Undersized gray iron test bars cause reduced tensile strength.

Tensile tests remain the primary basis for proving that the properties of gray iron castings are adequate to meet customer needs. In many cases, tensile tests are taken directly from castings to assure compliance. Too frequently, the size of the test bar is made unusually small because the person conducting the tests doesn't understand the effect that the machined size has on the results. Those conducting the tests may even feel that they are giving the foundry a break because they are aware of the section sensitivity of iron.

Experienced iron foundrymen know that pouring into a smaller bar yields higher tensile properties, while cutting smaller bars from the same size cause lower results. Unfortunately, documentation of that effect isn't readily available.

Experiment Procedure

The AFS Cast Iron Div. Quality Control Committee (5-J) decided to substantiate the well-known but little-documented fact that tensile property results are significantly affected by the size of the machined tensile specimen. The committee performed a test to demonstrate the significance of the effect.

The mold described in ASTM A-48 to produce the tensile specimens produces two bars. For this experiment, two "B" size test bars were produced in each mold at Seneca Foundry, Webster City, Iowa. The assumption by ASTM and the AFS Quality Control Committee was that both bars should have identical tensile properties. In order to make statistical comparisons, 10 of these molds were filled as rapidly as possible from the same ladle of iron.

In order to negate any effect of different shakeout times, the molds were allowed to cool for 4 hr before they were shaken out. Each mold was shaken out individually, and the two bars were marked and banded together for transfer to the testing laboratory.

The testing laboratory, taking care to keep the identity of the matched sets, machined one of each set to a "B" size test bar and the other to the smaller "A" size, as described in ASTM A48.

Size Matters

The laboratory, using the procedures described in ASTM E-8, determined the tensile strength of the material in each of the bars. The actual results can be seen in Table 1, but the averages and standard deviations were as follows:

               "B" Bar      "A" bar

Average        43,320 ksi   41,120 ksi
Standard
Deviation      942          1290
Table 1. Tensile Strengths of A and B Machined Test Bars
(thousands, ksi)

A  39  41  43  41  41  41  42  40  42  42
B  43  44  44  43  42  44  42  43  44  43

Using the standard statistical difference of means test, these results provide a z value of more than 4. In other words, it can be said with a 99.8% degree of confidence that the bars machined to the "B" size have greater tensile strength.

An unexpected observation that showed in the results was the higher standard deviation of the results from the "A" size bars. Applying statistical techniques to this difference showed that it was not significant at the confidence levels usually applied by statisticians (greater than 90%); however, few things have that degree of confidence in a foundry.

This study provides documented proof to what many iron foundrymen have known for a long time: machining smaller test bars from a given size will cause lower tensile property results.

As with most experiments, more questions were raised than answered. As noted above, the variability of the results from the test bars machined to the smaller diameter was greater. While the results from the small experiment didn't show the variability was statistically different at a level normally used by statisticians, it is noticeable enough to consider further investigation.

In addition, while this test showed statistically different results in the tensile properties because it dealt with only two sizes, no effort could be made to quantify the relationship between the size and the loss of tensile strength. This would be a worthwhile effort for future investigations.

Finally, this work was done on class 40 gray iron. It may be a safe assumption to believe that these results are indicative of what would be found with all grades of gray iron; however, the question of the effect on ductile iron needs to be investigated.