Management - Introducing Automated First Article Inspection Series

Management have a difficult time in determining which equipment to invest in and which ones to stay away from. Each department usually comes with a seemingly sob story about why you should authorise some money to make his job easier and supposedly save the company enough money to make it worth it. Management have to weigh up the endless options from staff suggesting equipment and processes with technical reasons that sometimes are self motivated to the employees own agenda. These can sometimes be as simple as building an inspection dynasty for themselves and as devious as hiding some inability to perform a job and the equipment will hide this.

An Automated First Article Inspection System is a very easy system to justify financially except it is a difficult one to suggest because the department that will use it, and is providing the technical reason for the purchase, has to come clean that basically they cannot do their job properly without it. This will expose the fact that there is NO traceable system in place that 100% covers this function. They will have to admit that they do not perform a 100% FAI and/or are not exposing the time it takes to do it. They are afraid the obvious question will arise... "What the hell do I pay you for if this is the first I hear of this hole in our SMT Inspection area".

Somehow this impasse has to be broken if the FAI process can be fixed. The seemingly simple process of FAI is that prior to switching the machines to full speed building product in SMT after a changeover is that someone 100% checks the first PCB coming off the line to ensure it is built to the customers specification. When the SMT builds it wrong the management step to in ask the question, what went wrong. Usually there are a variety of excuses like, Someone did not check the feeders properly and management decide that a 2nd person should check and consider the problem solved and everyone pats themselves on the back, job well done. The other is to train the operators on inspecting the feeder setup, or simply fire someone, or both.

The problem will not go away, the fact is that in any manual process there will be human error. In the FAI process there are 2 places for error, in the conversion of the customer CAD/BOM data to Pick and Place file and placing the feeders in the right places. When checking these places at FAI the operator is left to check this without any help and the result of his inspection will load 100's of thousands of parts onto PCB's, potentially wrong. And he is being pressured to perform this task in a ever reducing length of time. In some factories I have looked at they are given 10 minutes to perform FAI for a PCB that contains 700 parts. This means the person performing this FAI must locate the part on the PCB and check if it is the right part in under 3/4 of a second average per part. That is the simple math. Since this is not possible they must NOT be inspecting the PCB 100%. Which parts are not being checked, does it matter, any part missed at FAI means that every PCB assembled will be wrong. Does this not worry you, it would scare the hell out of me.

At Cluso we have our own CEM facility here because we started as a CEM and designed the Cluso system for ourselves many years ago. We know the problems and we have solved them with the Cluso system. In the 7 years we have been using the Cluso system we have not loaded a single part in the wrong location... 

Can you say that??

The ROI on the system has been calculated in weeks not months. One of the largest CEM's calculated that the system would save them $70KUSD per week at the one facility when measured against 100% manual FAI methods. Needless to say we have sold many around the world to this company. Maybe you can wait until the next time you have to explain to your customer that you loaded 5000 PCB's with the wrong component and it was in a location that the test did not pick up...

Good Luck.

Greg Ross

Cluso Vision Systems