A technical analysis of check valves.
Through the years, injection molding machines have continually improved. As modern control systems have been incorporated, machine size has grown to accept larger molds, and accumulator systems have been added to improve cycle times. If the injection molding industry is to continue to progress, however, a number of goals must be met. They include future improvements in parts quality, reduction of scrap levels, and more efficient utilization of regrind. Resin properties must be maintained throughout the plasticizing process, and statistical process control (SPC) and statistical quality control (SQC) must be fully utilized.
Accomplishing such goals requires injection molders to examine and eliminate variables that are creating variations. This article examines typical problems and offers practical, inexpensive solutions. Its purpose is to properly classify non-return valves that are available, and technically describe their operation. Proper technical analysis of valve design should help eliminate confusion regarding valve operation and performance capabilities.
Non-return valves are classified according to type of closure (flow and preclose) and sealing (seat and sliding).
Types of Closure
Flow closing valves are ball-, poppet-, and ring-type check valves that close on the inject stroke. Until 1985, no others had been available.
Initially, preclosing valves were poppet types, with a spring assist that closed when the screw stopped rotating (U.S. patent no. 4,512,733, Eichlseder). They were followed by patent no. 5,151,282 (Dray), a valve that features unique preclosure. Rather than springs or mechanical assist, it has only pressure declination that is created as screw rotation is stopped in the normal machine cycle.
Another spring actuated valve, patented in November 1992, features a thrust bearing on the poppet. It is known as the Auto Shut (patent no. 5,164,207, Durina). Patent no. 5,258,158 (Dray) is a pressure actuated, preclose type. However, as in patent no. 5,151,282 (Dray), the piston is outside rather than inside. Also, the valve does not incorporate any mechanical components for its preclose.
Flow closing can be described as "leakage closing." As the screw moves forward with the valve still open, leakage or upstream flow occurs. All ring, ball, and poppet valves require flow (leakage) for closure. Following is an examination of the variables involved in flow closure.
First, the pressure drop required for closure at a given flow rate (inject speed) must be determined. The following formula may be used to calculate pressure drop:
Delta P = Q12uL/W[h.sup.3]
where the following are true:
1) Q = Flow rate in [in.sup.3]/sec (determined by settings of initial injection speed) if decompression is used. Upstream resistance can be neglected if closure occurs prior to filling the void created by decompression.
2) Shear rate, slit orifice = 6Q/W[h.sup.2]
W = Phi D (ring valve);
h, or clearance, open to close (high compression), is normally 0.100 inch; h, or clearance, open to close (free flow), is normally 0.150 inch or greater.
3) Viscosity can be determined by consulting curves showing shear rate vs. viscosity for the applicable resin. If closure occurs prior to filling the void created by decompression, initial pressure drop on a given valve design may now be calculated.
Predicting closure on flow closing valves is difficult at best. If closure does not occur prior to filling the decompression void, upstream resistance must be incorporated.
4) Upstream resistance is the screw configuration or metering section helical channel length, width, and depth. Pressure drop for this section can be calculated with the same formula:
Delta P = Q12uL/W[h.sup.3]
Tan (helix angle) = lead/(phi) (D);
L = helical length per turn: (phi) (D)/sin (helix angle); and u = viscosity. Before viscosity can be determined, shear rate must be calculated:
shear rate = 6Q/W[h.sup.2]
For purposes of calculating shear rate for the partially closed valve, the distance open is unknown. It is known, however, that shear rates may vary: At 0.100 h, shear rate = 42.46 Q (diameter of ring valve = 4.5 inch), and at 0.010 h, shear rate = 4246.00 Q. The first problem is determining the distance h. As upstream movement is initiated, h decreases. We do not know when movement is initiated or the rate at which h declines, but we do know the initial h value.
Viscosity will decrease as shear rate increases (viscosity = shear stress/shear rate). Therefore, large viscosity variations and degrading can occur at high shear rates with sensitive resins. The viscosity variations are introduced to the resin in the screw metering section.
Pressure drop required for closure varies with valve type. Distance to closure is very important; upstream flow will be reduced as the screw configuration, filled with resin, resists upstream flow. Pressure drop decreases if flow decreases, thus increasing the time to closure. Excessive upstream leakage results in reverse screw rotation, which is the same as partial closure or nonclosure. The greater the distance to closure, the greater the chance that closure variations will occur as a result of upstream resistance. Pressure drop required for closure is thus reduced.
Although injection velocity should be determined by viscosity and mold fill requirements, it is often determined by valve requirements because closure requires rapid inject speeds.
Seat sealing valves include all valves except patents no. 5,151,282 (Dray) and 5,258,158 (Dray). Following is a discussion of variables involved in seat sealing.
Seal alignment is primarily a function of machining accuracy. Most valves have softer seats that allow for any misalignment. Ideally, the harder portion of the valve will cause the softer seal to fail slightly in bearing to form a perfect seal. Unfortunately, the hardened portion of the valve is normally movable, requiring constant reseating. If wear is incurred in the downstream face of a ring-type valve, misalignment occurs, which can cause partial closure and additional barrel wear.
Valve wear and seat contamination. Normally, wear in ring valves is on the downstream face; it is a result of flow (in filled resins) and peening that is caused by recovery startup forces. These forces are a function of screw flow capabilities and ring face area. Because screw startup is normally rapid, high ring velocities are generated that cause wear and, occasionally, ring breakage.
As the downstream faces become worn, distance to closure increases and closure becomes increasingly erratic.
With the variables involved, predicting the life of a flow closing valve is difficult at best. Occasionally, valve life is determined when the screw rotates in the opposite direction during injection. This is caused by excessive upstream leakage resulting from problems of valve seating or valve O.D. wear.
It soon became obvious to machine designers and molders that flow-type valve closure and seat sealing were uncontrollable variables. Consequently, cushion became a solution.
Cushion is the amount of forward stroke remaining after mold fill and pack; its primary purpose is to compensate for machine and valve inconsistencies. Excessive cushion can cause color bleeding and degradation, as flow is normally to the center. With excessive cushion, material that remains in the cushion area tends to cool and increase viscosity during recovery. Even if it is removed during injection, the material increases the potential for residual stress; in many cases, it causes color bleeding or degradation. Self cleaning during each shot, which can be achieved by minimizing cushion and streamlining flow profiles, is desirable because it eliminates potential for degrading and streaking. Today, most molders use excessive cushion because of variations in performance of flow-type closing valves. If minimum cushion were used, short shots would occur.
Today, two modes of molding machine operation are commonly used. One is precise molding, which normally makes use of statistical process control (SPC). When cushion is penetrated, alarms warn the operator of parts that are, potentially, off-specification. (This requires excess cushion.) In some operations, parts are then diverted to be ground for regrind. In others, parts are checked to see if they are still acceptable. In either case, additional cost is incurred. In nearly all cases, observed part weights are only randomly checked--normally, only when cushion has been penetrated.
The other common mode of operation is nonprecise molding. In overpacking, cushion is not utilized; stroke length is longer than necessary for filling the mold, and more than the required amount of injection pressure is used to ensure mold filling. Because cushion variation cannot be tolerated, the mold is, essentially, overfilled. This type of molding places additional stress on machine components. It also imparts nonuniform stresses in molded parts that may cause warpage after cooling.
In most cases, both modes of operation produce acceptable parts. However, if off-specification parts are found at the customer's facilities and subsequently rejected, high scrap rates and even greater costs will follow.
The solution for ideal operation of injection molding machines is the same today as 40 years ago: precise stroke length, with no cushion variation if cushion is being used. This requires complete valve closure on every stroke. If cushion is not being used, complete valve closure permits the molder to precisely fill the mold instead of overpacking it. In either case, partial closure or nonclosure may occur at any time with valves that close by flow and use a seat--even with new machines and valves. This is commonly referred to as a short shot.
Many attempts have been made to improve valve performance. Literally hundreds of variations of ring and ball check valves exist, including three- and four-piece streamlined flow types and poppet types. The two areas that have not been improved are seat-sealing and flow- or injection-closure. Although pre-closing has long been seen as a solution, it was previously thought to be possible only through downstream accumulation in a separate cylinder.
Patent no. 4,512,733 (Eichlseder) introduced a mechanical solution, which did not address the seat problems. Instead, mechanical solutions introduced other problems--spring life in the unfriendly plasticizing environment, and the degrading potential of "dead spots" necessitated by the design.
The same environmental problems pertain to patent no. 5,164,207 (Durina), which also includes a spring for preclose coupled with a thrust bearing. Mechanical failure of such designs in the plasticizing and injection environment is an obvious consideration.
Patent no. 5,151,282 (Dray) was issued in September 1992 for a revolutionary non-return valve. A second patent was issued in October 1993 (no. 5,258,158; Dray). The design consists of a two-diameter piston that closes by pressure prior to the inject stroke. Closure is of a sliding type without a seat.
The valve eliminated problems associated with flow-type closure into a seat. Setpoint of stroke length is achieved rearward at the end of recovery by stopping screw rotation. (Normally, inject cylinders are relieved of pressure at the same point.) This causes upstream pressure to rapidly decline, which quickly closes the valve. Closure distance is determined by recovery variables. Three recovery variables--viscosity, flow rate, and piston area difference--determine the valve open position and, therefore, closure distance.
Following is a calculation of valve open clearance for a valve of 3.5-inch diameter:
shear rate = 5.58 Q/phi(Ro + Ri) [(Ro + Ri).sup.2] = 50,000 [sec.sup.-1]
Q = phi (Ro + Ri) [(Ro - Ri).sup.3] (delta P)/12uL
Q = rate = [in.sup.3]/sec = 3 or 300 pph LDPE [approximately] 48 lb[/ft.sup.3]
h = clearance; u: viscosity = 0.002 lb sec/[in.sup.2]; L = land length = [similar]0.030 inch; delta P = 320 psi, assuming set back pressure of 50 psi hydraulic or 500 psi plastic. Screw required pressure = valve area ratio x plastic back pressure or 1.64 x 500 = 820 psi - 500 psi = 320 psi.
Rearranging and solving for Ro - Ri, we find that the clearance would be 0.02 inch. This calculation is for one exit hole where we have two; the shape of the hole will be part elliptical (valve body) and part circular (piston). Therefore, the clearance will not be exact, but rather a close approximation.
In this example, therefore, the piston is required to travel only 0.020 inch to completely close. Closure is complete when the small diameter of the piston covers the exit holes. Leakage during closure is virtually unmeasurable and certainly repeatable.
Valves that require precise closure and cannot allow for seat contamination (water, pneumatic, and hydraulic) normally do not have seats--their closure is a sliding type. Closure in the Dray valve is accomplished by sliding over the exit holes; contamination is not a factor. Seat problems are eliminated by this type of closure. Prior to injection, the amount of captured material in the accumulation area will be determined by three factors: length of machine stroke, pressure established at the end of recovery, and viscosity.
This valve design eliminates partial closure because scaling occurs by sliding. Also, as the screw stops its rotation, upstream pressure declines, initiating closure by forcing rearward movement of the piston. Rearward force is established by the difference in areas. Therefore, closure may be defined as follows: Closure = screw pressure x piston small-diameter area [is less than] accumulated pressure x piston large-diameter area.
In billions of cycles, the valve has never had a partial closure. Its design permits nothing other than complete closure. But what is required of molding machines to utilize this revolutionary valve?
First, consistent stroke length. Few machines are capable of holding position at the end of recovery; instead, they drift. The higher the back pressure, the greater the drift. Drift can be reduced by minimizing back pressure, and can be eliminated in machines that maintain back pressure at the end of recovery. Drift tends to compress material upstream on the screw, and therefore can affect shot size by influencing closure time and position.
Back pressure is always negative. If high back pressures are required to make a part, the screw design is wrong and should be replaced. In most molding machines, 80% of total energy is used by the screw drive. Raising back pressure reduces efficiency of the screw design and therefore increases overall energy consumption.
Because the valve opening is established primarily by operating conditions, the shear stress applied to the resin furnishes a controllable source of mixing and dispersing that is not available in any other valve design. Therefore, back pressures greater than minimum machine requirements are seldom used.
Molding machines that have the proper screw design and are capable of delivering consistent pressure to the accumulation area can eliminate the second most common reason for shot variation--pressure variation resulting from screw surging, as exemplified in recovery time variations. In a precise volume, varying pressures with viscoelastic fluids alters the mass weight in proportion to their variation.
Variations in viscosity are also caused by poor screw design. "General purpose screw design" is, at best, inadequate. It reduces efficiency and, in many cases, causes off-specification parts. The screw design should determine the amount of viscosity variation and should require minimal, if any, adjustment of back pressure. The Dray valve design reduces viscosity variations and permits dispersive mixing with minimal increase in energy. With proper screw design, back pressures higher than 100 psi hydraulic are not required.
The following examples describe valve performance.
The first involves high-precision molding that utilizes SPC. The machine (a Netstal) does not include a modern screw design as exemplified by recovery variations. In past performance, the machine included a general purpose screw, poppet-type valve; it required running without a cushion and overpacking the mold. Cushion could not be utilized because cushion variation would cause variations in the mold fill. The mold was a multicavity, 4 x 4 stack mold--four tops and four bottoms. The part is the most complex part being molded at the Nypro plant in Clinton, Mass.
Various machines in Nypro used this molding system, which was accompanied by high scrap rates resulting from parts distortion. The Dray non-return valve (DNRV) was incorporated, permitting use of a minimum cushion because with this valve, cushion is never penetrated. The valve also permitted pressure sensing in the mold for precise mold fill, resulting in parts with no distortion and zero scrap.
The following is a typical run sheet from the operation utilizing the DNRV:
Material: ABS Quality limits; standard deviation (3) Metering time 0.022 sec Start injection 0.11 mm Injection time 0.047 sec V/P changeover 0.14 mm Melt cushion 0.14 mm Cavity pressure after changeover 2.4 bar
Although minimal, the variation in cushion (0.14, S.D. 3) can be attributed to the variations in start injection S.D. (0.11) and metering time S.D. (0.022), which total 0.132. The balance of 0.008 could easily be attributed to viscosity variations caused by screw surging, which is directly related to variations in recovery or metering time.
The second application, involving one of the Big Three automakers, molds without cushion. Results in SPC format are as follows:
Material: 6-MF PP; standard deviation (3) Weight: 0.122 lb Skewness: --1.401 Kurtosis: 0.947 CPK: 0.656 CP: 1.045
After installation of the DNRV, the following results were achieved:
Material: 6-MF PP; standard deviation (3) Weight: 0.0067 lb Skewness: --0.2459 Kurtosis: 0.0621 CPK: 6.3681 CP: 6.4611
In this example, results show a vast improvement in weight variation. This was accomplished by installing the DNRV, lowering back pressure to 50 psi, and not using pullback.
New technology requires new thinking. Molders and machinery manufacturers who do not willingly embrace new technology will be bypassed by those who do. New technology must eliminate known problems and offer practical solutions to the problems. The valve technology (patents 5,151,282, Dray, and 5,258,158, Dray) described in this article meet these criteria.
Independent studies conducted by researchers at The University of Massachusetts at Lowell (SPE ANTEC '93 Tech. Papers, 39, 2804), machinery manufacturers, and progressive molders (Nypro) have all achieved the same results with this valve technology. Benefits of the technology include cost reduction through substantially reduced scrap rates, vastly improved quality as a result of preclosing, and the elimination of seat sealing, which permits implementation of meaningful SQC.
Polymer Composites is the engineering and scientific journal serving the fields of reinforced plastics nd polymer composites. PC focuses on the rapidly advancing technical areas of particulate, fibrous, flake, and skeletal composites and laminates. Published bimonthly, Polymer Composites brings to you the details of applied developments of specific interest in the field long before they are commercial realities. Each article is reviewed by a panel of scientists and engineers.
A one-year subscription costs $95 for SPE members, $200 for non-members (add $20 per year for all delivery outside North America). All subscriptions are on a calendar-year basis. Call the SPE Subscriptions Dept. at (203) 755-0471, or write to SPE, 14 Fairfield Drive, Brookfield, CT 06804 U.S.A.
|Printer friendly Cite/link Email Feedback|
|Author:||Dray, Robert F.|
|Date:||Jun 1, 1994|
|Previous Article:||ISO 9000.|
|Next Article:||Advances in fluoroplastics.|