these materials are applicable in layered large resin patterns, but the problems with air closing are analogous there.

It has a good effect the vacuum treatment of the compound as well. The average growth of the strength for the samples with Fe-dust has the following values: 6% for bending strength and 14% for impact strength in case of vacuum treatment before pouring (Fig.25, columns V-r1), and 14% for bending strength and 21% for impact strength in vacuum treatment in the period of the polymerization of the resin (columns V-r2). It is clear, all that derives from the improving on the structure of the samples.


« « « previous page             next page » » »