Plastic and nonmetallic article shaping or treating: processes – With measuring – testing – or inspecting
Reexamination Certificate
2000-07-13
2003-02-11
Heitbrink, Jill L. (Department: 1732)
Plastic and nonmetallic article shaping or treating: processes
With measuring, testing, or inspecting
C264S040500, C264S328100, C700S200000
Reexamination Certificate
active
06517754
ABSTRACT:
CROSS REFERENCE TO RELATED APPLICATION
The present invention claims the priority of German Patent Application 198 01 881.9, filed on Jan. 20, 1998, the disclosed content of which application is also hereby expressly made the subject-matter of the present invention.
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to a method for evaluating injection-molded parts, which are produced in an injection-molding machine to determine if the molded parts are good or bad, and more especially a plastic injection-molding machine for processing plasticizable substances, such as plastics materials, pulverulent substances or ceramic substances.
2. Prior Art
Such a method is explained in an article by M. Hoyer and E. Duffner entitled “Grafische Unterstützung vereinfacht Bedienung” in: Kunststoffe 84 (1994), Book 10, Pages 1388-1394, and it is illustrated, more especially, in
FIG. 6
there. A corresponding illustration of prior art is also found as
FIG. 1
belonging to this application. In this prior art, a plurality of parameter curves are initially determined as a function of a second parameter, for example the internal mold pressure over the cycle time. The curves, whereby non-defective good moldings have been produced, are stored in such case. A mean value curve M
1
is then formed from these curves. From the values which are available in the y direction, a standard deviation is now calculated in known manner, e.g. according to the formula:
Δ
y
=
z
*
1
m
-
1
(
∑
i
=
1
a
y
i
2
-
1
m
(
∑
i
=
1
a
y
i
)
2
)
with:
z=standard normal variable, e.g. z=3.29 for 99.9% control limit,
m=number of measuring cycles,
y
i
=value of the first parameter in the i
th
measuring cycle.
For a constantly calculated curve deviation, the span width R is now calculated by:
R
i
=Max
x
{|y
i
(
x
)−
y
(
x
)|−Min
x
(|
y
i
(
x
)−
y
(
x
|)}
with:
| |=absolute amount,
y(x)=mean value of the first parameter y in dependence on the second parameter x.
Since an actual value curve y
i
(x) is equal to the reference curve y(x) at at least one point, R
i
can therefore be calculated, so that the tolerance can be calculated according to the formula:
Δ
y
=
z
*
1
m
-
1
(
∑
i
=
1
a
R
i
2
-
1
m
(
∑
i
=
1
a
R
i
)
2
)
An assessment can now be made as to whether or not, in a subsequent production cycle, the parameter curve lies within the permissible range given by the tolerance band. If the permissible range is departed from on one or more occasions, the molding is judged to be a bad molding, according to the assessment criteria, and it is rejected.
However, the disadvantage of this way of determining the tolerance band in a manner known per se is that the tolerance band, which is determined only in the y direction, becomes very narrow. In particular, often considerable fluctuations in the value of the first parameter, that is to say, for example, the internal mold pressure, the velocity or the screw path, such fluctuations possibly occurring, for example, during the transition from the mold filling phase to the dwell pressure phase. As such, plurality of injection-molding cycles therefore depart from the tolerance band at this point, with the consequence that moldings, which should still be accepted in terms of their quality when properly considered, are wrongly eliminated.
Similar methods are known, for example, from DE 35 45 360 A1. A good molding curve is also determined there, and such curve is also provided with a tolerance band possibly in sections. This tolerance band is formed by an upper envelope curve and a lower envelope curve, and the curves are produced from a plurality of good molding parameter curves. This envelope curve therefore includes all of the injection-molding cycles which have previously already resulted in good moldings. However, no tolerance band is determined hereby, and no consideration is even given to any influences which are not reflected in the curve pattern, such as, for example, a high level of energy consumption. Above all, however, no recognized statistical evidence regarding the quality of the moldings produced is hereby possible.
Apparent from EP-0 452 513 B1 is a method wherein pressure curves are plotted in dependence on one or more parameters. Parameter values for determining a permissible range can then also be inputted, if desired, in sections in dependence on these curves and the values determined thereby. After a certain lead time, the last three pressure curves are preferably considered, and corresponding values are selected by the user from a numerical table representing the various pressure curves and inputted by hand. No tolerance band calculation is hereby effected.
A means for regulating product properties is known from DE 41 08 992 C2, wherein the particular weight of the molding is correlated with the work applied by the piston to produce the molding. The molding, which is produced, is judged to be a good or bad molding in dependence on the work actually required. The bad molding is, therefore, only evaluated in dependence on one parameter.
A method of assessing good or bad moldings is known from EP 0 317 992 A2, wherein the temperature of the injection mold assembly, the molding substance, the plasticizing cylinder or the nozzle is correlated with the time. Here also, a molding is only assessed in dependence on one parameter.
It is known, from U.S. Pat. No. 5,550,744, to select a plurality of quality parameters and to determine, in dependence on the control parameters, data fields where the values of good moldings have to lie at the intersections of the fields. However, a tolerance band over the entire range is not, therefore, determined in dependence on only two parameters which equally determine the parameter curve. In consequence, the control work needed there for monitoring purposes is disproportionate to the evaluation obtained thereby, especially since the known problem also still arises here, i.e. that the tolerance band becomes narrowest at the precise location where the parameter fluctuations are the greatest.
SUMMARY OF THE INVENTION
On this basis, the basic object of the present invention is to calculate a tolerance band, by means of which calculation a suitable tolerance band is provided, and more especially, when there are considerable fluctuations in the monitored parameter, in order to reduce the reject rate further to the actual bad moldings.
This object is achieved by setting the permissible range using a standard deviation calculated in dependence on a second parameter.
Whereas, in known methods, the standard deviation has usually only been calculated hitherto in the y direction, in the present invention an additional calculation is now also effected in dependence on the x co-ordinate. This causes a wider tolerance band to be set, more especially in the regions where the monitored parameter fluctuates considerably in the y direction, so that injection-molding cycles now also still produce good moldings which would hitherto have been eliminated. However, since the tolerance band could already be exceeded in such manner, with minimal delays, e.g. delays with respect to time, solely by shifting the fluctuation range in the illustrations to the right, the reject rate is noticeably reduced thereby. Nevertheless, it is ensured that parameter curves, which lie outside this permissible range, cannot lead to good moldings, so that the products there can be reliably eliminated.
REFERENCES:
patent: 5173223 (1992-12-01), Kamiguchi et al.
patent: 5291423 (1994-03-01), R{umlaut over (oo)}sli
patent: 5550744 (1996-08-01), Steinbichler
patent: 5800748 (1998-09-01), Kamiguchi et al.
patent: 5815397 (1998-09-01), Saito et al.
patent: 5871676 (1999-02-01), Fujita et al.
patent: 6019917 (2000-02-01), Ryckebusch et al.
patent: 35 45 360 (1987-06-01), None
patent: 41 08 992 (1996-06-01), None
patent: 0 317 992 (1989-05-01), None
patent:
Heitbrink Jill L.
Rabin & Berdo PC
LandOfFree
Method for evaluating injection moulded parts does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for evaluating injection moulded parts, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for evaluating injection moulded parts will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3128677