Visual Inspection

Six Sigma – iSixSigma Forums Operations Manufacturing Visual Inspection

Viewing 8 posts - 1 through 8 (of 8 total)
  • Author
  • #55982


    Dear Users,

    I am currently partially responsible for the inspection of produced goods and have an unusual problem to solve – it will be probably one Six Sigma Project. The inspection department (visual inspection) checks the produced parts and the all parts. Those which fail, are scanned with the problem description (into ERP) and are moved to a certain place in the department.
    All these parts are later being checked by one engineer, who makes the final decision, whether the parts need to be reworked or if a complete new part is needed.
    So far so good, but round about 70% of these parts are passed from the engineer as good ones.
    There are almost zero customer claims (2-3 in a year) so it means that the engineer and the inspection department do a really good job.

    It seems like the stuff in the inspection department needs to be additionally trained. But every part, which they exclude looks to be out of tolerance / norm (visual inspection).
    There exists a β€˜β€™grey zone’’ and it looks like to me like following:
    β€’ Everything within the tolerances / Norms
    β€’ Out of tolerance – but still somehow acceptable (from engineer’s experience)
    β€’ Out of tolerance

    I was wondering to teach the stuff in the inspection department, how to evaluate such a critical parts to be consider, but on the other hand to teach them, to send the parts out of tolerance out. I can understand their feelings, to do not risk any customer claim and do not have such a responsibility as the engineer. The invention to change the situation is obvious – to shorten the throughput time and costs for the engineer.

    I am aware, that In this kind of problems opinion of inspection people is very important, periodic meeting discussing action plans and results help a lot, give confidence to them. I can imagine, to work together with an engineer by evaluation of suspected parts to gain a sense to make a decision. One of the other option could be providing a board with Go-No Go samples, but I am still somehow confused and would like to know any proposal, how to manage this situation.




    Rather than looking at ways to improve the inspection process I’d ask you to consider that inspection is wasteful. It doesn’t change the product. It just finds defects that shouldn’t have occurred. It’s much better to reduce if not eliminate the need for inspection by improving the production process to prevent defects. Then and only then can you reduce inspection to sampling, to verify that the production process is operating within specifications, as opposed to inspecting every part.


    Mike Carnell

    Be very clear there is no such thing as 1005 visual inspection. That is called sorting and as Strayer said it is a waste. You don’t trust your process because you don’t understand it and consequently don’t know how to control it. That is what you need to work on.

    Inspection exists to find defects. If they don’t there is no reason for it. Therefore no mater how good you make your process you will still get rejects because nobody wants to lose their job. Chart it. Chart the number of rejects and tag improvements. rejects drop and then drift back up. A self full filling prophesy.

    That engineer needs to get to work cleaning up and mentoring the inspectors and you need to learn how often they need to be retrained to keep them on track. I went through this whole thing an a Navy Weapons Spec program in the 1980’s. Everything was not only inspected but it was inspected under a microscope. Get a copy of Mario Perez Wilsons book called Six Sigma. He documents how it was handled there.


    What Mike said is correct. You need to keep visual inspection running until the manufacturing process will be improved to ensure customer requirements. It makes good sense you go working on process making the parts. The time you will spend on doing will be adding value to your company but the one you are spending on visual inspection no, it is simply a waste.


    Matt Townsend

    We went through this problem a while ago. We were spending a great amount of time and effort on trying to increase the throughput of our inspection process to ensure quality of our parts.

    This was a GIGANTIC waste of time.

    We are now using the defects we find to fix our process to eliminate the errors from occurring in the first place. It really is the most effective measure to increase throughput throughout your entire plant.



    Just as an aside, for those in the IT industry: Rigorous testing, which is often more than half of software development time and effort, is mostly waste. We have to do it because our development process is incredibility error-prone, from requirements, to specifications, to design, to coding. This is why agile development methods have become so successful. With agile, incremental development we trap errors almost as soon as they occur. It incorporates poka yoke into the process.


    Anyone have experience with torque marks across fittings and bolts? Fittings are currently my biggest concern as per manufacturers recommendation should be made up until wrench meets resistance, marked with a line across the fitting then wrenched a certain number a flats past the initial mark. Having to stop making up the connection to mark the fitting, keeping track of the pen to mark, knowing the correct amount of flats to go past for different size fittings all create issues.

    Just wondering what others have done to solve this issue.



    Oopps thought I was up a level in the forum…sorry disregard.

Viewing 8 posts - 1 through 8 (of 8 total)

You must be logged in to reply to this topic.