The finished product samples performed well in the moisture testing oven. The in-process samples did not. The in-process sample was 40 degrees hotter. This was the root cause of the problem. The higher starting temperature caused the sample to “pop” in the moisture testing oven. This “pop” expelled oil from the sample pads, skewing test results. We only wanted moisture removed from the sample.
As a result, liquid margarine batches were continually resampled, corrected, resampled, etc. Six batches were mixed together before packaging. Each batch took 45 minutes to crystallize (emulsified by cooling rapidly). Once the product was crystallized, it was finished product.
Luckily, Tim was helpful. He agreed to get a second in-process sample after the batch started the crystallization process. Testing on the crystallized samples revealed the production process was outstanding. The margarine batches were consistently on target; very little variation existed between batches which was expected with the state-of-the-art equipment. A deficient in-process sampling procedure slowed production to the point everyone was working weekends.