Tech Articles

Does the part dimension really increase with packing pressure?

In my seminars, when I ask the question ‘How can I increase the dimension of a part?’, one of the answers is ‘Increase packing pressure’. It is correct is saying that if the relationship between the packing pressure and a dimension of a part is of direct proportion or in other words, higher the pressure higher is the dimension, lower the pressure, lower is the dimension. But is it technically correct to say, ‘The part dimension increases with increases with packing pressure’? The key words are ‘technically correct’. Let us investigate and understand the basics behind this. Read On

Plastic melt has a higher volume compared to solid unmolten plastic. Consider a mold cavity of dimension A. Refer to the figure. If it is filled with melt it will have less number of molecules compared to if it was solid unmolten plastic. As the melt cools down, it will occupy less space because of the shrinkage resulting in a part with a dimension smaller than the cavity dimension. Let us call value of the value of this dimension as B. During the shrinkage if more plastic is added (packing phase) and the added molecules reduce the inter molecular distance restricting the molecules to shrink to the previous levels. The part will therefore have a larger dimension and will be C which is greater than B.  If the pressure is further increased during the packing even more plastic is added further reducing the inter molecular distance and therefore the shrinkage resulting in D which is greater than C.

So technically speaking, the part dimension did not increase but ‘The part shrinkage was reduced with an increase of packing pressure’.

Back Pressure – The misunderstood molding parameter

When I teach the concept of Back Pressure in my seminars almost always one of the attendees will say ‘Increasing back pressure will increase the melt temperature’. That is absolutely true! However, there are a few things we must understand about back pressure. One should only use the minimum needed and never to fix dimensional problems.

So what is back pressure? As the screw rotates to pick up the plastic for the next shot, the accumulated material pushes the screw back. When the screw reaches the shot size, the screw rotation stops. The volume of the barrel that is filled with the plastic should be equal to the volume of the runners plus the cavities in the mold. (Ignore cushion for now.) Back pressure is the pressure applied to the back of the screw to compress the melt to a consistent volume. Since Part Weight = Melt density X Volume, consistency from shot to shot can only be achieved when the melt density is consistent from shot to shot. Since volume is fixed it plays no role in consistency. The back pressure maintains this shot to shot melt density. Increasing it to more than required will only make the plastic slip on the screw, increasing shear and therefore the melt temperature. A certain percentage of the heat is generated, and must be generated by shear but it should be capped off once the balance between the barrel heats and the shear heats is reached.

Back pressure also serves the purpose of compressing the melt to remove any volatiles that will get into the melt and cause splay. Again, a minimum amount of back pressure must be used to get rid of the splay.

During first shots - Never look at your part dimensions unless …

One of my customers called me since he was having issues with one of their new molds and parts. He had visited the mold maker for first shots that resulted in parts being out of spec. The mold steel dimensions were adjusted, 10 shots that met the specifications were sent to him, he accepted the mold and brought the mold over to his shop in the US. When he began molding parts with a process that the mold maker had given him he was unable to mold parts to spec and meet the required quality requirements. So what happened??

The failure lies in the fact that mold makers try to mold good parts during the first shots and/or make steel adjustments based on non-robust processes. The first step should always be find the a good robust molding process. The mold should first be adjusted for gate sizes, runner sizes, cavity balances, venting, pack pressure process windows, part sticking and so on. Once the mold is acceptable, a process with a wide process window must be defined and then the mold steel dimensions must be adjusted based on a robust process.

I follow a two-stage process that can be defined on this simple mold qualification diagram below. It is a pdf and so click and save it for the future if you like.

(To download the pdf, right click and select ‘Save link as’)

Complaint – DOE Takes too long to perform and measure

I recently did a survey with some of my software customers and ask them their biggest hurdle in performing DOEs. The biggest one was – DOE takes too long and it takes even a longer time to measure all the parts and all the dimensions. What do I start off with my seminar?? – Basics, Basics, Basics … Understand this and the DOE will get very simple, less time consuming!

Why do we do a DOE?

For two reasons: To find the effect of the processing parameter on the part quality and to then determine if a process can be selected such that it is robust, repeatable and reproducible (The 3 Rs). Now let us look at the complaint:

Complaint No.1: DOE takes too long to perform. There are only 11+2 molding parameters on the molding machine that will influence the part quality. Of these there are only 5 that will have the biggest effect on the part dimensions and quality and of these 5 there are mainly only three. I have written a paper on this and will be happy to share this upon request. Please send me an email. So, if you use 3 factors and do all the experiments, you will end up with only 8!  +1 for confirmation and that is only 9 experiments.  On an average size mold (75 to 300 Tons), this should take you  no more than 1.5 hours. Now you have your parts for measurements within couple hours.

Complaint No. 2: The measurements take a long time. First think Cavity balance. If you have got good cavity balance, then that means every cavity is going to act the same way. Therefore, in an eight cavity mold Cavity 1 will respond (not have the same dimension) the same way as Cavity 8. Since you are investigating the effects of the parameters, if the holding pressure will increase the length by 0.010” on Cavity 1 then it must also increase the length on Cavity 8 by 0.010”. It is simple and plain science. This tells us that to study the effects, you can study just one cavity and determine the important factors and the trends. If you have an imbalance, then you will probably need to study more cavities. But remember the worse your cavity balance, the smaller your process window and so that must be fixed sooner or later to get the 3R Process. Next Step: Once you select a process or look at the data for one cavity, then measure all the cavities at any one process or the selected process and record the dimensions. (This is another importance for Cavity Balance).

At what temperature does the viscosity graph need to be performed?

A question that frequently gets asked in my seminar is ‘If there is a recommended temperature range, at what melt temperature does the viscosity graph need to be performed – the high end or the low end or in the center?’

To answer that question let us understand that the reason to perform the viscosity study on a molding machine. It is to find the injection speeds where the viscosity remains fairly consistent and therefore the fill. The actual viscosity number is not of any value at this point. If the temperature increases the viscosity drops and if the temperature decreases the viscosity increases but the profile of the viscosity curve stays the same. Refer to Figure below that shows the viscosity study done on a HDPE material at 175 deg C and 220 deg C. It is clear that the two profiles are identical. The viscosity is higher at 175 deg C and is lower at 220 deg C. The curves run parallel to each other.

Since the reason for developing the viscosity curve is to find the region of the curve that is most consistent, the profile of the curve is all that matters. Since the profiles are identical, the selection of the temperature for the study is not critical.

Understanding Splay – It is not always because of moisture

In Injection Molded parts the most common cosmetic defect is splay. I have been to several organizations and have been shown their defect charts. Splay is always the issue that tops the charts. To solve any problem one needs to understand the source of the problem. So let us understand what is spay to begin with.

Injection molding is the process of injecting molten plastic into a mold. The cavity steel has a desired texture that gets picked up by the molten plastic and is replicated on the part. Depending on the base polymer that is being molded the plastic melt temperature is anywhere between 350 deg F to up to even 750 deg F. At these temperatures water turns to steam and some of the low molecular additives can burn to produce volatiles. The speed of injection of the plastic into the mold will also shear the molecules. Excessive shear can degrade the molecules. Steam and volatiles (from now on collectively grouped as volatiles) from degradation then flow with the plastic into the mold. Because of the fountain flow of the plastic in to the mold cavities the volatiles get to the surface and prevent the molten plastic from coming in contact with the mold steel and at the same time smearing the volatile on the interface of the melt and the mold steel. This shows up a streaks and is called as splay. Splay is also called as Silver streaks.

Anything that can excessively shear the plastic can cause splay. For example, if the gates are too small or have sharp edges the plastic can get excessively sheared and give out volatiles. In that case, mold design is the culprit. Sometimes a worn out screw and/or barrel can also cause excessive shear resulting in splay.

There will always be some volatiles and definitely air that gets trapped in the melt stream. This air is the air that is present between the pellets as the pellets go from the feed section to the compression section of the screw. Back pressure applied during the screw recovery process helps get rid of this air and volatiles. Back pressure should always be kept to a minimum since excessive back pressure can also increase shear and result in splay.

Regrind can also cause splay if the regrind has excessive fines. Fines don’t convey well since they can get stuck to the screw and finally degrade causing volatiles.

The vents in the mold provide an outlet for the volatiles. If the vents are plugged up and/or are not deep enough and/or are not sufficient in number, then the volatiles have nowhere to go resulting in splay.

Following are the solutions to take care of spay. PLEASE DO NOT FOLLOW THEM BLINDLY! Find the source of the problem!

  1. Check moisture in the material
  2. Decrease melt temperature
  3. Decrease injection speed
  4. Increase back pressure
  5. Check to see if regrind has excessive fines
  6. Check and clean the vents in the mold
  7. Check part designs and mold designs to see if there are any areas that can excessively shear the plastic.
  8. This is a living list of issues and are not the only reasons and solutions, add your own, build your own list for your parts and industry.

How many molding machine parameters truly make a difference in the quality of the part?

During my seminars, I ask for a quick answer to a question: ‘How many molding parameters truly make a difference in the quality of the part?’ I get answer that range from 4 all the way up to 30 or 40! What do you think? Here is my take ….

There are 11+2 molding parameters that make a difference in the quality of the part. Why the +2? The first phase is the Injection Phase where the cavity is filled with molten plastic. The second phase is called as the compensation phase where the shrinkage is being compensated. The compensation phase is further divided into the pack and the hold phase. Most processors set only one number for the compensation pressure and time and call this as the holding phase. Not accurate, but it has become acceptable. So if you consider pack and hold as different phases then you end up with 13.

Following are the 13 molding parameters. These are machine inputs. (See outputs below)

  1. Barrel Temperatures
  2. Mold Temperatures
  3. Injection Speed (Velocity)
  4. Injection Pressure
  5. Pack Pressure
  6. Pack Time
  7. Hold Pressure
  8. Hold Time
  9. Screw Rotation Speed
  10. Back Pressure
  11. Cooling Time
  12. Shot Size
  13. Transfer Position

There are a few other parameters that are secondary parameters that do not get adjusted as often as the 13 mentioned in the previous section. They also do not have a substantial effect on the part quality.

  1. Decompression or Suck back position
  2. Pack Speed

Here are the machine outputs that should be monitored. Outputs are more important that the inputs.

  1. Melt Temperature
  2. Actual Mold Temperature
  3. Fill Time
  4. Peak pressure during fill
  5. Pressure at transfer
  6. Cushion Value
  7. Screw Rotation Time
  8. Cycle Time

Please post your comments here

Tonnage Calculations: Why the formula of projected area is not correct and is only an estimate?

I get asked about tonnage calculations all the time. For years molders have been using the formula below for calculating tonnage:

Tonnage Required = Total Projected Area X Tonnage requirement for the given material

The reason this is not accurate is because there are several factors that do not get factored into this formula.

Refer to the figure below and see how the projected area of the part is the same in all cases but the tonnage can vary based on the factors below.

  1. Wall thickness:  Thinner parts need more pressure to fill the cavity whereas thicker parts will require more packing pressure to compensate for the shrinkage. Two parts can have the same projected area but the thicker part will require more tonnage since it needs to get packed out more than the thinner part. However, in a part such as in a laptop cover, a very thin wall with a long flow length will also require more tonnage to withstand the high injection pressures required to fill the part. Thin walls constitute parts as thin as 0.5 mm (0.020 in) and thick walls are those above 7 to 8 mm (about 0.3 in). Nominal walls are usually between 2 and 5 mm (0.080 to 0.200 in) thick.
  2. Number of gates: More the number of gates, easier it is to fill the mold and the less is the pressure required to pack the cavities out. Two parts can have the same projected but the one will more gates will require less tonnage
  3. Position of the gates: If the part edge gated, it will require more tonnage compared to if it is gated in the center. The flow length is cut into half when the part is gated in the center.
  4. Sequential valve gating: Molds that are sequentially gated require less tonnage since the force is being applied only in the areas that are influenced by the open gates.
  5. Orientation of the part in the mold: In the second figure, the same part is shown to have injection points from two different directions. Using the above formula, the required tonnage when the plastic is injected from the side will be lower than when injected from the front. This does not mean that the part can be run on a lower tonnage press. The flow length would then play a role in the tonnage.

Challenges in predicting Tonnage: 
At this time there is no accurate way of calculating the required tonnage. Flow simulation programs have come a long way but too do not give accurate results. The quality of mold building plays a major role in shutting the mold off during injection. This variable is impossible to predict. The role of vents is also a factor. The lesser the number of vents, the slower the air can get out building in pressure in the mold and thereby increasing the tonnage. All simulation programs consider perfect venting.

Screw Rotation Speeds Rule of Thumb - Throw this out of the window

During my last seminar, one of the attendees mentioned that he sets the screw recovery speed such that the screw recovery time is always 2-3 seconds less than the cooling time. He has been molding like this for a number of years and with all due respect to his years of experience I told him that he must throw that rule out of the window.  Here is why…

What is the function of the screw?

  • To convey the material from the base of the feed throat to the front of the barrel.
  • To melt the plastic during this conveying action. This happens because of the heater bands around the barrel and the frictional heat or shear heat from the rotating screw.
  • To achieve a homogeneous mix of molten plastic. This happens because of the dispersive and distributive action of the screw on the melt.
  • To achieve a good consistent melt temperature throughout the mass of the molten melt.
  • To avoid the plastic from burning and degrading during this process
  • To avoid any additives from burning and degrading during this process
  • To avoid any fillers, especially the fibrous fillers such as glass to be broken down during the process.

If the screw rotation speed is high, the material could degrade, the fibers could break and the additives could get destroyed. On the other hand, if the screw rotation speed is low then there is the issue of loss of melt homogeneity and insufficient melting. The rotation of the screw also provides the frictional heat or shear heat that helps the plastic to melt. The crystallites in crystalline materials need a lot of energy to melt  and therefore screw rotation speeds need to be higher for crystalline materials as compared to amorphous materials. All these factors must therefore be considered  when setting the screw rotation speeds and so as you may have figured out by now, screw speed is material specific and not part specific.

See the figure below. As a made up example, let us say a you need to mold a 10 gram business card like part with a wall thickness of 0.100” from a Nylon.  Assume that this nylon needs a screw rotation speed of 50 rpm on a 35 mm screw and at this speed the screw rotation time is 7.5 seconds. Assume that the required cooling time for the part is 10 seconds. Now consider that the wall thickness of another similar part is 0.250” and is being molded from the same Nylon with the same part weight of 10 grams.  (Obviously some of the the part dimensions are smaller but the volume of the material is the same.) Since this wall thickness is more than the previous part, it will need more  cooling  time. Let us say it needs 20 seconds.

If the required cooling time for the thicker part is now 20 seconds, based on the rule of thumb the screw should rotate of about 17 -18 seconds. The screw rotation speed must be therefore slowed down. So instead of the 50 rpm, one will need to set it to may be 20 rpm. This reduction can easily lead to a loss of melt quality and homogeneity leading to several issues and defects in the molded part. The screw speed must therefore be kept at 50 rpm even if there are still 12.5 seconds left in the cooling. If there is a concern of material degrading in front of the screw, then one should add a charge delay time of 10 seconds and then start the screw rotation.

Please email us for more info about how to optimize screw rotation speeds.

7 Top reasons NOT to perform a Viscosity Curve

The data obtained from a Viscosity Study during Mold Qualification is very valuable. The 1st Step in the 6-Step Study (Scientific Molding Studies) is the Viscosity Study. It is also called the In-Mold Rheology Study or simply the Rheology Study. Those new to the area of molding take a seminar, learn this technique and the first study they try to perform is the Viscosity study. They probably end up with a good looking graph but then fail to use it in practice. Or they the graph they generate just does not ‘look good’. They then throw all that they have learnt out of the window and go back to the ‘old school way’. I have seen this happen a number of times. And honestly, they may be justified in doing so!

One must remember that there are always exceptions to these rules/studies.

Following are some of the exceptions where the viscosity study must not be done, or could be done but it is not necessary that the data must be used.

  1. Insert Molded Components
    When inserts are loaded into the mold they are held in the mold only at certain places. The rest  of the insert is unsupported. When the plastic is injected at higher injection speeds the inserts can move out of place and the molded part can then be out of specifications. In some electrical connectors the pin to pin tolerances can be less than 0.001” and in such cases the injection speeds must be set at lower values.
  2. Cosmetic Components
    As the injection speeds are increased the shear at the gate increases. This shear can either cause gate blush, color separation and/or similar cosmetic defects. To avoid these defects the injection speeds must be set at lower values and profiling of injection speeds may be necessary.
  3. When the material is shear sensitive
    In materials like PVC higher injection speeds will cause the plastic to burn since the polymer is shear sensitive. To avoid the burning the injection speeds must be set at lower values.
  4. When the mold is pressure limited
    If the required pressure to fill the mold equals the max available pressure then the process is said to be pressure limited. In such cases, the injection screw will never reach the set injection speed for lack of pressure. Therefore the values of the Fill time will not be accurate and the obtained viscosity graph will be of no value. (Viscosity = Fill Times X Pk Inj Press X IR)
  5. When the % usage of the barrel is less than 20%
    To get accurate Fill Time and Peak Pressure readings the machine must have enough time to respond. Using lower that 20% of the barrel does not allow for the screw to have enough time to consistently build the plastic pressure and/or output the fill time. Remember that plastic melt is compressible and so responses can get inconsistent.
  6. When the parts are small (less than 0.25”)
    This follows the same reason of Bullet No.5. In addition, since the part is the last to fill (after the sprue and runner) even if the runner is large causing the %shot size used to be more than 20% the inconsistency is fill, again because of the compressible melt may not give you a good viscosity graph.
  7. When the machine is not capable to having digital outputs for peak pressure and fill times
    I often get asked this question, ‘We have a old gage machine, what to we do??’ – Sorry, the data is not reliable, forget the study.

So in such cases how does one set the injection speed?
Answer: Find the fastest injection speed that will not give any cosmetic defects. Give a little cushion on the lower side of the speed and set the injection speeds.
Remember: Inject only as fast as you NEED and not as fast as you CAN or the viscosity curve shows you.

Why implementation of Scientific Molding fails for some companies

I have had a number of companies and their employees attend my seminars. There are also a number of people who attend seminars on the same topic conducted by some notable companies and consultants such as RJG, Mr. Bozzeli, Mr. Routsis, Mr. Tobin and so on. When they leave the class they are all excited and fired up about Scientific Molding and DOEs. But then when they go back to their companies and try and implement what they learnt, they do not succeed as much as they should. Sooner or later are back to turning the knobs and dials the way they used to before they attended the seminar. They are frustrated and say ‘Scientifc Molding does not work’. I often feel disappointed about the outcome but have now begun to realize some of the reasons for this failure … (If you have your own reason, please contribute them with a link I have shared below)

  1. There are 5 pillars to a successful molded product. The Part Design, The Plastic Material, The Mold Design and Construction, The Molding Machine Selection and the last one is the Molding Process. If the first 4 are not optimized for molding and holding the tolerances, the molding process is not going to be stable. For example, A thin walled part requires a lot of injection pressure and so needs a higher pressure machine, a small shot being molded in a large barrel will never give you process consistency and/or may degrade the plastic, Bad mold shut off can result in shorts, sinks and flash at the same time and with very low statistical capability (Cpks), molding a large tight tolerance part with an olefin, and so on. In such cases, no matter what you do achieving consistency is almost impossible. Performing Scientific Molding is futile.
  2. Even if you have the right combination of the above 5 pillars, there are still exceptions to every one of the Scientific Molding rules. For example,
    1. for an insert molded part performing a viscosity curve study (Rheology), is not required since faster speeds will move the inserts out of place. One must inject as fast as REQUIRED and not as fast as you can or on the flat portion of the viscosity curve.
    2. A gate seal study cannot be performed on hot runner molds (I have an procedure in for this, please email me.)
    3. Cavity balance is important but some get stuck on the actual % imbalance number rather than the significance of the imbalance. I have written about this on one of my earlier posts down below.
  3. ‘Why change it if it is molding good parts’ syndrome. We accept the fact that the ‘good parts’ are the best parts and with the most efficient process we can mold. I have been to multiple companies and reduced cycle times by as much as 50% on 40 – 70 second cycle and reduced the scrap rates to almost nothing.  In fact, in my seminar I tell the attendees not to experiment with new mold but to start the implementation process by learning on an existing production mold.
  4. Management support – This is very important but remember that the management is looking at the dollar figures and so you have to prove to them on at least a couple projects that these systems work. Only then will they support you in the future. They are investors and so they need a money back guarantee.
  5. Remember that Scientific Molding is not the act of following procedures. Scientifc Molding is understanding the science behind molding and implementing the priciples to achive process consistency and efficieny.

Please do share your thoughts on my free online forum here.

Setting Acceptable Machine Tolerances during Production

My friend Fernando Quintero (who also happens to be my customer taking one of my seminars) asked me ‘What are acceptable tolerances I should set on the machines for my molding process during production?’ Again a very common question and here is my answer to this –

First let us begin with what most companies have. Most companies have a ‘generic’ tolerance for each of their process parameters. For example, Melt temperature: +/- 10% or Cushion Value +/- 0.05 inches. The problem in using this ‘generic’ method to set the tolerances is twofold.

  • Using a % is always misleading. If the melt temperature is 350 deg F, then a +/- 10% is +/-35 deg or a range of 70 deg. If the melt temp is 650 deg F, as is in case of a PEI then a +/- 10% is +/-65 deg or a range of 130 deg F. So higher the value, wider is the tolerance range; clearly does not make sense.
  • Process Tolerances are applied to keep the product within the product quality requirements or within the LSL and the USL. If one puts a generic tolerance of +/-10%, then one is assuming that if the parts are molded at the lower end (-10%) and the parts are molded at the high end (+10%) then both these sets of parts are acceptable. This may or may not be the case and so again, this clearly does not make sense.

So how should one set tolerances of the machines?

  • To begin with, first note the dimensional tolerances of the part.
  • Look at the DOE results and pick the top two process contributors for the part.
  • Plot the contour plot using these two process parameters.
  • Draw the Dimensional Process Window and look the limits where you can mold parts within the acceptable limits. (See Figure)
  • Apply these as the process tolerances

Note:
The above procedure is for one dimension and one cavity. One must consider all the cavities and all the dimensions. This can be a challenging task especially if the mold cavities are not balanced and/or the Cosmetic Process Window is narrow. Most DOE programs will predict an optimum process (or Gold Spot as it is called in the FIMMTECH Nautilus Software). One can look at that process and using the software simulate the process and then set the tolerances.    In case of tight tolerance parts, these process tolerances will be much narrower compared to commodity molded products such as buckets, pails, forks and spoons (where dimensions requirements are almost non-existent).

Please post any comments on the free molding forum here.

Importance of adding Charge Delay Time when determining the ‘fill only’ (95 – 98%) phase

I learnt this the hard way a few years ago. I was working with a cold runner mold trying to mold PP. I wanted to find the 95-98% full part with zero pack & hold pressure and time. I do not remember the exact numbers but lets us make some up. Shot Size = 10 inches, Transfer Position = 2 inches. Pack and Hold Pressure = 0 psi. Pack & Hold Time = 0 seconds. The part was full. So I changed the transfer position to 3 inches. The part was still full. So I changed it to 4 – no change. So I changed to 5, 6, 7 and still NO CHANGE !!! It was driving me crazy when I suddenly realized ….. The gate was not frozen at the end of injection. There was no pack & hold phase and so the screw began to rotate and recover immediately. The back pressure that is applied during screw recovery was pushing the plastic into the cavity and filling it. With a transfer position of 7, I added a charge delay time of 5 seconds and the next shot was only about 20% filled!. I left the charge delay time at 5 seconds and then I experimented with the next few shots to find that the transfer position of 3.5 inches got me a part that was about 95% full.

It is very important that a molder knows how much of the part is being filled in the injection phase. This is also called the ‘Fill only’ part and is a very critical piece of info to evaluate processes and to transfer processes from one press to another.

Questions or Comments? Post them here.

Injection Phase - 95-98% by Weight or Volume?

In my last seminar I was again asked another common question. The same question has been discussed a few times in the past on my forum.

Question: If one uses the Decoupled Molding technique (service mark of RJG Inc), should you fill the mold 95-98% full by volume or by weight?

Here is my take on this. To begin with let me give you my theory on filling the part less than full. As the plastic melt cools down it shrinks. As it shrinks the volume reduces. Molten plastic is injected into a cold mold where the melt begins to cool down as soon as it touches the mold. Therefore we need to get the plastic inside the cavity as soon as possible. This is the injection phase. In the injection phase, we fill the mold cavity with molten plastic. But if we stop there, the plastic will shrink and the part that is ejected out of the mold will have sink. To compensate for the shrinkage and eliminate the sink, the second stage called the pack stage is applied. During this stage, the plastic must enter the cavity at the rate of the shrinkage that is occurring. The third stage is the hold stage that is applied such that more plastic does not get into the cavity nor does the plastic that is under tremendous pressure get out of the cavity. It is applied till the time the gate freezes off.  So in summary, we fill the cavity, compensate for the shrink and hold till gate freeze.

When we say ‘fill the cavity’ – in theory we need to fill it up a 100% with molten plastic … and that is where the problem arises. Plastic melt is highly compressible like a ball of rubber bands. So we do not know if the volume of plastic that was injected into the mold was equal to a 100% of the volume of the cavity or was it more than a 100%. It could have very well been more that a 100% since the melt could have been compressed. Compressed melt is not desirable since it can lead to issues such as stress and flash in the parts. For that reason, the mold must be filled just a little less than a 100% and then the pack and hold should be turned on. That is where the number of ‘less that 100%’ comes from. An added benefit of this is also to slow down the injection speed before hitting the point of transfer into pack and hold to achieve consistency of the point of transfer. It is like slowing down at the traffic light before coming to the light and not slamming on the brakes at the zebra crossing.

So does it then matter if we are going to transfer by weight or volume? To begin with, Weight = Volume X density. Since the melt density is going to be constant, 95% by weight = 95% by volume! I think the question is answered. It does not matter. This can get confusing in some situations:

  • On thick parts you may fill the skin first and then the inside will get packed out. So there is not visual sign of the less that 100% in injection.
  • On thick parts once the skin is fully filled, the vents are closed and there may still be air inside the cavity. So you may need to inject slowly to vent out the air and then pack and hold. In this case a ‘visual short’ may be the way to go. This may appear to be 98% short by volume but may be only 90% by weight.
  • On thin parts, since there is not much shrinkage a 98% by volume will be very close to 98% by weight.

Comments on the % number: So should it be 95% or 98% or why can it not be 90%? The answer is that there is no definite number that must be followed. Here are some general guidelines:

  1. The thinner the part, try to use closer to 100%. In some cases, where there are very thin sections such as filter mold I was working on, I had to get 100% of the material in there in the injection phase with no pack and hold pressure but with some hold time to seal off the gate.<

The difference between dimensions being In-Spec and being Capable

We processors are not Statisticians. When I started my second job as a process engineer, QC measured the dimensions of the parts from a mold I had sampled. The dimensions were within the required specifications between the Lower Spec Limit (LSL) and the Upper Spec Limit (USL). I thought I was done! But then Terry Hertz, the quality engineer came to me and said ‘Sorry Mr. New Guy, the dimensions are in spec, but they are not capable. You need to make them capable!’ I was taken aback since I had no idea what he was talking about. I asked him to explain, and he did. That was the day when I started to understand quality systems and developed an interest in statistics. So let us understand the difference between being in spec and being capable.

Variation is natural. If you say the time it takes you to get to work is 30 minutes, then that 30 minute number is the average of the times it took you get to work in the last few months or years. There have been days you got to work in 29 or 28 or even 25 mins on the low side or may be as high as 35 mins on the high side. This is natural variation that we cannot control. For example, traffic, traffic lights, road conditions, weather and so on. If you weigh 100 consecutive shots from a molding machine, you will see that there is a variation in the weights even though you have not touched the molding process.

If you extend this concept of variation to part dimensions you will see the same result. If you measure 100 consecutive parts for a given dimension you will see that there is a variation on the part dimensions. Refer to Fig 1. Most of the dimension will be centered around the average but there will be dimensions at the low and the high end. If you measure three parts from the green area of the graph, you would think all the parts are good and so the entire production run is good. However, there will be parts out of specifications that are in the red area of the graph. Unfortunately, these will be the one that your customer will find!

To take into consideration this variation, statisticians perform certain calculations called as Process Capability calculations and predict if this variation can be fully contained between the LSL and the USL. If it can be contained, they say that the process is Capable. In short, the process is always capable of molding all the parts to the required specifications. The terms Cp, CpK, Pp and PpK are used for quantifying this. We will discuss these in the next article.