Among all the issues an engineer has to address when it comes to designing an accurate and reliable scale, temperature and creep characteristics are two important parameters he has to constantly work with throughout the design. Temperature changes physical characteristics of a material. A material expands or shrinks as temperature changes. For instance, a lever ratio or resistance in an electrical circuitry may change when it is subjected to a change in temperature.

Creep test is a test that indicates the long-term stability of a sensor. When a material is put under a stress, it deforms and reaches a new equilibrium position. However, if such a stress continues to exist in time, the newly reached equilibrium position may not stay the same but may continue to deform or change gradually. When a material is placed on the weighing pan, the scale or sensor in the scale reaches a new equilibrium position indicating the weight of the material. But when the material is left on the pan for a period of time, the weight display may shift as the equilibrium position may gradually change in time. This phenomenon is called creep. Creep of a scale is defined as a change in reading when a weight is left on the pan for a fixed amount of time. A typical creep test will be to leave the weight equal to the capacity of a scale on the pan for five minutes and record a change in reading.

A great amount of engineering and know-how are required to minimize the effects of temperature and creep so that the scale will give accurate and reliable readings at all times. Under normal circumstances the ambient temperature stays the same or in a reasonable range while in weighing, besides you record a reading soon after you see the reading becomes stable, therefore, the effects of temperature and creep will not be critical factors in weighing. (For a type approval like EC, NTEP and NSC these two parameters make an important part of testing procedures.)

Though a moisture balance does not fall under the scales that require type approval, these two parameters become critical for accurate reading. By nature of its use the material to be tested is left on the pan of a moisture balance for a period of time while the material is being heated or dried. In other words the moisture balance is put under a very severe environment or such an environment that one wants to avoid in case of using a scale for weighing. It is being heated up to a few hundred degrees from the room temperature in a short period of time and a material is left on the pan for a lengthy period of time, which means that any amount of creep becomes a source for error. In other words a moisture balance is subjected to tests of temperature and creep by nature of its use all the time. It is equivalent to using a balance in an oven with a pancake on the pan. Thus, it will make a showcase of the know-how and philosophy of a company designing balances and scales regarding accuracy and reliability.

Our new moisture analyzer MX/MF should be the showcase how A&D designs or the degree of our know-how in terms of controlling temperature and creep characteristics. It has been commonly understood in the industry that 0.01% reading is not necessarily absolute in accuracy unlike in the case of weighing. When a moisture balance shows the result by 0.01% it does not mean its accuracy is definable in terms of 0.01%. Our comparative study suggests that our MX is the first one that truly shows the 0.01% accuracy in standard deviation. Below shows the result of our MX and two other competitor's models. We used sodium tartrate dihydrate, which has moisture content of 15.66%. We had 5gram sample and repeated the test 6 times.

 

Average reading Standard Deviation     Max   Min   A&D's MX-50     15.72%    -  0.015%    0.004%   Competitor A    15.79%    -  0.058%    0.014%   Competitor B    15.76%   -   0.053%   0.013%

Last week I attended our European Partners Meeting held in Dublin, Ireland where we had over 60 people gathered from all over Europe. On the last day of the three-day meeting we conducted product training on the MX/MF and WinCT-Moisture, software that shows real-time graph display of weight changes and helps one find optimum conditions for moisture analysis. We used Sodium Tartrate Dihydrate and cornstarch for testing. We also tested liquid or milk samples by using glass fiber sheets. When we had all the tests we planned for the training done and were waiting for the summary of the training, people in one group poured water from a bottle on the table into the container that has glass fiber sheet and began testing or drying it. They obviously thought of testing the moisture balance if it would show the obvious result or 100%, as water poured into the container should evaporate completely, should the balance work accurately as A&D people claim! "Will it be the case with MX?" was a question they had in mind. In other words they incidentally began conducting a test on the moisture balance to see effects from temperature and creep. It greatly attracted the attention of others, and out of curiosity people in other groups followed the test.

We noticed 4 tests were being conducted simultaneously; the very test the engineer wants to avoid being done in uncontrolled conditions. Two groups of people got 100.00%, and the third group got 100.02%. The fourth group got 102%. When the fourth group got 102%, all the attention of the people gathered around the moisture balance of the fourth group. When the lid of the balance was raised, it was found that there were two containers sitting on the pan. Suspicion was there was some water left between the containers below suggesting the unmeasured water was the cause for the additional 2% over 100%. Naturally we repeated the test. This time with one container having around one gram of water absorbed by glass fiber filter. To our relief we got 100.00% in reading, and everybody was convinced of the fact our MX truly gives 0.01% in accuracy.

This test with water is a simple, straightforward test. Yet it will be a nice test to show the accuracy of a moisture balance or a balance under the effects of temperature and creep. It incidentally highlighted the superiority of the 
SHS (Super Hybrid Sensor) sensor and the design of our new moisture balances. 
Our slogan of "Test with The Best - True 0.01%(S.D.) stood the test.