In our last issue, we discussed using a diagnostic flow chart. This doesn't work for all systems or all organizations. An alternative approach is to do sequential testing against a series of possible causes (Check for A, if not A then go to B, etc.). The key to making this approach work is to limit the number of causes to be checked. How can you decide which ones to include?

Make a spreadsheet and list eight to 12 failure modes across the top (you can pull these from your repair logs). Include an "Other" column. Now comes the tricky part. You need to sort these from most likely (on the left) to the least likely being (on the right). One way to do that is to have several senior repair techs rank these based on their judgment. For example, the most likely gets a 1, next most likely a 2, and so on.

Plug the rankings in under each cause and have the spreadsheet total them. Next, plot the results with causes sorted from lowest score to highest. As you look from left to right, you should see a "break point" in column height. Stop there, and don't include the least likely causes.

Repeat this exercise every couple of years or as conditions indicate.