A better Lot/bin system

[Hybrid]

Hey y'all,



I am a lone in-process inspector who does inspections for 20 machines (stamping, CNC, screw machine)

Operators are suppose to give me samples every hour. Problem is that some do give every hour and some give every 3-4 hours. We cap bins at 2,000 pieces. I maybe check about 100 pieces of a bin before it sent for cleaning and onto final inspection. In final inspection they will find visual defects that I do not find. (Hate to sound cocky but I am real good at my visual inspections per JEDEC spec)

This has recently caused management wanting me to sign off on each bin plus inspections before hand. So I am inspecting samples until a bin gets to 2,000 pieces, plus inspecting a AQL of the entire bin before going to cleaning. Anyone have tips/ideas on how to improve this? I already have a meeting with management to let them know that this won't work and will only cause bottlenecking of each bin.


This makes me feel like I'm not doing my job enough.

 

[John Predmore]

I have a few thoughts based on what you wrote. You did not describe the parts or what kind of defects, so I must use my imagination. Tell me if this is helpful.

I accept your statement that you are good at inspection. Yet, final inspection finds visual defects that you "did not find". How could this happen?

One possibility is the parts changed after you inspected them, perhaps defects were introduced in the next operation, which you said was cleaning. Could the parts be damaged in the cleaner? Are the visual defects something like scratches? Does the cleaner machine have any rough edges the parts pass over?

Another possibility is the parts are damaged in transportation. Do the parts have sharp corners that can scratch other parts? Are bins cleaned before parts are put in? Are the defects caused by heavy force? Are the parts with defects found on the bottom of the parts bin? Are parts bins stacked one on top of another anywhere downstream of inspection?

Another possibility is the visual defects are hidden from view until the parts are clean. Do you clean the parts you inspect?

Another possibility is there are only a few defective parts and you inspect a sample, so the probability is low that you select a part which has a defect. Does final inspection inspect an AQL sample at the same sampling level that you use, or do they do 100% inspection? One idea is to mark every part you inspect with a paint pen or a prick punch in an inconspicuous location. Then examine the defective parts found by final inspection. This experiment would tell if final inspection looked at the same parts that you previously inspected

 

[Tagin]

In addition to what John Predmore write, I'd only add that since it is apparently a visual inspection: are you and final inspection both "calibrated" to the same visual criteria?

I.e., do you agree that the defects that final inspection finds would also have been labeled as defects by you? Or are the two inspections using different criteria?

AQL means “quality level that is worst tolerable”, and is no guarantee of 100% absence of defects in the lot, so it is unsurprising that final inspection found some defects. It is more important how many defects final inspection found, and whether that defect rate exceeds the realistic expectation of remaining defects resulting from your inspection, based on the AQL level you use.