Indentifying human error as a root cause

The OP needs to try to implement AOI for through hole devices. Many AOI machines are optimized for surface mount and don't always see through hole parts well. And even then some parts (some capacitor types for instance) have their markings on the sides in a place that other components can block and therefore the AOI has zero chance of telling a 123 from a 456 because the size is the same, the orientation is the same, and the color is the same.

Yes, ideally the process would be improved. But currently AOI machines are close to doing what the OP needs while through hole machines are expensive, limited in the types of components any one machine can handle, so a company needs several, and not economical for small to medium size production runs.

Better customer design would help, but the reality is most designers know little about PCB assembly and therefore churn out designs with numerous design for manufacturability problems. What worse is many don't want to listen when you try to tell them how to make their board easier to build.

outdoorsNW said:

What worse is many don't want to listen when you try to tell them how to make their board easier to build.

Click to expand...

The sales people should learn to decline the work. Not knowing the impact on the cost of goods sold by dealing with customer-induced errors is the slip 'n' slide to going out of business...

The problem is so prevalent that it would be hard to stay in business if a company rejected everyone with poor designers. The problem is so bad it is almost like dealing with the weather. You just have to build the cost into you pricing to the extent you can.

The big problem is PCB design is rarely taught in colleges. I checked several local engineering programs along with Stanford and MIT and only found one elective class at one local school. Most designers only care about the electrical functionality and not DFM. Many frequently ignore the IPC design guides in the 222x series of IPC documents.

As someone who runs every aspect of a business on a daily basis...what I read is:

Guest said:

The sales people should learn to decline the work. ...knowing the impact on ... going out of business...

Click to expand...

Time to fire those guys.

Hi, did you try to solve this issue before looking for cause analysis? My questions are the plant's overall production within the limit of your customer specification? are the missing parts related to a specific shift? Is it related to a particular product? Did you have a meeting with the inspector and operators associated with the missing features? If yes, what did they say? is the lighting system good enough for their tasks? What about their vision strength? Is it suitable for this job? Do they work the same shift? What about the workload? If you have a positive result for these questions, the problem is a behavior problem. I suggest concentrating on the inspector because he is the final stage in this process unless you have special monitoring equipment, such as an X-ray. Thank you

In my opinion, there are 3 key things that I can say about it, according to my working experience on spacecraft company:

1. Such operational errors will happen at each enterprise in which people and technology interact

2. The non-use of punitive measures for mistakes made creates the best basis for identifying local conditions that give rise to mistakes within the organization
Also I had experience of working at airplane engineering company and there I've come with such conclusion:
Airlines that are tolerant of operational errors and have a non-punitive error policy are better able to equip flight crews with appropriate countermeasures to deal with errors in flight operations

IMO, with a background in human factors, the only "human error" is that of the designers, who designed a product that could be misused.

I think of "human error" as the medical version of "It must be psychosomatic. (Because it couldn't possible be that the doctor can't diagnose.

At the very least, you should have a clear definition of "human error."

While not related to the "assignable root cause is a person", I want to use this piece of insight to emphasize a point I make in other sub-forums:

outdoorsNW said:

The big problem is PCB design is rarely taught in colleges. I checked several local engineering programs along with Stanford and MIT and only found one elective class at one local school. Most designers only care about the electrical functionality and not DFM. Many frequently ignore the IPC design guides in the 222x series of IPC documents.

Click to expand...

This assessment (which I agree with) is precisely why I think it is foolish to abandon the concept of pre-control risk assessment (i.e. "left side") of Detectability in Design Failure Modes and Effects Analysis. When engineers apply standardized design techniques, including those from "classroom/laboratory learning" or "enforced design from a CAD package", I 100% think that these sorts of practices go into improving the Detectability of some types of failure modes with the possibility that those failure modes may not require additional risk controls on the "right side" of the analysis. Verification of Effectiveness would still be required. Ad hoc engineering practices deserve poor "left-side" ratings for detectability of failure modes, and as a corollary: standardized application of good engineering practices deserve better ones.

I am disgruntled that so many folks have (my words) thrown up their arms with a statement of "we don't even known what detectability of a design failure mode would mean!" and thus want to sacrifice a perfectly useful rating scale that can actually drive improvements in design efforts. If we don't expect excellence we won't get it. I see tossing out detectability from DFMEA as basically saying that there are no effective design techniques and we have no choice but to test quality into the design.

Tidge said:

I am disgruntled that so many folks have (my words) thrown up their arms with a statement of "we don't even known what detectability of a design failure mode would mean!" and thus want to sacrifice a perfectly useful rating scale that can actually drive improvements in design efforts. If we don't expect excellence we won't get it. I see tossing out detectability from DFMEA as basically saying that there are no effective design techniques and we have no choice but to test quality into the design.

Click to expand...

I think there are a couple of factors at work here. First, the traditional AIAG DFMEA form is far from being the ideal instrument for design purposes, and many make the mistake of conflating the form with the actual DFMEA process. The map is not the territory. The other issue is that it's rare in my experience for design authorities to include manufacturing people in the process, so the process becomes a self-fulfilling prophecy. One more factor is that (again, in my experience) electrical/electronic design engineers tend to be beset with a god complex and very fragile ego. You'll play hell trying to get them to admit a mistake or shortcoming in their work. When the perfect storm occurs and all three of these things come into play at once, nothing good can come of it.

I second Jim's comment about electronic design engineers being unwilling to admit their work has flaws or could be improved.