Process control tips for a phenolic urethane nobake foundry.

Recognizing these 12 facts and practicing the 15 tips that follow can help you make the PUNB process work - as intended - for your foundry.

"There is no substitute for sound foundry practice" is one of the tenets listed in the preface of the AFS Chemically Bonded Cores and Molds Handbook. Foundrymen who've made chemically bonded mold and core processes work to their ability realize that such wisdom is one of the truths for optimizing the phenolic urethane nobake (PUNB) process.

Success with the PUNB process boils down to two words - process control. With a variety of inherent variables such as sand, binder, additives, mixing efficiency, compaction, and sand and binder temperature, a lack of control will slow down your foundry's overall effectiveness.

Above all else, remember that success starts with enlightened foundry engineering. The design of gating and risering and the construction of the tooling have a profound impact on the effectiveness of any nobake operation. Most casting quality problems are solved by engineering changes. Further, the most critical step, and often the least consistent, is pouring.

Following is a short list of facts and recommendations for improving your foundry's nobake process. For additional details, consult AFS Transaction Paper 97-151.

SAND

Fact 1. No New Sand Sample Taken at the Foundry Will Ever be Truly Representative

* Tip - Require that the supplier furnish quarterly summaries of the screen tests (mean and standard deviation on each screen) and then use them to establish the purchase specification (mean +/- 3 sigma). If you must sample within the foundry, it is best to "cut" a moving sand stream periodically and combine these samples, and then reduce the sample's size by using a sand splitter.

Fact 2. Fines Content is the Most Important Sand Test Result

From a control-charting standpoint, a percent fines test will probably suffice. Fines should be less than 1%, since their presence increases binder needs, reduces sand permeability and contributes to expansion defects. AFS Clay isn't a precise test and is "messy" to run. The surface tension of urethane coated sands is different than new sand, which inhibits its use on reclaimed sand.

Fact 3. Grain Fineness Number (GFN) is Meaningless

GFN is the weighted average diameter of the sand grains in the sample, expressed in mesh. It has no place in foundry process control and should never be "charted." It should only be used to compare the relative sizes of different sands.

* Tip - A good way to look at sieve test analysis data is to calculate the surface area of the sand in the sample (usually in [cm.sup.2] per gram.) This can be done by assuming that the sand grains are spherical and that the average diameter is equal to the screen opening through which it passes. Table 1 lists a sample calculation method.

Most suppliers will furnish the calculated surface area on each shipping report. It shouldn't vary more than 5%, and less is better. For a typical molding sand (52 GFN with a 4-screen distribution and 0.3% fines content) run through a 500 lb/min mixer, the surface area of the sand to be coated is about 1 acre a min, or 753 sq ft/sec. That is like a 15 x 50-ft panel, or coating the surface of a billboard in 1 sec.

RECLAIMED SAND

Fact 4. You are Your Own Customer

Once a foundry begins to reclaim its sand, it becomes its own sand supplier, and must accept responsibility for incoming quality control.

* Tip - A sampling program must be developed in which samples are taken consistently during the operation of the sand reclamation system. Design access into the sand system for taking these samples. Sand samples must be transferred into carefully marked, clean containers that can be tightly sealed. Sealable plastic bags work well - keep a supply at the mixer and the reclaimer. Sand samples should be taken from an entire cross section of a moving stream at the system's final discharge, and must be reduced using sample splitters ("rifflers").

* Tip - A useful quality control test is to check bulk density by weighing a standard volume (usually a heavy walled "dry" quart container = 101 [cm.sup.3]). A bulk density test (usually expressed in centigrams per [cm.sup.3]) will serve to alert if there are any significant changes in size distribution, sand grain shape or a mineral contamination.

* Tip - Examine your sands. Use a low power microscope (20-50 magnifications) and take several samples daily (including ones from the mixer). A single observer will give the most meaningful results, and he/she should note anything "out of the ordinary" (contaminants, discolorations or irregularities in grain shapes), which should trigger further investigations and/or testing. Low power (about 50 diameter) SEM photomicrographs can be valuable and should be obtained for any important final samples. Higher magnification SEM photomicrographs are difficult to interpret.

* Tip - The percent magnetics (weight loss due to dragging a strong magnet through a sample) can be important - especially if penetration is the problem.

BINDER

Fact 5. Water is Always Unwanted

Watch out for moisture-condensation (usually a seasonal problem). Water kills the polymerization by reacting with the part II isocyanate and results in "weak" sand.

Fact 6. Rely on Supplier Testing

The testing required to characterize the binder components is beyond the capability of foundry laboratories, and is best left to the supplier.

* Tip - Always keep a supply of chemical mailers (furnished by the chemical supplier and approved by common shipping agents) for submitting samples of questionable binders back to the supplier's lab. Once a year, incoming material should be compared with the test certifications.

* Tip - When in doubt about a binder's ability to work, use the gel test. It's difficult to run accurately (it's very temperature dependent and is best run in a water bath). Caution is needed, as it's very exothermic. Generally, it isn't a good quality control test - but at least it lets you know that the binder will set.

Fact 7. Free Formaldehyde Content is Usually Related to Odor Problems

The odor associated with PUNB, and especially eye irritation, is usually a result of formaldehyde emissions at very low concentrations. The newer "low odor" formulations either chemically tie up the free formaldehyde in the phenolic binder, or the chemicals are processed differently in the kettle to result in a lower free formaldehyde level (see sidebar). Solvents used can also affect the odors at mixing.

Table 1. Surface Area Factors for Sands (AFS Mold & Core Test
Handbook)

USA Sieve Sizing       Silica      Divine      Chromite      Zircon

-6 + 12                 x 12        x 12          x 8         x 6
-12 + 20                x 24        x 24          x 16        x 12
-20 + 30                x 40        x 30          x 30        x 24
-30 + 40                x 60        x 60          x 40        x 35
-40 + 50                x 90        x 85          x 60        x 50
-50 + 70                x 130       x 120         x 85        x 70
-70 + 100               x 190       x 170         x 120       x 100
-100 + 140              x 270       x 240         x 170       x 140
-140 + 200              x 400       x 350         x 250       x 200
-200 + 270              x 600       x 500         x 350       x 300
-270 + 400              x 900       x 800         x 600       x 500

Note: Use the following factors to adjust the calculated surface
area: Rounded: 1.2 Subangular: 1.4 Angular: 1.6

Fact 8. A Dirty Little Secret

Binder companies certify refractive index and free formaldehyde, but their results are often manipulated chemically to give a consistent analysis (and color) to the liquid resin. Binder certifications showing final refractive index or formaldehyde content are not meaningful. Suppliers must be trusted to deliver a consistent product.

Fact 9. Watch Solvent Content

Solvent content (or its inverse, the solids content) is a measure of the resin's dilution with the solvent (and must be known to work with LOIs). Generally, the higher solvent content resins ("low solids" systems) are more sensitive to sand temperature because they experience a higher evaporation rate. As a rule of thumb, the more solvent there is, the less developed ultimate strength.

The traditional PUNB catalysts are based on pyridine. The amount of pyridine introduced to the sand (along with sand temperature, and assuming chemically neutral sand) governs the reaction rate. Adding higher levels of pyridine makes the reaction go faster, because it is better distributed.

To slow the reaction, select a more dilute form of pyridine (the so-called "slow" catalysts). These catalysts are the same, only with more solvent. They make pumping relatively low volumes much easier. The solvent content in the catalyst can be critical, and if it is high, it can affect the volatile organic compound (VOC) emissions at mixing and can promote the formation of lustrous carbon at the mold-metal interface.

A new generation of catalysts based on "vinyl imidazoles" has been developed for its anti-sag characteristics, and these catalysts have different solvents that may reduce VOC emissions and lustrous carbon formation.

* Tip - With foundry lab evaluations, the best way to test raw materials and reclaimed sand is to follow a standard laboratory mixture and mixing procedure. When in doubt, make a set (six) of dogbone tensile test specimens under controlled conditions. The work time and strip time should be determined for these test batches. The results shouldn't vary more than 5%. For testing the binder, it is necessary to use "standard" sand - and it is often best (most consistent) to have several bags of new sand set aside in the lab for this purpose.

* Tip - When the foundry is running well and casting quality is excellent, set aside samples (5 lb) of new sand, reclaimed sand, any additives and binder components for comparison purposes in the event that problems occur later. Send perishables (binder components) out for analysis from their suppliers.

* Tip - Always conduct evaluations with fully blended sands from the mixer - with additives. That is what counts.

* Tip - Document work time. Make a 12-in. diameter disk about 1 in. thick - pie or cake pan. Roll over and expose the "rammed" surface (cover with plastic wrap - doing so better represents the sand inside the mold). Determine the mold hardness B every minute until it is [greater than]40 (arbitrary end of work time). Plot and laminate and keep at the mixer - check at least once a day. Reclaimed sand can reduce work time (preset sand will result in lower strength).

* Tip - Temperature of the sand at the mixer is a major factor in determining the work time and strip time for the nobake binders. Wide swings in sand temperature ([greater than]+-10F) can't be easily tolerated. to illustrate this impact, a 17F (10C) increase is sand temperature will double the chemical reaction rate for phenolic urethane binders.

The best way to test sand temperature consistency is to buy a simple kitchen timer and a probe thermometer. Every hour, take a sand temperature from the next mold made. During the test period, write down anything that might be important (such as "switched to new sand," "sand settling too fast") on the control chart. After at least one test a day, record how long it takes the sand to harden (judged by when a nail can no longer be stuck into the sand).

Plot this data on a control chart(X-bar/R-plot of individuals) for a week (30 tests) and then review for consistency. Calculate the mean and standard deviation - to determine temperature variation. On a well-controlled system, the standard deviation is 3F. This survey should be repeated quarterly.

SAND TESTING

Fact 10: Most Sand Tests are Meaningless

Most sand tests have nothing to do with casting quality, and the results often are due to variations different than those being studied. Sand tests are guides only, and generally don't define "acceptable" or "rejectable" conditions, especially when predicting the production of good castings.

There are three main reasons for sand tests:

1. To fulfill a quality assurance function (provide a record of how the system is operating and show that it's within established control parameters so that quality molds/cores can be expected);

2. To evaluate new or alternative materials;

3. Troubleshooting (run when casting quality deteriorates with no assignable causes).

Sand tests can be used to eliminate reasonable suspicions so that foundries can concentrate problem solving efforts to the real causes. The trick is to have a plan ahead of time.

MOLDING

Fact 11. Good Molds Require Good, Consistent Compaction

* Tip - A device call a mold quality indicator (MQI) has been developed to quantify mold compaction. A portable permmeter can also be used.

Fact 12. All Foundry Molds are Made Better Via Cope Venting

* Tip - A minimum recommendation is to vent on 2-in. centers.

RELATED ARTICLE: Phenolic Urethane Binder Chemistry

One of the most significant advances in foundry technology in this century has been the development of chemical resins for bonding sand. In the early 1970s, the phenolic urethane resins were introduced as nobake formulations, followed by the coldbox formulations. Phenol-formaldehyde resins were some of the first developed for the foundry (first as thermally cured such as used in the shell process, and then as acid-cured resins).

Chemically, each of these phenolic resins is different. The phenolic urethanes employ the so-called "pep-resins" (more formally referred to as polyether polyols - such as the polybenzylic ether-phenolic resin). These "pep-resins" have been engineered or highly modified to contain both ether and methylene "bridges" (very highly ortho-substituted). It is the phenolic hydroxyl groups that form the critical cross linkages that give these resins their useful properties.

These resins are unique in that while they have a highly molecular weight, they also exhibit relatively high fluidity and stability. This fluidity allows them to efficiently coat the aggregate, which decreases binder demand.

An important chemical characteristic of these "pep-resins" is their slight excess of formaldehyde. This "free formaldehyde" level is carefully controlled to minimize odors during sand mixing.

The other important consideration in binder formulation is the selection of solvent(s). These tend to be complex mixtures of aromatic hydrocarbons that dissolve the phenolic resin monomers but don't interfere with the polymerization reaction. Rate of evaporation is also an important consideration.

Resins also contain additives such as release agents. Silane is commonly added (especially to the coldbox formulations) to improve humidity resistance.

The second part of the phenolic urethane resin ("Part If") is a polyisocyanate resin of the methylene diphenylene diisocyanate (MDI) type. These resins have two or more reactive isocyanate groups in their structure. In reality, they tend to be blends of similar isocyanates.

Water reacts strongly with any isocyanate group to form substituted urea compounds - with a single molecule of water actually cross-linking and destroying two of the resins' important isocyanate sites. For this reason, it is important to protect the Part II resin from exposure to moisture in the air (by using vent desiccates) and to use absolutely dry sand ([greater than]0.1% moisture).

There is a chemical concept called "functionality" that must be considered by the resin formulator. Aromatic polyisocyanates give more rigidity to the final resin that bonds the sand; but there are stability problems that are best addressed by using aliphatic diisocyanates. Polymer chemists have solved this by blending these two types of isocyanates, along with appropriate solvents and release agents.

Generally, a high functionality resin is used in the nobake formulations, as opposed to the low to mid-functionality resins that are used in coldbox formulations. Higher isocyanate contents generally provide higher hot strengths and better sag resistance.

By definition, the isocyanate resins contain nitrogen, and functionality is to some extent a measure of the isocyanate content. However, the total nitrogen content of commercially available Part II resins varies only a small amount.

The Part III catalysts used in the phenolic urethane binders are tertiary amines that promote the formation of the crosslinks between the resins. This addition type polymerization reaction is unique among the foundry resins in that it occurs quickly, and nearly simultaneously through the sand mass. Once the urethane reaction takes place, it is irreversible. The rapid cure allows for rapid strip times characteristic of these binders.

The cross-linking reaction doesn't produce by-products (that subsequently can inhibit the reaction rate), nor do they require additional reactants, so the cure proceeds at nearly the same rate throughout the sand. This also makes the binder more sensitive to the catalyst.

In most applications, the nobake catalyst is a liquid amine called pyridine. This can be premixed with the phenolic resin (Part I) to assure good distribution.