Why Do I Get Inconsistent Moisture Test Results?
Posted by Art Gatenby on Fri, Oct 14, 2011 @ 01:32 PM
“Why are my Moisture Test results inconsistent?”
That is an issue for many of you who test for moisture. We discussed the complexities and multiplicity of issues involved with moisture content determination 
in our “Loss-on Drying Moisture Analysis and other Moisture Mysteries” series.
In addition to intrinsic properties of test samples that may adversely affect moisture testing systems, automatic equipment parameter set-up, operator oversights and sample handling contribute to seemingly intractable moisture test result inaccuracies.
Common test sample vagaries include:
- Volatiles other than water that are released at close to the same vapor pressure as water evaporation.
- Strange conditions of entrapped moisture that release water at capricious times.
- Samples not representative of the principal batch
It is important to determine if these test sample quirks are responsible, because test protocols may require changing. Variations can also be minimized by running more tests to 
statistically reduce variation effects. In some cases, changes in test methods may be needed (i.e. Karl Fischer rather than Loss-on-Drying).
While trouble-shooting such problems, it is important to check automatic setting level and operation, which determine when a test is completed. With the Loss-on Drying method, these settings relate to measuring sample weight changes. If the instrument is set to stop too soon, the weight-loss curve will slope steeply and the moisture result will be subject to irregular variations from test to test. If the end-of-test calculation is based upon too small of a weight change, there is a potential for burning the sample -- another cause of inconsistency.
Similar problems can arise if widely different test-to-test sample weights are used in a timed test environment.
Another source of inconsistent analysis can be that a small amount of moisture requires detection. A small amount of sample and a very sensitive balance are frequently used for this type of test. Operators must carefully follow test procedures or the balance’s high sensitivity will yield wide result variations from test to test ( Often A switch to the Karl Fischer Method will Solve this Problem).
As equipment manufacturers, we will ultimately consider the possibility of instrument malfunctions. Our experience with instrument service leads us to either clear instrument failure 
or consistent high or low results signaling equipment problems and not just inconsistent results.
Our troubleshooting protocols require duplicating clients’ problems in our lab. When we cannot, experience leads us to consider environmental conditions at our customers' facilities. Frequently, we find electrical power conditions to be responsible. Special power-conditioning equipment will usually solve this problem.
On occasion, we cannot find the sources of variation.
In summary, issues relating to test sample properties and/or or incorrect test parameters normally yield inconsistent results. Occasionally it is operator error. However, on rare occasions, test site environmental conditions are responsible.
Sometimes it requires painstaking investigation to find the cause. When found, we can usually then develop workable solutions.
I hope this sheds some light on the sources of inconsistent test results.
As usual, there is still a bit of witchcraft 
and folklore needed to solve the more elusive measurement problems.
Still trying to get answers, I remain a puzzled,
Art
P.S. Subscribe to these articles by clicking the RSS buton or by inserting your email address in the space provided.