Indentifying human error as a root cause

[Lee Taylor]

This is my 1st post and I hope I have it in the right forum.

We’re a sub contract assembly manufacturer and we make various PCBA’s and assembly enclosures for all industries.

I’m struggling to identify the route cause for operators missing the fitting of some conventional parts during assembly (Resistors, Diodes, Capacitors etc) This may only be one component here and there but these also being missed during the inspection process.

All the operators and inspectors are experienced. (IPCA-610 & J-STD) yet PCBA’s with defects are getting dispatched to our customers. Further to this, the defects are not isolated to any particular customer or assembly. Although, I have narrowed it down to two operators and one inspector.

Once these assemblies get dispatched it can be a month or two before we are informed of the defects making it difficult to identify the route cause.

All documentation in relation to the job is correct and clearly detailed. Which leaves human error, something I don’t like recording this as a route cause. There are several human factors that could be the cause of the defects, Inattentiveness, Fatigue, Distraction, Rushed, Habit, Stress or even all of these.

How can I overt these issues when I can’t pint point the exact root cause?

 

[Marc]

Do you mean ROOT cause?

 

[Lee Taylor]

Yes, Thanks for pointing that out

 

[somashekar]

AOI
Automated optical Inspection
Its time you install this.

 

[Lee Taylor]

We use Automated optical inspection mainly for surface mount components, but the AOI had limited internal height restricting its use for through hole components.

Using an AOI would'nt help prevent the the operator not fitting the part in the 1st place. This would only assit in finding the defect.

 

[somashekar]

Using an AOI would'nt help prevent the the operator not fitting the part in the 1st place. This would only assit in finding the defect.

I agree with you.
The approach is to first arrest the defective from going out of your gate.
Then the result of the AOI can help in indicating the place where such errors happen.
Then it is a case of building competency, reducing the component load per operator, or anything else best suited to your context.

but the AOI had limited internal height restricting its use for through hole components.

This is a constraint of your AOI which has to be addressed to meet to your requirements.
Good Luck

 

[John Predmore]

There is ongoing discussion here at Elsmar Cove about human error as root cause. I liked the HERCA worksheet shared recently by
Johnnymo62 that allows root cause investigation to go deeper than blaming the operator. Many times, experienced operators make errors because of distractions, and improvements in workplace layout and work environment policy can go far to minimize distractions and errors.

You asked how to pin-point what happened when the error is discovered weeks later. You could, as a temporary measure, do audits or added inspection, and try to catch errors closer to the point in time when they are created. Depending on the culture in the workplace, you could ask downstream workers to "inspect" each PCBA before they accept it, at least look for the Top 10 most common problems. You could ask workers to keep a log, date and time when their workflow is interrupted by distractions. In order not to become too "big-brother"ish, maybe keep the logbooks private until there is a specific problem incident discovered, then explain you are trying to correlate a specific failure occurrence on a specific day and time to a specific root cause. Good Luck. Let us know if you find a solution that works!

 

[Sidney Vianna]

1. Ask the operators what could be done to ameliorate the situation. Without blame, engage the people directly involved with the work and do not isolate them from the problem potential root causes and possible solutions.

2. Assign the kits with very precise quantity of components for the hand-assembled, through hole pieces. If there are 2 resistors left after assembly, you know they were not installed on a board.

3. How are the defects identified downstream at the customer? A functional test? Replicate the method at the end of your line. Many times the customer will assign a test fixture to a supplier exactly to identify problems early on and upstream of their lines.

Good luck.

 

[outdoorsNW]

You may need a better AOI machine. Some are better than others at handing tall parts.

Also, can electrical test detect the missing components? I know not all nets are testable. Again a better machine can test difficult boards that a less capable machine cannot.

 

[Ninja]

It may also help to pre-measure part quantities.
If they have to mount 34 resistors, 22 diodes and 16 caps...and that's what they have...if they have any left over, they know they missed something before AOI even gets run...
Is this an option?

Also, since it is circuitry we're talking about, electrical testing can determine open circuits because a resistor is missing, or a cap missing. can the right leads and probes be identified to do this testing?