Why did our accepted cover become three scratch examples?

LuisBrown0956 · 5 Oct 2025, 15:53 UTC

Closed
LU
LuisBrown0956
I opened the UR5e camera trial folder and found three reject images of the same polished cover. The bright line is in a different place each time. Quality accepted the physical cover, but the image labels were added by somebody looking only at the screen. We need a useful comparison set before tuning starts.

15 replies

MI
MinaArcher0424
Replying to LuisBrown0956

Retain all three views with their setup details, but remove the unsupported defect labels from the tuning set. Can quality identify the exact cover and confirm its present condition, rather than relying on an earlier acceptance?

17 points
HA
HarishBrooks0837
Replying to LuisBrown0956

I'd put the actual cover under the agreed inspection light before ordering anything. We nearly bought our way out of a reflection nobody could find on the part.

10 points
FA
FarahAli0243
Replying to HarishBrooks0837

Harish, yes, but a genuinely marked part can be hard to see under the wrong light too. I would not let one quick glance replace quality's viewing method.

21 points
HA
HarishBrooks0837
Replying to FarahAli0243

That's why I said agreed light. I'm arguing against a new camera being the first move, not against doing the physical inspection properly.

13 points
VI
VictorAli0185
Replying to LuisBrown0956

Add confirmed subtle rejects to the comparison. A perfect cover and an obvious gouge make a very forgiving test; the faint mark near the viewing limit is often where the proposed image stops being useful.

24 points
LU
LuisBrown0956
Replying to VictorAli0185

Quality rechecked the identified cover and still accepts it, Mina. Harish and Farah, they used the written viewing method. Victor, they've added a faint scuff reject. The same holder lets the accepted cover sit two ways, which we also need to fix.

18 points
JO
JonasBrooks0807
Replying to LuisBrown0956

Which surfaces locate it? Two seats suggests a holder question before a lighting purchase.

18 points
RA
RaviCarter0998
Replying to LuisBrown0956

Keep the awkward seats documented. They explain why the old examples disagreed.

16 points
FA
FarahAli0243
Replying to RaviCarter0998

Ravi, would you keep those in operator training once the holder changes? I worry about teaching people to accept a seating problem as normal variation.

5 points
RA
RaviCarter0998
Replying to FarahAli0243

As labelled diagnostic examples, not the new acceptance references.

23 points
VI
VictorAli0185
Replying to FarahAli0243

Farah, I use that sort of pair to show why seating matters, separately from the pass and fail examples. It works only if the distinction is obvious in the training material.

12 points
LU
LuisBrown0956
Replying to JonasBrooks0807

Jonas, the temporary holder located against a loose edge strip; tooling has replaced it with a checked locating arrangement. Repeat images are more consistent. The faint scuff still needs a better lighting comparison, so thresholds remain undecided.

9 points
HA
HarishBrooks0837
Replying to LuisBrown0956

That's progress without pretending the whole inspection is done. Did the holder change make loading more fiddly? A camera trial can miss that when the designer places every cover like it's jewellery.

20 points
MI
MinaArcher0424
Replying to LuisBrown0956

Luis, who owns the reference covers after commissioning? They need condition checks and replacement rules, especially the faint reject everyone will want to handle during training.

24 points
FA
FarahAli0243
Replying to VictorAli0185

Victor, thanks for explaining the diagnostic pair. I can use that distinction in our changeover training without mixing old holder behaviour into the new acceptance examples.

7 points

Discussion closed

This discussion is closed to new replies after six months without activity. Last activity: .