Baume: complete coating control?

AFS Molding Div. Mold-Metal Interface Reactions Committee (4-F)

Inside This Story:

* A comprehensive analysis of the Baume test for refractory coating process control is performed.

* Using the Baume test as a sole quality control test inherently introduces a wide range of actual percent solids, resulting in fluctuating variations in surface and casting quality.

For years, foundries have struggled with the concept of which control tests should be included in a sound refractory coating process control program. Debate has centered on the relative importance of the tests available, the balance between economics and number of control point locations, and the recommendations by suppliers and industry experts.

In recent history, the coating control test that has received the most attention is the Baume test. This article will deduce how accurate the Baume test actually is, and explain any inconsistencies present in the use of this process control test for refractory coatings.

What is the Baume Test?

The Baume test is the most common test used in foundries to control coating operations because it is both quick and easy. The test is performed with a hydrometer, which is a sealed glass tube that contains a calibrated scale in degrees Baume.

The hydrometer should be clean and dry with a resolution of at least 0.5 degrees. The gauges reference the material they ate to be used for and the temperature the material should be to achieve measurement accuracy. The Baume scale of numbers relates to the specific gravity and body of a coating.

After mixing the coating sample thoroughly, the hydrometer is immediately floated into the coating slurry. When it stops sinking, the degrees Baume can be read direct[y from the hydrometer. This test requires that the coating be homogeneous, at the correct temperature have no air bubbles and be completely still.

Phase I

In order to perform a comprehensive analysis, two water-based coatings were chosen, One was a low solids ceramic "automotive type" coating between 28-32% solids (32-34 Baume) demonstrating a significant yield point with some thixotropic nature. The other was a high solids zircon coating between 71-73% solids (90-92 Baume).

The high-resolution glass hydrometers that were utilized had a resolution of 0.1-degree Baume. Readings using these hydrometers were estimated to the nearest 0.05 degrees. Low-resolution plastic hydrometers representative of many typical foundry applications also were used. These hydrometers had a 2-degree Baume resolution and readings were recorded to the nearest 0.1 degree.

The dry weight gain for each Baume value, weight per gallon and percent solids values were collected and recorded for both coatings. The weight per gallon cups used in the experiment featured recordings to the nearest .01 lb/gal. Three replications for each operator were performed. Percent solids values obtained represent the shortest possible delay period and the longest.

The experiment was designed for up to four operators. The operators collected all samples from a carefully mixed common lot of refractory coating. For each sample, the operator recorded multiple Baume readings first with a glass hydrometer and then with a plastic hydrometer. There were three prescribed delay periods prior to starting any tests: 30 sec, 120 sec and 240 sec after mixing. This procedure was repeated for three test cycles.

The data generated suggests some interesting observations. First, the standard deviations for each operator across all delay times and hydrometer dwell times ranged from 0.292-0.363 (average of 0.324) for the high solids coating and glass hydrometers. In contrast, the standard deviations for the plastic hydrometers ranged from 0.50-1.30 (average of 0.80). So, for the high solids coating, plastic hydrometers produced Baume readings with more than twice the average standard deviation value than glass hydrometers.

A similar analysis for the low solids coating showed that the standard deviations for each ranged from 0.519-0.712 (average of 0.609) for the glass hydrometers. The standard deviations for the plastic hydrometers ranged from 0.80-1.30 (average of 1.05). Again, as was the case for the high solids coating, plastic hydrometers produced standard deviations that were considerably higher than their glass counterparts for the low solids coating (Fig. 1).

[FIGURE 1 OMITTED]

The coatings themselves demonstrated measurable differences in Baume reading variability. Readings taken for the high solids coating showed less variability than readings for the low solids coating tbr both hydrometer types.

A comparison between experimental ranges also was made by hydrometer type for each coating. Highest and lowest Baume values for each test start time, hydrometer type and coating were used to calculate the ranges. The purpose of this was to quantify any hydrometer sensitivity effects on these range values, The results, presented in Fig. 2, indicate that a 20 fold increase in hydrometer sensitivity ([+ or -] 0.05 resolution versus [+ or -] 1 resolution) reduced the variability in testing results by more than 60%.

[FIGURE 2 OMITTED]

For the high solids coating, the range mean was 0.92 for the glass hydrometer compared to 2.4 for the plastic hydrometer. Standard deviations for each were [+ or -] 0.286 and [+ or -] 1.1 respectively. The same analysis was carried out for the low solids coating where the range mean was 1.13 for the glass hydrometer and 2.9 for the plastic. Standard deviations were [+ or -] 0.159 and [+ or -] 1.0 respectively.

The effect of hydrometer settling time also produced some interesting results. The high solids coating demonstrated decreasing variability from start to end, with some stabilization realized as the hydrometer was allowed to settle. This trend was not evident for the low solids coating, suggesting some influence by the coating itself.

Weight per gallon and percent solids showed much less variability as expected and were less time sensitive than the Baume test. No conclusive relationships were observed between the dry weight gain results and the time delay considerations used in the experimental design. There also was a difference in the dry weight gain standard deviation values between the high and low solids coatings. The values across delay to test start times and hydrometer dwell time for each operator was much less for the low solids coating. In fact, the average dry weight gain standard deviation for the low solids coating was about hall of the same value for the high solids coating.

Phase II

The second experiment featured six operators, three each diluting two lots of high solids and low solids coatings to two different Baume levels with three replications for each dilution. Each operator completed four dilution sequences in triplicate for a total of 12% solids tests being run by each operator. Following each sample dilution, a percent solids was determined for the sample. The low-resolution plastic Baume hydrometers were selected and used in the experiments to represent the typical foundry hydrometer.

The experimental results are presented in Table 1 for both the high and low solids. Two sets of values are reported: the "actual" values are taken directly from the raw experimental data, while the "statistical" ranges are based on the statistical mean value [+ or -] 3 sigma.

Table 2 presents a case study of the relationship between Baume (as a process control) and the percent solids of a coating in a flow coat applicator. The foundry's production target Baume for this low solids coating is Baume 34 with an established process range of [+ or -] 1 and a requirement for hourly Baume checks. The data in Table 2 was collected by having quality assurance personnel collect a sample from the production area and then test the sample in the laboratory for percent solids.

Actual values for the high solids coatings progress from 0.25% "as made" to 1.48% at 90 Baume, and finally to 2.46% at 85 Baume. These values demonstrate a trend toward a wider range in solids as the coating is diluted to lower Baume values. The low solids coating demonstrates a similar trend progressing from 0.32% "as made" to 2.96% at 30 Baume and finally to 4.75% at 25 Baume. Both the high and low solids coatings demonstrate a significant increase in variability as they are diluted to lower Baume values.

Upon comparing the statistical ranges at mean values [+ or -] 3 sigma to the raw data it is noted that the same increase in percent solids is present. The fact that the statistical ranges are wider than the actual ranges more accurately reflects the reality of the test results. The [+ or -] 3 sigma values indicate that there is a greater variation associated with the test results than is apparent in the descriptive range statistic. The increase in range of percent solids demonstrated by the statistical values is a consequence of the variability associated with using the Baume hydrometer.

For a sand foundry using coatings, the more a coating is diluted, the wider the range in percent solids of the coating. Ideally, a coating control program based on the Baume test should function within a strict ideal operating range, however, the actual percent solids will be outside the ideal control window for what would be thought as good quality control. Therefore, using the Baume test as a sole quality control test can result in fluctuating variations in surface and casting quality.

Baume Test Variation

Sources of Baume test variation can consume a foundry's operating range, with the magnitude of variation changing from coating to coating, as evidenced by the difference in numbers seen between the low and high solids coatings. There was less variability in the high solids coating Baume values for both hydrometer types. As suspected, there is a definitive difference in the amount of variation and resultant confidence levels between both types of hydrometers, with the glass models showing marked superiority.

Variation between multiple operators was also demonstrated, as a reduction in variability occurred as dwell time increased. The percent solids and weight per gallon tests offer a higher degree of process control and are more reproducible.

It is important to note that the tests performed were conducted under ideal laboratory conditions and not in a foundry environment. It is reasonable to assume that the data presented is a "best case" scenario. An attempt was made to cover many variables like multiple combinations of operators, delay to test start times, hydrometer types, dwell times and coating solids contents.

Using Baume as an in-process control test requires extra cost to maintain casting quality. This additional cost manifests itself in engineering design to compensate for a wide fluctuation in percent solids, and is a burden on the casting designers, tooling designers and coating manufacturers.

A density measurement test, such as weight per gallon or percent solids, should be run in conjunction with Baume. However, if a foundry uses Baume as its principal refractory coating control tool, variability can be reduced by using a high resolutions hydrometer as well as specifying the lag time until the start of the test and the hydrometer dwell time. MC

For More Information

"Evaluating Refractory Coatings: A Practical Approach," S.G. Baker, MODERN CASTING, October 2002, p.21-23.

Mold & Core Coatings Manual, 2nd Edition, American Foundry Society, pp. 60, 64, 72, 77, 86.

Table 1. Experimental Range of % Solids (Actual Ranges)

                            Highest      Lowest     Total Range
                   Baume    % Solids    % Solids     % Solids

High Solids Experimental Coating
As Made                      81.43       81.18         0.25
First Dilution      90       80.88       79.40         1.48
Second Dilution     85       79.40       76.94         2.46

Low Solids Experimental Coating
As Made                      45.37       45.05         0.32
First Dilution      30       32.57       29.62         2.96
Second Dilution     25       30.92       26.17         4.75

Table 2. Experimental Range of % Solids (Statistical Ranges)

                             Highest      Lowest     Total Range
                   Baume     % Solids    % Solids     % Solids

High Solids Experimental Coating
As Made            100.56     82.50       80.40         2.10
First Dilution      90        81.5        78.94         2.56
Second Dilution     85        79.94       75.98         3.96

Low Solids Experimental Coating
As Made             66.45     46.04       44.84         1.20
First Dilution      30        33.36       28.79         4.57
Second Dilution     25        30.94       23.66         7.28