Die Casting Defects 

Defect Definition

The defect definition is simple: “there is not enough sound casting material in a certain location.” Yes, it is that simple. Take a cracked casting - not enough sound material in the cracked area; cold runs – missing sound material which merges / weld together during cavity filling; shrinkage porosity – there is a hole in the casting due to missing sound material; air entrapment –sound material is replaced by air. If you need more examples feel free and use a web browser of your choice or search for “Die Casting Defects - Causes and Solutions. Multiple books and training classes are offered by many companies, organizations, or schools around the world.

Why is the Defect still out there?

Thousands of die casters fight defect castings on a daily basis. They work on the same issues over and over and it seems they never get tired of it. What makes a high pressure die casting or the casting process so special that a failure free casting is seen as often as being hit by lightning?

The answer to the question is not simple. Many castings are driven by design and function but not by casting / production requirements. However, casting development does not have to exclude anything if considered from the start of early development. Many times, a die caster has to give a quote for a part that has a fixed design, which will create “challenges” during production. Furthermore, companies take on castings they do not have enough experiences with or projects that do not fit into their production line.

The casting defect is created by the casting developer / designer, the die caster and the tool shop, and (do not let us forget) the casting buyer.

Why the buyer, you ask?

The casting buyer introduces the new part to the die caster as soon as he has information to share and he does not make any engineering changes. Without adjustments, the die casting will not be successful. By selecting the casting supplier close to the final due date, the casting engineers do not have sufficient time left to work with the development department to optimize the casting design. The semi-optimal design is released and production has to live with it. In effect, the final decision who gets to produce the casting is made far too late. The period between the first quote and final decision is typically more than 1 year, but the die itself has to be completed in less than 12 weeks, which reduces the process engineering duration to less than 4 weeks. Now, the engineering must be completed in a very short time with a reasonable result, but, without cooperating with the casting development department, 6 to 9 months is wasted. Time would be better spent focusing on building a prototype together.

Defect Avoidance

The strongest tool we have for disposal is fluid flow, solidification, and stress simulation. The high end software programs provide all the information to detect the die casting defect and point the engineers to the root causes. Don't let the software fool you by performing a simulation overnight. By doing so, you easy find some defects like shrinkage, air entrapment or hot spots. Varying some parameters to find out how much influence parameter changes create makes the engineering process much longer.

If you are interested in relations between melt pouring temperature and die temperature or a possible correlation between process cycle time, lubrication time, and die cooling layout, many more simulation runs have to be executed.

The time for result preparation and discussion should not be underestimated. What value does information have when it is not shared? It takes time to work through all the material, prepare a presentation, get all involved team members together, explain the results, discuss them, and agree on the next steps.

For some projects, a simple fill and solidification simulation may be enough. A complicated casting may need fill approaches from different gate locations, multiple cooling / heating line designs, lubrication pattern variations, residual stress and deformation simulations including the casting heat treatment. Doing this a couple of times for parameter optimization and 4 weeks is nothing. Yes, not all of this is required for each new casting design or process, but gathering more information in the beginning creates a more stable production process.

If enough time and engineering effort provided the die casting does not show any kind of defect, we have not accepted during casting and process development. This defect will be seen always (long term). An acceptance standard has to be prepared and released with the casting.

And now let's see what the production process has for us

Once the Process FMEA is written and the die is ready for production in the machine. This includes an adequate venting / vacuum system which keeps the maximum efficiency until end of cavity fill. The 3D printed water flow system cools just the needed areas without extended cooling lines never properly fitted into the die steel. The die thermal control system is connected to balance the water volume for the optimal cooling rate. Temperatures are measured using an infrared camera system. The cooling / heating system as well as spray times are regulated to keep the die in the required temperature field.

As some experts are telling you, “when the temperature pattern is the same and the shot parameter are the same, you will have the same casting quality all the time. If the casting quality changes just bring the temperature or shot parameter back where they belong to.”

Production starts and the process is monitored with frequently visual and x-ray controls. From this point on die casting is a piece of cake! No, more defects! The die stays in the machine for the next hundred thousand shots until it is replaced by another one. Congratulations.

But wait, you work in high pressure die casting. Here, the world is not that easy. Yes, the majority of processes run effortless. Just the other ones keep your full attention. Even the good running processes have a scrap rate after machining of up to 3%. Far too much in the parts per million failure rate your customer thinks. He is right. Nobody can explain why we produce 3% scrap castings when we can make 97% good ones. Where are the 3 bad ones coming from? How come we have the same shot parameter and thermal image but different casting quality just by changing the die casting machine or die? The answer is pretty simple most of the time: we do not know.

What else influences the casting quality?

At first, we did not engineer the process as good as we should (see above). For example, the production die casting machine is not as strong as anticipated for simulation input, or the possible cycle times may not be as fast as recommended.

Second, we still have some parameter in the process left that we are not able to control as well as we want. Metal pouring temperatures vary more than expected, or the spray nozzles for die lubrication lose their position to an uncontrolled spray pattern. This may not be noticed as an effect in the temperature profile but will influence soldering on the die surface.

Warming up the die too quick or under an insufficient temperature profile can lead to flashing and fast die wearing, which reconditions our previous process parameter that was developed.

Doing some minor adjustments to the parameter set found for a new die, may require major adjustments for a die which is in production for a long time, showing wearing on the surfaces or scaling in the cooling channels.

Some plants have quality issues due to different cooling water temperatures (colder in winter times and warmer in the summer, warmer when all machines are in production, etc.), or the melt transfer between furnace and the shot sleeve can be significant unstable. Having an open furnace and a ladle system needs more tolerances for volumes and temperatures as an automatic dosing system, which is even easier to monitor and control.

The casting companies know their short comings best and how tough monitoring the process is when full automatic control units are not available. This high end unit needs proper maintenance and trained engineers to evaluate the measured data. A sufficient Quality Control System has to be in place. Collecting data at the die casting machine may be helpful for the operator but is by far not enough to control and improve the casting process.

Measuring at the machine, but not alarming anybody, when the parameters are out of tolerance does not help at all. The data have to be collected and looked through on a regular basis to find the correlation to the castings made. Experiencing a high number of failed castings due to a defect brings all attention and responsible personal to that die casting operation which takes the people away from other important tasks. This trouble shooting has to come to an end by proper process monitoring and control. If a quality system is not in place or not strong enough to find parameter at the die casting machine which are out of tolerances, the casting defect will not be eliminated.


Every die caster has its own position with strength and weakness. Some are specialized in long, automatic production runs and able to invest in many support systems, others work in very short manual production runs that do not need the most utilities recommended for automatic production. Between these two extremes, the other plants work with their unique equipment in their own niche.

Long story short, it is important to have the engineering experts of the die casting plant at the table at early process development. They know the situation at their shop floor the best.

Having this insider information is tremendously relevant before starting the casting / process development. Doing it right, together, and starting early enough will ensure success is reached and Die Casting Defects are not an issue.


Sampling Dies is Wasting Money 

A die or a mold is a two component steel tool, one is the cover half, the other is the moving half. These parts form a cavity when locked together, and melt can be poured into the cavity. After solidification, the die halves move apart and a casting is ejected. To produce complex castings, slides can be added. For faster solidification and a constant die temperature, cooling channels are drilled into the die.
In the past, experienced toolmakers designed the die components, drew them on paper, machined the steel, and fit the components together. This craftsmanship was costly, time-consuming, and required years of training. Due to machining variations, tolerances were wide and each die was unique. Changing components from one die to another was almost impossible.
Today, dies are designed on computers, based on native casting shapes. Computer simulations are used to optimize fluid and heat flow in the mold. Machining steps are programmed and transferred to a multi axle machining center which cuts autonomous die components with precision and consistent accuracy. Today, dies are identical, and single components can be changed from one mold to another without restrictions.

Die Sampling

In the past, die sampling was a necessity. Due to the die complexity, certain requirements like wall stock could not be determined without pouring melt into the die and measuring the dimensions on the real casting. With manual craftsmanship came variations and differences, so each die had to be sampled and measured.
Today, dies are as equal as it gets. Toolmakers take pride in the fact that all components are interchangeable with each other. It is understandable that process engineers want be sure to have the right casting quality and sample the first die. However, not running a die under production conditions will not accurately reflect true casting quality, because production dies get significantly hotter than dies under sampling conditions. Steel expansion is based on temperature, which is the main reason dies do not last and dimensions are out of control, which is not taken into account. Sampling replacement dies does not make any sense when these dies come from the same state-of-the-art tool shop, and were made on the same machines. Sampling replacement dies just costs a lot of money, time, and resources.

Die casters who don
t agree should ask themselves how many replacement dies they have sent back to the vendor after sampling? Why do die casters sample dies in the first place? Because its procedure! It is required by OEMs based on a 50 year old procedure that nobody is questioning. As if nothing has changed in mold building since. Does the finance officer know why they pay all this money for sampling, or how die sampling and process development works? Lets review the current process.

A. Build a die into the die casting machine and pour less than 50 castings for dimensional control only. Now, the measurements are taken from a cold die, which is a condition that does not match what occurs in production.

B. The mold sits in the die casting machine, cooling lines are attached and up to 300 castings are made. Shot parameters are not defined, spray head and program is not set up right, the cycle times are far from target and the trim tool is not implemented into the process.  At this point, the die caster can get an impression about the die conditions and the process he can expect, however, this insight is still not enough to make comprehensive improvements.

C. The die sits in the original machine, cooling lines are attached, spray and shot parameters can be adjusted and cycle times are defined. With a robot, castings are extracted, quenched, and trimmed. Production should have at least 3 shifts to work on the process. This is the first time flaws in the process and its components can be defined, and then improved. 

D. Run production and improve the process if necessary. 

So let’s ask ourselves: Why should it be necessary to sample a replacement die which is close to the first die anyway? What differences are to be expected? Where are the risks putting replacement dies into the machine and making castings? Of course, if there are concerns about quality, you can sample a batch of 300 castings which are available already.Ultimately, production on original equipment and processes will give more information about the quality to expect, than any sampling can do.

I strongly recommend the consideration of the following questions:
Do we sample again if a core breaks and has to be replaced, or after changing a slide, an insert, or a retainer. When it comes to sampling, how do we know when enough is enough? And for comparison, why is the sampling process so vigorously required on replacement dies and not on dies in production that have been in the machine for some time, deformed, welded, heat cracked, repaired and far from the original condition?

I believe there is a large opportunity to save money by not requiring the sampling of replacement dies. So when do we start?


Thermal Diffusion Surface Transformation Increases Die Life 

What is Die Life?

Dies are made of hot work tool steel which was melted, remelted under vacuum, forged, cut into size, machined to the required shape and finally heat treated. Even with the best efforts put in to increase the steel hardness, strength, toughness, ductility and thermal conductivity die tools can start wearing off after 10,000- 20,000 castings made. Imaging now how much money to be saved by finding a way of increasing the die life to hundred thousand or more castings.

Why Does Steel Wear Off?

The best steel production process can't avoid internal microscopic cracks, impurities or distortions in the grain structure or vacancies in the crystalline. During casting production the steel surface gets heated up to 600C (1200F) over some seconds and cooled down back to room temperature to get heated up again right after. With the highest temperature changes right at the die surface and much less some millimeters into the steel stress creates due to thermal expanding and contraction. Over time the stress builds up, the steel starts cracking at the impurities and microscopic cracks open. With cracks increasing the poured liquid metal will be squeezed into the cracks and worsens the already high stress level. Cracks grow further. Depending on the casting surface requirements the piece of die steel has to be replaced sooner than later.

How Does Thermal Diffusion Works?

Compared to Nitriding which creates a surface layer onto the die steel, the Thermal Diffusion Surface Transformation process instills a highly complex chemistry to the surface of the steel which will resist many types of stresses. Diffusion takes place in the first 0.35mm (0.015 inches) below the die surface. By not changing the die component dimensions or the surface finish micro cracks close and toughness adds by increased plasticity and tensile strength. A positive side effect is the increase in wear resistance or in other words minimized erosion to the liquid melt.

Soldering in High Pressure Die Casting

As much as soldering is needed in multiple industries it is not wanted in a die casting die. Due to high affinity of aluminum to iron there will be a chemical reaction, the aluminum will stick on the die steel surface and has to be removed mechanical which decreases die life.
Thermal Diffusion Surface Transformation is not only improving the die steel properties as toughness, strength and wear resistance the used chemistry prevents free iron reaction with aluminum and protects against soldering.

Die Wearing Prevention

Dies showing wearing can be repaired using Thermal Diffusion Surface Transformation the better way would be to prevent cracking and soldering at the first time. Fluid and heat flow simulation software are very handy to help. Simulation results will point to locations with high die wearing using melt velocity, temperature, time and residual stress results. Areas prone to these defects should be treated before starting casting production and be maintained frequently during production breaks to get the most out of the die.


How to Compare Fluid and Heat Flow Simulation Software  

Wouldn’t it be great to have a simple model you know already how the filling and solidification pattern should be? A model to use that everybody understands and immediately shows the simulation rights and wrongs? Fortunately with simulation we can do more as in the real world and we have an object to use.  Image a globe with 12 inch (300mm) diameter and a shell thickness of about a 1/8th of an inch (3mm).  Connecting a 6mm diameter cylindrical straight runner to the shell the filling conditions following the shell must be the same in all directions. The melt front should move in a circular pattern and merge at the opposite pole. If multiple streams or voids appear or the melt does not merge opposite the runner the filling simulation failed. The most likely reasons would be a mesh inconsistency or a flaw in the mathematical model used which can be detected easy on a globular shape. A minimum of 10 mesh layers should be in the shell to get a real 3 dimensional result, an understanding about the real simulation time and possible variations in the filling pattern. Multiple mesh layers (proposal of 1, 5, 10, 15 layers) should be simulated to find and understand mesh influences.

This simulation set up is the easiest thinkable. Do not hesitate and request a simulation set up and calculation in your office to learn how it is done. Use a computer similar to yours including explanation about the mesh creation and result quality.  You don’t see what you expect or don’t like what you see move on to the next software until you are satisfied.


Process Parameter 

“when the plunger moves we can’t do anything anymore”, “during fast shot we can accelerate and decelerate the plunger up to 5 times”, “when we have good castings we make a temperature image and measure the plunger speeds. If we get quality issues we just go back what we documented to make good castings again”. Three statements I heard from a die caster, a die casting machine vendor and a professor. What did I learned from it? There is a kernel of truth in all statements.

The first scenario is what we see the most in the industry. The shot is set and checked with monitoring systems. Casting poor quality the set up will be changed, sometimes the parameter are stored and used for future production runs.

Parameter changing during fast shot is not used at all. The closes to this doing is the plunger deceleration at end of fast shot to reduce plunger impact load at end of fill. There are several reasons this opportunity is not used in further details. Velocity reduction reduces the force behind the melt and fronts would not merge creating cold runs. As well it increases the fill time resulting in colder melt and cold runs. Variations in the poured melt volume create deviations between switch distances and melt fronts. In other words we have an instable process.  

So and why is the third approach not working?  Checking plunger speeds is an easy task and done in many facilities today. Measuring the temperature is much more complicated. The camera is very expensive and die caster shy away based on the costs. The camera is big and therefor can’t be directed to capture all areas of interest. Depending on the angle the measured values will change. To make images the production has to be stopped and the die cools during capturing the temperature profile. As longer the ways around the machine and as more complicate the camera position as further away is the measured temperature from the real temperature.  Adjust the die temperatures with water or oil intakes through the channels in the die maybe not possible. Cooling a new die is different to cool a die with scaling in the water channels.  And on top there is more than temperature and plunger switches the professor told me. Like clocking of wind vents, changing in clearances between components in a new and an old die. Not to mention to keep the same plunger stroke but have a plunger with a different diameter. Or plain the die inserts at their surfaces for spotting purpose and reduce the gate area at the same time. Now the casting quality and shot parameter change even with the same plunger strokes and velocities.

As usual there is not a single solution for the right process parameter set for the complete life of the die. With deterioration of the die process parameter may have to be changed. As more documentation is done and as more information the die caster has at hand as easier it will be to find the optimized machine set up.


Quality Inspection 

First casting inspection is done by just looking at the casting. When the result is promising, the casting will be cut with a saw to find inside porosities. Cutting castings for quality inspection is common and used in every shop and there is nothing wrong with as long it is not the only device to judge what is inside the casting. Modern x-ray systems should be the number one equipment for quality control. These systems allow the detection of internal porosity, tears and inclusions without destroying the casting. Full automated systems need a minimum of man-power and can be used for 100% real time production inspection with immediate feedback to production. Images can be made and stored.  Computer tomography can be integrated and the size of porosity measured and documented for later statistical process controls. Cutting x-rayed castings in defined sections for visual checks and training for engineers and operator personal is fine and recommended. Running production based only on cutting castings is far away of state of the art and increased customer satisfaction. Porosity shift from shot to shot and cutting castings is like playing roulette. Porosity can be in the casting, but the cut passes it and quality inspection releases a process making bad castings. Similar results are given when the porosity is just touched by the blade and only a very small area of the porosity appears.  Cutting slightly aside a huge void could be visible and the process will be stopped. More samples have to be cut with similar results. Two more disadvantages are automated inspection is not possible but worst of all good castings are destroyed during inspection, too. And the numbers add up. Making 200 shots a shift and inspect twice a sample size of 2 castings are 4 castings or 2% production loss.  Not to mention the 100% inspection automotive customers requires. 


The Only ONE Quality Issue in High Pressure Die Casting 

The only quality issue we experience in high pressure die casting is missing metal in the produced casting. Unfortunately the reasons are multiple. Metal can be too cold creating cold runs and overlaps. Melt fronts end aside each other but do not weld together. Between these fronts air and dirt will be found. Hot spots in the die or overheated melt create sinkholes. Fill lines will be seen when too much grease or lubrication is on the casting surface. Melt shrinkage and insufficient feeding leaves holes in the bigger casting areas. Air, water, grease or lubricant mixed with the melt creates porosity. Trimming runner and overflow systems could cut or break into the casting. Even contraction during cooling leaves its missing mark by moving material out of areas and results in rough marks after machining. And to lift the challenge up these quality issues appear when they wish, that makes it much easier said than done to assign a root cause for the poor quality. Finding the trim cut not working properly may be one of the easy ones, but finding the reason for elevated numbers of entrapped air or miss-fill can be hard work. Sometimes issues can overlap and cause harder to find results. Like situations when melt shrinkage and entrapped air are in the same locations. Reducing the volume of entrapped air may not show an effect on the porosity as well as working on the solidification pattern only.

The best is to avoid casting problems in the first place and don’t let them appear at all. Second best is finding the root causes very fast if something goes wrong.  One way to do this is bringing as much experience together as possible. A development team of the die caster’s specialists, the casting customer and tool maker engineers should work on the casting design and die/process layout prior casting design release. At this stage the development system has the most flexibility. The casting can still be optimized, the die designed to an optimal filling pattern and heat flow through the die. Necessary process items like the use of vacuum or special die steel can be selected. As more the process progresses as more restrictions are given and changes become more complicated and expensive. Die layout ideas can be simulated, optimized and compared before die steel gets cut. A toolmaker with simulation software in house should be able to present results in no time. Main design flaws will be avoided and equipment, tool and process costs can be estimated more precise.  Now is the moment to know about the need of a vacuum system? What is the risk running a die without proper cooling line design? Is a die heating necessary or is water cooling enough? What is the process cycle time to achieve? Faster production should make more money, but should not elevate the level of break downs or maintenance. Designing the most stable process with the biggest range of process parameter will create the good castings the industry expects.


Do I need Vacuum to Produce my Casting? 

Entrapped air in castings does not leave room for metal to fill that volume and causes porosity. Porosity created by air appears in a more global shape with smooth surface and stays local. Compare to shrinkage porosity which has a rough surface and an undefined shape. Shrinkage can appear throughout the casting creating leakages. Air porosity mainly reduces weight and saves metal but does not create too much trouble. Though it looks bad and is not accepted in the most cases. If air can’t be pushed out of the cavity by the melt front the air has to be reduced before die filling. A nice side effect is reduced plunger pressure and a faster fill time.  Melt velocities can elevate local in the cavity over the critical speed and increases die wearing. Higher velocities are minor issues and can be fixed easy.  More “tricky” is sealing the die to create and hold the vacuum. As more slides and as lower the level of air as more engineering has to be allocated to it. An experienced tool maker will be a good help.  Process simulations help finding locations of last fill where the vacuum should be pulled off. All vacuum systems on the market start pulling when the pouring hole is closed by the plunger. Main difference in the vacuum systems provided is at the point of closing the vents. Every vendor is pulling to the required low air pressure level before starting the fast shot. One approach stops pulling the vacuum before starting the fast shot, another is closing vents using the melt stream at end of fill.  A third design does not use vents at all. Vacuum is pulling through wind vents which stay open until the melt solidifies in chill blocks and closes the vacuum channel. Atmospheric pressure at 1013mbar defines air pressure, below as vacuum or better under pressure. Low vacuum is in a range of 500 to 800mbars, super vacuums come down to 50 – 100mbars rest pressure.  Typical areas to use vacuum are thin wall castings to support filling, castings to be welded or powder coated and all other casting where critical air porosity can’t be reduced without it.