By Gregory Frenklach
I am not a big fan of the contradiction matrix as a means for searching the appropriate principles for problem solving. I prefer more accurate tools like elements of ARIZ or the inventive standards. Even the 40 inventive principles themselves – without the matrix – when reorganized are preferable in my opinion.
Nevertheless, trying to solve problems with help of the matrix can be useful. Trying to match real system parameters to the matrix's parameters enables a problem solver to get deeper into the problem and understand it better. And better problem understanding is a step in the right direction. The aim of the article is to make using of the "conventional" matrix easier and more algorithmic.
The following is the (more or less) conventional method of problem solving with the contradiction matrix:
Let's concentrate on the UDE and the known method. This pair determines each problem. But any problem situation has a number of such pairs. Therefore we have to find a way to determine other pairs of the problem situation. Problem situation specification and problem mapping can help. [4, 5] According to problem specification we have to determine the element connected with UDE and function of this element. According to problem mapping we have to define additional UDEs:
We also can define two more UDEs:
Each UDE should be treated as the original UDE.
Now we can match the UDEs of each contradiction to parameters (improving and worsening) of the contradiction matrix.
The test fixture is for measuring high frequency surface-mounted electronic components (couplers, filters, etc). The test fixture is installed on an automatic testing machine. The measurement is performed with a device designed for such work. The components are measured and then, depending on the results, are packed or thrown to the defect bin. During measurement, the components are placed on springy contacts of the test fixture printed circuit. The printed circuit is a three-layer sandwich – epoxy glass that is covered with thin metal layers as shown in Figure 2.
The test fixture measures about five components per second. The problem is that the printed circuit is very sensitive to the "strikes" of its measures. Metal layers get cracks and lead to incorrect results. This problem occurs after tens of thousands of measurements. The test fixture costs a lot of money and has to be removed from the testing machine and repaired. Repairs demand a lot of time, special equipment, high qualifications for repair personnel and costs a lot of money. What can be done?
|Table 1: The Resulting Six Contradictions|
|Contradiction number||UDE to be eliminated||The known method to eliminate UDE||UDE' that appears if the known method is used|
|1||Short life of test fixture||To change the printed circuit||It takes a lot of time and printed circuit is too expensive|
|2||There is no contact between component and measurement device||Printed circuit||Short life of test fixture|
|3||Cracks in metal layers of printed circuit||To change the printed circuit||It takes a lot of time and printed circuit is too expensive|
|4||There is no RF signal||Metal layers of the printed circuit||Cracks in metal layers of printed circuit|
|5||Low measurement reliability||Re-measurement||Poor productivity|
|6||There is no contact between component and measurement device||Printed circuit||Low measurement reliability|
Now we can match these contradictions to parameters' contradictions of the matrix. At this stage a problem solver familiar with problem formulation, problem mapping and re-formulation knows to make the change in the printed circuit like in the real solving process. It is clear which principles should be applied, but we will use the contradiction matrix to find the appropriate inventive principles.
|Table 2: Recommended Principles Using the Contradiction Matrix|
|Contradiction number||Parameter to be Improved||Parameter That Worsens||Recommended Inventive Principles|
|3||30||34||35, 10, 2|
|4||24||30||22, 10, 1|
|5||27||39||1, 35, 29, 38|
|6||24||27||10, 28, 23|
The printed circuit is turned from a sandwich of three layers firmly connected to a sandwich with layers that are not connected. Such a design eliminates the strains that cause cracks in the metal layers because of repeat contact during measurement. As a result, the test fixture is ten times cheaper, reliable during millions of measurement and easy to repair.
Using the presented algorithm for problem formulation makes it easy to see which parameters to use and which principles to use to develop solutions
Gregory Frenklach is a R&D engineer at Medinol in Israel. Contact Gregory Frenklach at gregory_f (at) 012.net.il.