How detect visual scratches on mirror glass surface?

Six Sigma – iSixSigma Forums Old Forums General How detect visual scratches on mirror glass surface?

This topic contains 15 replies, has 7 voices, and was last updated by  visvet 11 years, 6 months ago.

Viewing 16 posts - 1 through 16 (of 16 total)
  • Author
  • #47041


    We have problem with certain detection (100% operator’s visual inspection) of scratches on mirror glass surfaces (outside mirrors for cars) on assembly line.
    (mirror glasses are input material from external supplier)
    No each operators successfully detect those deffects!
    Have you any experience about those mechanical deffects (scratches) on glass surfaces, please?
    (e.g. any special light lamp, filters, etc.)
    Toyota customer dishonours any visible scratches on complete mirror glasses…
    Remark – those rejected scratches are very hard visible (we have light 500 lux and distance 0.5 meters/about 20 inch from eyes).
    Thank you for any suggestion.



    I need advice for easy detection of those scratches on mirror surface – on top of mirror glass is very sensitive Crome layout – it is very easy scratch it…
    Sometime those mirror glasses (intup material) are dirty/dusty, so visual detection is harder…
    any suggestion?



    Use  simple  Chek-sheets



    it could be your operator not well trained, thus not being able to catch defect accurately ….it could be the environment issue (dust count particle need to control to prevent from being easily dusty and make inspection even more difficult) ….it could be due to lack of inspection kit ….or why bother inspecting? why not preventing defect from happening at the first place? …..etc
    The best solution for your problem is to start doing your own DMAIC project on this visual defect improvement. Good luck!



    Why don’t you do what Toyota does? Reject scratched glass to you supplier. This is the simplest way.
    Or go to you supplier, start a DMAIC project to find the ROOT CAUSES of scratched glass. I see it as this should be your supplier’s project. Try to find if it is scratched due to production, storing (you mentioned dust), transportation, manipulation…
    I believe improving inspection is not the way out of your problem.
    Take care



    You can build your own verification system.
    You’ll need a single, bright light source mounted inside an open box with the inside of the box painted black, so you can observe bright defects against a dark field.
    You should check the incoming glass for polishing defects, such as scuffs and digs, and also for handling damage – by far the most common cause of glass defects.
    You’ll need a different system for checking post chrome deposition.



    OK, there are two faces to this issue:  1st s the process-focus.  2nd is the mechanica/technical solution.
    Process Focus:   Do you and the Operators know what the actual/measureable optical standard is?   What is the tolerance?  If none exists and it’s all visual/gut-check, then you may want to conduct a VoC/QFD with your Customer to try and quantify pass/fail criteria to help mitigate within/between subgroup variation amongst Operators, Inspectors and Customer incoming Inspection.   Bottom line, what is the standard to which you are held and are your folks adequately trained and certified to meet this standared?
    You can purchase a special refracting gauge.   These are used in Optical environments for checking several things including clarity (polish), concentricity and surface damage.   try googling optical supply houses.
    Basically, you have a black-lined cabinet with a bright lightsource.   You take the gauge, in this case, a “tuned” prismatic lens made from glass/quartz round stock.   You place the gauge on the part being inspected.   any anomaly in polish, concentricity, or surface topography will be come immediatley apparent.   it will take the form of warped circular lines in a field.  
    Again, try Optical supply houses, or try benchmarking Optics shops.
    Google is your friend.



    Hello Satan,
    what do you mean “simple Check sheets”?
    PLS, can you describe it in detail?



    Hello Ferry Waskita,
    thank you for yoyr advice.My dream is eliminated any scratches and eliminate visual detction, but those way is very hard (maybe impossible).
    Our external supplier (Germany) of mirror glasses supply to us glasses in small packages filled in plastic foil (due to prevent dust pollution), but when we open those packet (25 glasses), mirrors are dusty with finger marks!
    I think, root cause are operators in extern.supplier, but supplier refuse it…
    I’m affraid, our SQA deppartment have contract without clause “mirrors surface must be without finger markes and dusty…”For specify:1. When glasses come from supplier, our incoming inspection make statistic visual inspection of some random selected glasses from all batch and storaged it.2. After glasses goes to sub-assembly station (stick mirror glases with heater foil and press it on plastic holder). Glasses are placing into plastic box (capacity 90pcs) separated paper (insulator paper prevents generation scratches from other one). Visual inspection for scratches is 100%.
    3. After pressed glasses are stocked again.4. After warehouseman deliver those mirror glasses to assembly line – operator insert one to the complete external mirror glass for car. Again – 100% visual inspection for scratches and chips.
    5. On last station 100 % quality wall operator checks complete mirror for any deffects (one of all is visual check of scratches on mirror surface again) => the 3th 100% visual inspection of scratches on mirror glass!!! Inspection = Non value added activity.It’s incomprehensible for me – I want change it, but I need find out the best solution.I need to eliminate any scratches in source and keep it non scratched till dispatch to customer.
    Thank you for advice.



    Hello Silviu,
    thank you for interesting comments.Yes, reject scratched glass to you supplier is good idea, but supplier claims those our rejected mirrors – he say that mistake is on our side (in our company).
    Now I gather data from each batch of glasses (I need foud out were scratches arise) – from incoming inspection to last quality wall station before dispatching.
    But I’m not sure our operators well distinguish good mirror glass (without any visual defects = scratch) and reject parts (with visual scratch error) on all stations (2 stations on sub-assy line, 1 station on assy line, 1 station on last quality wall station of assy line).
    Therefore we need to develope any “black box for reliable and simply detection of visual deffects” by each operators equally! Now we have normal fluorescent tube lightning 500 lux above operator only.It’s BIG PROBLEM when operator of assy line found scratch on glass, but for operator of sub-assy line it is good product!I think we have very strick criteria for visual detection of those scratches – during last 6 month we had 1 complaint from Toyota only. Operators looks for scratches instead of simple detection (pass/non pass).
    We tried make simple Gage R&R test (20 mirrors) with OK & fail mirrors and operators long looked for defects on good mirrors.
    Repeated G R&R was worse like previous one dou to boundary samples (big indecision=random result).
    Next week I’ll try change internal packaging between sub-assy station and assy line (eliminate separated paper between glasses).
    Thank you for new advices.



    Hello Photon.Thank you – your advice is very interesting (black box inspection box).
    Could you describe it in more details, pls?What dimensions of inspection box, when mirror has about 18 x 12 x 1 centimeters? (divided by 2.54 for inch)
    And what kind of lightning (we need intensity lightning 500 lux and distance mirror from eye 50 centimeters). I think type “D65”  (6500 kelvin) is suitable, isn’t it?Optics and visual deffects on glass is new for me.
    What do you mean there: “You’ll need a different system for checking post chrome deposition.”
    Our mirrors = glass + thin and sensitive crome layer on glass (TOP)
    thank you for any advices.



    Hello HBGB B^2 (good nick),
    your 2 assumptions are 100%!
    First = yes, our operators have mishmash (hotchpotch) between OK and not good parts!Customer isn’t helpful – he strictly say: mirror glass surface must be any deffect free…
    We had 1 complaint during last 6 moths only, so I think, we have very strick criterias for rejection of parts by operators (they looking for deffects intead of detection).Unfortunately – I have new quality enginer (he is absolute new) for this project (previous left 1 month ago), so communication with customer is very difficult…SECOND – a black-lined cabinet with a bright lightsource is think what I need! But optics is new domain for me.
    What do you mean “tuned prismatic lens” from glass round stock?Could you describe it in more detail, pls?What dimensions of inspection box, when mirror has about 7 x 4.7 x  0.4 inch?
    And thank you for your google phrase – I try find out more information about industrial solution.Thank you again for your suggestions.


    Heebeegeebee BB

    Heebeegeebee BB

    go here,
    scroll half way down the page…look for inspecting for flatness.


    Heebeegeebee BB

    The box can be any size you want…whatever is comfortable for the operators to use…we were inspecting quartz sensor wafers/diaphragms that ranged from 1″ Dia up to round glass wafers over 12″ in Dia…Our lightboxes were 3 ft cubes minus the front panel for access/observation mounted with a fairly powerful white-light source…   a pair of 100 watt light bulbs are just fine…



    Hello Heebeegeebee BB!
    Thank you for hyperlinks.
    Super – I found this info. It seems you worked in this area…
    I have to read your hyperlinks (during weekend) and I write you my potential questions.
    If you have any more info about this problem, thank you for your time for sent it to me…

Viewing 16 posts - 1 through 16 (of 16 total)

The forum ‘General’ is closed to new topics and replies.