JPH02120643A - Method for estimating pressure loss in die and method for planning die flow passage using such method - Google Patents

Abstract

PURPOSE:To accurately estimate the pressure loss in a die by calculating the shape resistance of each of a plurality of divided sections and replacing the shape of each section with a round pipe shape to perform flow simulation and calculating the temp., viscosity, flow velocity and average apparent viscosity of a resin at each part. CONSTITUTION:When lead frames 1 loaded with chips are placed on the cavities 5 of a lower die 4 and a flow passage performing the filling of a resin 7 from a pot 8 is planned, the flow passage is divided into a plurality of sections to calculate the inherent shape passage in each section while, under an initial boundary condition given in a round pipe wherein each section is replaced along a flow direction, a fundamental equation describing the viscosity change transport phenomenon of the resin is solved to calculated the temp., viscosity, flow velocity and average apparent viscosity of the resin at each part. Subsequently, the pressure loss generated in each section is calculated and cumulated to calculate the synthetic pressure loss. By this method, the flow state of the resin in a die is simulated, and the optimum die data and a molding condition can be set.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a mold for molding a thermosetting resin, and particularly to a method for designing a mold flow path suitable for reducing defects in products.

(Prior art) A conventional method for constructing a flow path for a mold with multiple cavities for thermosetting resin is to gradually move the runner bottom of the mold in the direction of resin flow, as described in Japanese Patent Publication No. 17697/1983. The flow path is formed so that the constriction angle of the gate increases gradually with each cavity moving away from the pot along the runner, so that the total pressure loss between the runner and Day 1 is constant for each cavity. I was trying to make it happen.

[Problem to be solved by the invention]

The above-mentioned conventional technology does not consider a method for predicting the flow behavior of resin that flows through a complicated flow path shape while changing temperature and viscosity, and does not consider how to actually fill each cavity with resin at the same speed. It is difficult to improve the quality of the product.

The purpose of the present invention is to quickly simulate the flow state of resin in a mold in a five-product stack, set optimal mold specifications and molding conditions, and significantly shorten product development time and improve product quality. The aim is to improve.

[Means to solve the problem]

The above purpose is to divide the mold flow path into multiple sections, calculate the flow path-specific shape resistance β and flow ff1Q in each section, replace each section with a circular tube shape, and use this shape to prevent flow stains. The temperature, viscosity, flow rate, and H1 uniform appearance of the resin at each part are determined by the viscosity η (l is 17, pressure t) loss Δ
It is achieved by calculating Y) by Δl) = βICLQ and performing flow 4iJ+ prediction. The configuration of this analyzer system is shown in Figure 1. In Figure 1, first, the human power department inputs the resin material f4 value, molding conditions, mold flow path 1 material required for simulation, and then enters the mold flow path specification data, which is the flow path volume. The calculation section and section division section calculate the body IR of the entire flow path and the section to be divided, respectively. The data processed by the section division section is input to a section volume calculation section and a section shape resistance calculation section, where the channel volume and shape resistance of each section are determined, respectively. On the other hand, the data processed by the section volume calculation section enters the circular tube channel replacement section, and each section is replaced with a combination of circular tube channels. Furthermore, the data processed by the channel volume calculation unit and the section volume calculation unit, as well as the transfer time data in the molding conditions input to the input unit, are the section flow rate,
Section passage time calculation 9: Enter section 9, where the flow rate of resin passing through each section and the time for passing through the section are determined. Then, using the data processed by this calculation unit, the data processed by the circular tube flow path replacement unit, and the molding condition data input to the input unit, a flow simulation inside the circular tube is performed, and the resin in each part of the flow path is Calculate the temperature, viscosity, flow rate, average apparent viscosity, etc. The calculation results from the flow simulation section are sent to the viscosity comparison section, and if there is a problem with the results, the simulation is stopped and a part of the input data is changed at the manual condition change section, and the process returns to the beginning. On the other hand, if there is no problem in the viscosity comparison section, the average apparent appearance of each section determined by the flow simulation is calculated as the pressure loss of each section using the viscosity, flow rate and shape resistance of each section. Find the total pressure loss by accumulating this. This result is entered into the pressure loss comparison section, and if the calculated pressure loss is larger than the predetermined pressure loss, it is entered into the input condition change section, where part of the manual data is changed and the process returns to the beginning. If the pressure loss comparison section determines that there is no problem, the output section outputs final mold channel specifications, molding conditions, and other necessary information.

[Effect]

According to the above method, by calculating the optimum mold flow path setting M and the necessary pressure loss, it is precisely determined based on the shape resistance specific to the flow path and the flow rate number. On the other hand, in order to calculate the average apparent viscosity by replacing the complex actual mold flow path shape 4 with a combination of circular pipe flow paths, which requires complicated flow analysis and a huge amount of calculation time,
The calculation time is extremely short, and the prediction accuracy of pressure loss that is sufficient for practical use can be ensured.

〔Example〕

An embodiment of the present invention will be described below with reference to FIGS. 2 to 13 and Tables 1 and 2. First, we will discuss the flow simulation method inside a circular pipe.

Express the isothermal viscosity equation for thermosetting resin using the following model.Time,
C: viscosity increase coefficient, T: absolute temperature, t: time. Also, η. (T) = ctex p (b/T) ”...(2
)to (T)=dez p (e/T) ””'
(3)c (T)=f/T-g...
...(4). (Lr br d+ et
f+g is a resin-specific parameter that is not affected by molding conditions. Equation (1) satisfies the following boundary conditions.

When 1=0, η=η. (T)・・・(5)t=t0
When (T), η=ψ (6) The characteristics of equation (1) at an arbitrary temperature T are shown in FIG.

Inside the mold, the resin flows while receiving heat from the tube wall, so it is in a non-isothermal state in most cases. Next, a method for predicting viscosity in this case will be explained. First, by rearranging equation (1) in a dimensionless manner, the following equation is obtained.

1+τ μ=1-0・・・・・・・・・(7) 17C[T] Here・μ=(η/η.(T)) ・・・・・・・・・(
8) τ=t/10(T) ・・・・・・・・・(
9). This curve is = 0 and μ = 1. τ=1 and μ=
It has the characteristic that ω. This curve is shown in FIG.

Now, in FIG. 3, it is assumed that τ=τ1 and μ=μm, and the time at this time is tll and the temperature is T2.

Then, when time passes △, the temperature also increases by ΔT,
Let us find the new viscosity when the time and temperature each reach t21'r2. According to equation (9), τ is a function of L and T, and the increment Δτ of τ up to the new state τ2 can be found by the following equation.

△,=tan 28 t to '2 △T... (10) a
δT Also, from equations (9) and (3), the following equation is obtained.

Δt and ΔT in equation (10) are known in advance as shown in Figure 3, and T=T in equation (11) and T=T in equation (12).
By substituting T, τ=τ, Δτ can be found.

Therefore, τ2=τ1+Δτ−・Φ・・−(13), and (7)
μ2 can be found by using the formula = 2.

Then, the following equation is obtained from equation (8).

η=η. (T) μ00T) ... (15) This (
15) In the equation, T=T2. Substituting the value of μ=μ2, η2=
η. (T2)μ2C(Tll... From (16), the viscosity η2 in the new state is determined.

By repeating this method from τ=0 to 1, it is possible to calculate the viscosity change from the initial state in a non-isothermal state to gelation.

In order to analyze the state in which resin is flowing inside a mold, it is necessary to solve the above viscosity prediction method in combination with the basic equations of various conservation laws.Each equation for a circular pipe flow path is as follows. Shown below.

Q=2πj" uzYd Z""・(17) This is a conservation equation for quantity and energy. In equations (17) to (19), Q: flow rate, R: circular pipe radius, υ2: flow velocity in the pipe axial direction, γ: Pipe radial distance, Z: Pipe axial distance, P: Pressure, η: Viscosity, P:
density, T: temperature, t: time, λ: thermal conductivity. (1
7) to (19) to the isothermal viscosity equations (1) to (4), (7
) to (16), and solve the given initial conditions and boundary conditions using numerical analysis methods such as the finite difference method and finite element method, it is possible to simulate flow in a circular pipe channel. Can be done. An overview of the simulation program 11 used in this example is shown in the fourth section. In the output, the average apparent viscosity value is also determined when the Newtonian fluid is assumed to be flowing in a steady isothermal layer.

Table 1 shows the physical properties of the epoxy resin for semiconductor encapsulation used in this example.

Table 1 Here, P I G and λ are values obtained using a commercially available thermophysical property I11 constant device. Also, a, b, d, e, f.

g is a parameter in formulas (2), (3), and (4),
These values are calculated by combining several types of circular pipe channels with different diameters and several types of mold temperature conditions, actually measuring the average apparent viscosity of the resin under each condition, and extrapolating based on these characteristic values. It was determined by the method and curve fitting method. FIG. 5 shows a comparison between the calculated value and the measured value when the average apparent viscosity ηa in the circular pipe flow path was calculated by inputting the physical property values in Table 1 into the simulation program shown in FIG. 3. The two values matched very well at each mold temperature, verifying the validity of this simulation method.

Figures 6-a to 6-d outline the manufacturing process for resin-sealed semiconductors. FIG. 6-a is an enlarged view of a state in which chips 2 are mounted on a lead frame 1 and connected with gold wires 3, and shows a multiple lead frame 1 on which a plurality of these devices are mounted. Figure 6-1) shows the process of resin-sealing the device described in -. First, the lead frame 1 is placed in one of the cavities 5 of the lower mold 4, and the upper mold 6 is closed and fixed within the mold 1. On the other hand, the thermosetting resin 7 preformed into a tablet shape is exposed to a high frequency F thermal bridge (see 1).
The plunger 9 attached to the molding machine (not shown) is lowered. At this time, it receives heat from the mold heated by the heater 10. The resin 7 is melted and the runner 11.gate 1
2 into the cavity 5. After filling of the resin 7 into the cavity 5 is completed.

After a predetermined period of time has elapsed, the resin 7 hardens and is removed from the mold in the state shown in FIG. 6-c. And lead frame 1
After passing through the cutting and bending steps, the final shape shown in FIG. 6-d is completed. In FIG. 6-b, if the filling time of the resin 7 into each cavity 5 is different, the viscosity, flow rate, and curing state of the resin will be different.

Due to these factors, there are problems such as residual voids and deformation of the gold wire 3. Therefore, next, a flow path design method for making the filling condition of the resin 7 uniform will be described.

In order to equalize the filling situation of resin into each cavity, it is necessary to create a mold flow path that can equalize the sum of the pressure loss at the runner and gate part to reach each cavity when distributing the same flow rate to each cavity. All you have to do is ask. Therefore, calculations are required to estimate the pressure loss in the flow path, but in Section 6-b
The cross-sectional shape of the runner 1] and the gate 12 in the figure is a semicircular bottom or an inverted trapezoid, and moreover, in most cases, the cross-sectional area changes along the flow direction. Like this 'f! i. The calculation time required to perform the flow simulation using the rough boundary conditions is enormous, making it impossible to create a practical design system. Therefore, in the present invention, the following method is used. In other words, the flow path is divided into multiple sections, the flow path-specific shape resistance β and flow rate Q are calculated in each section, each section is replaced with a circular tube shape, a flow simulation is performed with this shape, and the resin of each part is calculated. The temperature, viscosity, flow rate, and average apparent viscosity 7a were calculated, and the pressure loss ΔP was determined using the following formula.

ΔP=β了aQ (20) Fig. 7-a shows an example of one section of a flow path in which the cross-sectional area changes along the flow direction. X is the distance along the flow direction from the reference point,
X2 is the start point and end point of the section, respectively. Furthermore, W is the channel width and h is the channel depth, both of which are functions of X.

Now, in this interval, if W≧h, β can be found by the following equation.

Here, F is a shape factor determined by the cross-sectional shape of the channel and the ratio of W and h. In equation (21), W, h, and F are given as functions of X, and β can be found by performing approximate integration analytically or using Simpson's formula. In addition, W
<h is used. Furthermore, if a location where W≧h and a location where w<h coexist within the section, equations (21) and (22) may be used at each location and added together. Fig. 7-b shows a circular pipe flow path replaced with Fig. 7-b. section x
1 to X2 are circular pipes with -like pipe diameters. This diameter is calculated and determined so that the section volume in Fig. 7-a and the section volume in Fig. 7-b are the same. Moreover, the broken line in FIG. 7-b shows the shape of the section before and after the section X1 to X2 in the flow path tapering along the flow direction after replacement with a circular pipe.

With this substitution, the flow simulation method in the circular pipe described above can be used as is, and the average apparent viscosity r)a in each section can be estimated. Also, the section passing flow rate Q is the mold flow channel volume ■.

From the resin injection time tp of the plunger, first, calculate the flow rate Qp discharged from the pot as Qp = V t / t P
It is calculated in advance by calculating and further distributing Qp according to the number of branches of the flow path. Through the above steps, the pressure loss in each section is determined, and this is accumulated to give the total pressure loss ΔPT.
is found.

FIG. 8 shows a flowchart for designing a runner and a gate for equalizing the resin filling situation in each cavity in a multi-cavity mold having the structure shown in FIGS. 6-6. first,
Enter resin physical property values, molding conditions, and mold specifications. Among these, the resin physical property values are the physical quantities for each resin shown in the symbols in Table 1. Molding strip setting plunger maximum pressure PM, resin preheating temperature.

is the mold temperature. For the mold specifications, provisional values are input for specifications other than fixed factors such as the pitch between cavities and flow path layout for the first calculation. Next, the volume Vf of the entire flow path is calculated, and the flow path is divided into sections.

At this time, the number of joint surfaces in the section is at least the same as the number of branches in the flow path, so that there is no branching of the resin flow within the section. Then, the section volume Vn and section shape resistance βn are calculated. From the above, first, the flow rate Qp discharged from the pot
is calculated as Qp = V t / t p, and Qp is further distributed according to the number of branches of the flow path to determine the flow rate Qn in each section. At this time, it is assumed that the same flow rate is distributed to each cavity. Then, the resin passing time tn for each section is determined by Vn/Qn.

Also, from Vn, the diameter of the circular pipe flow path having the same volume is determined for each section. A flow simulation in the circular pipe flow path replaced in this way was performed sequentially from the upstream side under the given initial and boundary conditions, and the temperature, viscosity η, flow velocity distribution, and average appearance of the resin in each section were calculated. Calculate the viscosity ηa. Here, at a predetermined position in the radial direction of each section, the downstream and upstream η
, and check that η does not rise more on the downstream side than on the upstream side. If η increases on the downstream side, the molding conditions and mold specifications are changed and recalculation is performed. This is because under conditions where η increases during flow, a highly viscous resin that has undergone a hardening reaction will flow into the cavity, which is extremely likely to cause molding defects, and the purpose is to prevent this in advance. be. Once a condition in which η does not increase is obtained, first, the pressure loss ΔPRn in each section of the runner is expressed as ΔpR.
1=βn·1anQn, and the pressure loss ΔPRT in the runner is calculated by accumulating this. Next, the pressure loss ΔPJIS≦ΔPRT in the runner, which was set in advance for comparison, is compared, and when PR8≦ΔP, A, the mold flow path, specifications, resin injection time of the plunger, resin preheating temperature, Change the value of at least one condition of the mold temperature and re-enter it to find the 8PjlT at that time, change the conditions one by one and calculate ΔP.
When us>ΔPRT is satisfied, runner specifications and molding conditions are determined. This is because when the pressure loss approaches the maximum plunger pressure PM set in the molding machine, the plunger cannot descend at a constant speed, and the resin stagnates during flow.
There is a phenomenon in which molding defects occur frequently, and the objective is to first prevent this in the runner section. Note that 8PR5 is set to a value much lower than PM. Subsequently, the pressure loss at each gate section is set so that the sum of the runner section pressure loss and the gate section pressure loss up to each cavity is constant. At this time, compare the preset pressure loss Ps with the calculated total pressure loss ΔPT, and if ΔPS≦ΔPT, change the molding conditions and some of the values of the flow path specifications again and recalculate. I do.

Then, when ΔPs>8PT, the pressure loss ΔP+In that the gate portion should have is determined.

Note that ΔPs is an intermediate value between ΔPR5 and PM.

Then, △Pan and the average apparent appearance of the gate section obtained by flow simulation are the viscosity 'QLLa: and the shape resistance β that each gate should have from the gate section flow, f & QQn.
31 mouths β3. n=ΔPan/”i a a-Q
Calculate by an. Then, using the relational expressions shown in equations (21) and (22), the specifications of the gate section are calculated backward.

Furthermore, if the obtained specifications do not satisfy the constraints at the gate, such as the upper and lower limits of the angle of the tapered part along the flow direction, and the upper and lower limits of the depth, Part of the conditions are changed and the gate specifications are determined when the constraint conditions are satisfied.

Runner 11 of the mold designed using the five methods shown in Figures 9-a and 9-a. The structure of gate 12 is shown.

Figures 9-a and 9-a show the runner 11 when viewed from above in Figure 6-b. This is the structure of games 1 to 12. Here, the 15th
In Figure-B, the runner 11. Cavity 5 is omitted. Figure 9-b is a side view of Figure 9-a, and the runner 11 has a structure in which the depth is gradually decreased along the flow direction in order to reduce the time difference in the arrival time of the resin to each cavity. did.

Figure 9-a is the A-A cross section and B-B in Figures 9-a and 9-b.
This is a diagram showing a cross section, and the structure is such that the aperture angle θ of the gate is widened on the downstream side, and the sum of pressure losses in the runner part and the gate part up to each cavity is equalized. In addition, when there is a constraint condition on the gate aperture angle, the width and depth of the gate outlet 13 may also be changed to ensure the necessary pressure loss value at each gate h.

Next, the present invention will be explained in detail. Section 10-a.

Figure b shows the results of 9-a and e without using the flow 1h simulation method and assuming that the viscosity of the resin is constant.

This figure shows the state of resin filling in the cavity using the conventional method of designing a mold having the structure shown in C. In Figure 10-a, the vertical axis is the non-dimensional filling rate of resin in each cavity, the horizontal axis is the non-dimensional filling time after the resin starts entering the cavity, and the dotted line in the figure is the non-dimensional filling rate of resin in each cavity. In the conventional method, which shows ideal filling in which filling is performed at the same speed, the filling of resin into each cavity deviates significantly from the ideal filling. Fig. 10-1) shows the filling state of the resin 7 at the non-dimensional filling time of 0.5 in Fig. 9-a, in which the resin fills quickly from the cavity 5 on the upstream side.

Figure 1.1-a shows the resin filling situation in the cavity of a mold designed according to the present invention using a flow simulation method. Filling conditions very close to ideal filling were obtained in all cavities. Fig. 11-b shows the filling state of the resin 7 at the dimensionless filling time of 0.5 in Fig. 11-a, and all cavities are filled with the resin 7 in the same way.

FIG. 12 shows a comparison of the maximum resin flow rate in the cavity and the resin residence time. Here, tp indicates resin injection time. In molds designed using conventional methods, the flow velocity in the downstream cavity is high. This shows that due to the filling situation shown in Figures 10-a and 10-b, the flow rate shifts downstream every time the resin is completely filled from the upstream side. On the other hand, in the mold designed according to the present invention, as shown in Figures 1.1-a and 1.1-b, the resin is filled uniformly, so no flow concentration occurs, and the flow velocity exhibits approximately the same value in each cavity. On the other hand, in molds designed using conventional methods, the resin residence time increases in the upstream cavity. Here, the residence time is the time from when the resin completes filling in the cavity until the plunger stops. The mold designed according to the present invention has almost no residence time in any cavity.

FIG. 13 shows a comparison of the incidence of gold wire deformation defects.

Here, results are shown when two types of gold wires, gold wire A with a large diameter and gold wire B with a small diameter, were used. In the case of gold wire A, there is no defective wire deformation in either the mold designed by the conventional method or the mold designed by the method of the present invention. On the other hand, in the case of gold wire B, molds designed using the conventional method have a high incidence of cavities on the downstream side.

This corresponds to the increase in flow velocity in FIG. 12, and shows that as the flow velocity increases, the gold wire with a small diameter and low rigidity becomes more easily deformed. On the other hand, in the mold designed by the method of the present invention, there is no increase in flow velocity, so no defects occur even in the gold wire B.

Table 2 shows a comparison of the incidence of appearance defects of molded products.

Table 2 When tp is 18s, no defects occur with either method, but when tp becomes 30s, the defect rate becomes 100% for the mold calculated using the conventional method.

This is because under these conditions, the residence time of the gate shown in FIG. 11 becomes extremely long, and the resin hardens without sufficient pressure being applied inside the cavity. On the other hand, in the mold according to the present invention, which has almost no gate retention, no defects occur even under these conditions.

〔Effect of the invention〕

According to the present invention, it is possible to quickly and accurately design optimal flow path specifications for mass-produced molds and select optimal molding conditions on a desk, thereby shortening the new product development period by abolishing the prototyping process and reducing molding defects. The effect of reducing costs by reducing the diameter of the gold wire and reducing the diameter of the gold wire is very large.

[Brief explanation of drawings]

Figure 1 is a configuration diagram of the analysis system of the present invention, Figure 2 is a characteristic diagram of the isothermal viscosity model used in an embodiment of the present invention, and Figure 3
The figure is an explanatory diagram for calculating viscosity change in a non-isothermal state,
Figure 4 is a schematic flowchart of flow simulation.
FIG. 5 is a comparison diagram of the measured value and calculated value of the average apparent viscosity ηa, and FIGS. 6-u to 6-d are explanatory diagrams of the semiconductor resin encapsulation process. Figures 7-oL to 7-b are explanatory diagrams for calculating the pressure loss in the flow path, Figure 8 is a flowchart for designing the runner and gate of a multi-cavity mold, and Figures 9-α to 9- c.
The figure shows the mold runner and gate structure, No. 10-0,
Figure 10-b is a diagram showing the resin filling situation in the cavity using the conventional method, Figures 11-a and 11-b are diagrams showing the resin filling situation in the cavity using the method according to the present invention, and Figure 12 is a diagram showing the resin filling situation in the cavity using the method according to the present invention. Fig. 13 is a comparison diagram of the inner resin maximum flow rate and resin retention residence time, and Fig. 13 is a comparison diagram of the incidence of gold wire deformation defects. 4...Lower mold, 5...Cavity, 6...Upper mold, 7
...Thermosetting resin, 8...Pot, 11...Runner, 12...Gate. 1st er

Claims (1)

[Claims] 1. In designing a mold with a pot and a flow path connected to the pot, the flow path volume V_f is calculated from the flow path specifications set in advance, and the thermosetting resin poured into the pot is calculated from the flow path specifications set in advance. Set the resin injection time t_p of the plunger for injecting the resin into the flow path, and set the resin flow rate Q_p injected from the pot into the flow path as Q.
_p=V_f/t_p is calculated, Q_p is distributed according to the number of branches of the flow path, and the flow rate Q_ of any part of the flow path is calculated.
In addition to calculating n, the flow path is divided into multiple sections with an arbitrary cross section perpendicular to the resin flow direction, and in each section, the flow path width, depth, cross-sectional shape, and length are determined. Basic equations that calculate the shape resistance value β_n, replace each section with a circular tube shape along the flow direction, and describe the viscosity change transport phenomenon of the resin under the given initial and boundary conditions within the circular tube. Solve and calculate the temperature T of the resin in each part of the circular tube.
, viscosity η, flow velocity υ, etc., as well as the average apparent viscosity η
_α_n is calculated, and the pressure loss ΔP_n occurring in each section is calculated as ΔP_n=βn@η@_α_nQ_n, and ΔP_
A method for predicting pressure loss in a mold, characterized in that a total pressure loss ΔP_T is obtained by accumulating n. 2. In claim 1, the pressure loss ΔP_S preset for comparison is compared with the calculated total pressure loss ΔP_T, and when ΔP_S≦ΔP_T, the flow path specifications and the injection of the plunger are determined. Change the value of at least one condition among the time, the temperature of the mold, and the preheating temperature of the resin and re-enter it to find the ΔP_T at that time, change the conditions one by one, and ΔP_T.
A mold flow path design method that determines flow path specifications and molding conditions within a range that satisfies P_S>ΔP_T. 3. In claim 1 or 2, ΔP_S>ΔP
The channel specifications and molding conditions are within the range that satisfies _T and that the resin viscosity on the downstream side does not increase more than the upstream side at a predetermined position in the radial direction within the circular pipe replaced in each section. Mold flow path design method to determine. 4. In claim 1, the mold has a structure in which a plurality of cavities branch-connected to the runner connected to the pot via gates are arranged along the runner, and the number of joint surfaces in each section is at least Assuming that the number of branches is the same as that of the flow path and that the same flow rate is distributed to each cavity, for each cavity so that the total pressure loss in the runner and the gate is constant for each cavity. A mold flow path design method characterized by changing gate specifications.