Eliminate Porosity by Eliminating Test Variation

Removing hydrogen from aluminum and preventing the formation of internal porosity in castings requires an operator to degas the melt prior to pouring. Rotary Inert Degassing (RID) – the most common degassing method – distributes an inert gas, such as nitrogen or argon, throughout the melt. The inert gas combines with the hydrogen gas in the melt and carries it to the surface where it burns off .

In order to determine the effectiveness of degassing, most foundries perform a Reduced Pressure Test (RPT) on the melt prior to pouring. In the test sample molten aluminum is solidified under a vacuum of 25.5 to 27.5 Hg for a minimum of 7 minutes. Solidifying the sample under vacuum expands the hydrogen gas by approximately 10X. This expansion of gas porosity in the sample allows varying levels of porosity to be measured.

Unfortunately, the manual RPT test method used at most foundries is highly operator-dependent and can result in test-to-test variation. The inherent process variation in both sample collection and sample analysis of the RPT test often leads to difficulty passing internal, customer, and third-party audits.

The aluminum foundry can reduce RPT test variation by updating its procedures and implementing operator check sheets at each step of the test, however these controls can be hard to manage and often are frowned upon by quality system auditors. Many foundries striving to satisfy ISO, TS, QS, and other quality systems have turned to computerized RPT test equipment to eliminate operator variation in key steps of the process. Equipment like the Palmer PAS5000 automatically controls critical test parameters and RPT sample analysis, eliminating operator influence from key steps of the process.

To eliminate process variation in RPT testing, the source of variation in each step of the process must be understood and controlled. The following details will help you understand, control, and transform your RPT testing into a quantitative test.

1.  Sample cup. The sample cup can be made from either carbon steel or SS steel. To maintain a molten sample prior to placing it in the vacuum chamber, the wall thickness should be approximately 1 to 1.5 mm. If the sample cup is too thick or too cold the RPT sample will start solidifying before being placed under the vacuum, causing inaccurate results.

A typical sample cup is approximately 60 mm in diameter by 20 mm in height. To promote sample release from the cup and reduce heat transfer from the molten metal to the cup, coat the sample cup with a thin layer of high temperature release agent or coating, such as ZYP boron-nitride aerosol.

2. Collect the sample. Process variation when collecting the molten aluminum sample can be caused by a dirty melt, improperly filled sample cup, or insufficient heating of the sample cup prior to filling. The melt surface should be skimmed to remove the oxide layer immediately prior to collecting the sample. Once skimmed the sample cup should be filled with molten aluminum and held over the melt for around 10 seconds. Then molten aluminum is then poured back into the melt.

Once emptied, the cup should be slightly glowing. Now the test sample can be collected by “back-dragging” the cup across the surface of the melt to push back the oxide surface layer and filling the cup from the clean area of the melt. The full or nearly full cup should be placed under the vacuum dome within 30 seconds.

3. Produce RPT vacuum sample. It is critical that the sample is completely molten prior to starting the vacuum. Process variation during solidification of the RPT sample under vacuum is common when using a vacuum with a manual valve and gauge. In this case, the foundry is completely reliant upon the operator to run the vacuum test to the procedure. Test results can vary if the operator does not pull vacuum to the same level each and every test.

When using a manually controlled vacuum, the operator must ensure the vacuum is maintained at the target set-point and there is no leak in the system. In addition to ensuring that the vacuum reaches the correct level, the operator must allow the RPT sample to solidify completely before turning off the vacuum pump. Complete sample solidification takes a minimum of 7 minutes, and when using a manually set vacuum a dedicated timer should be added to the operator station to ensure the RPT sample is not removed prior to full solidification.

The vacuum target can be set within a range of 25.5 to 27.5 Hg. A typical vacuum setting used by many foundries is 26.5 Hg. The main thing here is to be consistent. If the foundry SOP states 26.5 Hg as the target, 26.5 must be used for all RPT tests. Inconsistent results will occur if the vacuum is set at 25 Hg for one test and 27 Hg for the next.

4. Analyze the RPT sample. The legacy method of analyzing the RPT sample has been to cut the sample in half, grind the surface, and compare it to a chart showing example photos of varying degrees of porosity. This method can be dangerous (saw cut of small samples), time consuming, and fraught with process variation.

For example, two operators visually comparing the same sample to a chart will frequently arrive at two different conclusions. The good news is that there is a safer, faster, and more accurate way to analyze the RPT sample. Since the density of the RPT sample is directly related to its percentage of internal porosity, the specific gravity (S.G.) of the sample can be measured providing an accurate repeatable result. With this method, the requirement of cutting the sample and visual analysis is eliminated.

Computerized RPT systems (e.g., Palmer PAS5000) use the specific gravity method to completely remove the operator influence from sample analysis. In addition, these systems allow the foundry to set a pass/fail S.G. target. For example, a foundry wanting to produce a low porosity 356 casting would assign a target S.G. of 2.60. In less than 30 seconds the S.G. of the RPT sample can be measured. If the result is 2.60 or higher, cast. If the result is 2.59 or lower, continue to degas. No need to saw, polish, or make an arbitrary judgement.