22
”average” moisture values that are close to the values being generated on the same
samples by standard oven. To obtain the best possible results for moisture, and
thus for fat, there are three factors to consider: 1) sample preparation, 2) sample
application technique and 3) power and time parameters.
After establishing moisture repeatability and achieving moisture results as close to
standard ”long” methods as possible, use the following procedure to identify fat
biases for different product categories.
1.
Select three or more samples (approximately two pounds each) from
different lots of the same product.
2.
Prepare each sample according to standard procedures (3 - 4 passes
through a 5/64” blade, with mixing between grinds, or equivalent).
3.
Divide each ground sample in half and package and label sub-samples in
two separate, air-tight containers.
4.
Analyze one set of the sub-samples (from step 3) in duplicate by standard
”long” methods – oven for moisture, Soxhlet for fat. If facilities and/or
personnel are not available for these tests, use an independent laboratory.
Request raw data (individual tests) from the laboratory, not the mean or
”average” value.
5.
After obtaining the results from the standard ”long” methods, analyze the
other set of subsamples in the CEM Moisture/Solids Analyzer for
moisture only. Adjust power and spread technique until acceptable
moisture repeatability is achieved.
6.
After obtaining acceptable moisture repeatability, use the Fat Analysis
System and perform 3 - 5 moisture and fat analyses for each sample. As
with moisture, fat results should be repeatable. To obtain optimum fat
results, use a sufficient redry time. Two minutes is typical, although some
samples may require more time. The sample weight should be stable
(<0.2 mg weight change in the last five seconds). An insufficient redry
time will yield artificially low fat results since the true dry, extracted
weight will not be reached.
7.
Subtract values obtained from the SMART System 5 from Soxhlet values.
Average the differences to identify the average bias for a specific product
or product group.
Sample A
B
C
Average Soxhlet fat
xx
xx xx
- Average CEM fat
xx
xx xx
= Bias
xx
xx xx
(A + B + C)/3 = Average Fat Bias
8.
The bias should be relatively less than 2.0% and must be added to the fat
results from the SMART System 5 to produce results that correlate closely
with official methods. Repeat this process on a regular basis to confirm
that the bias is consistent. Raw meats and blends will produce smaller
biases than cooked products with extenders.