Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Inspection Methods

#7318

Inspection Methods | 21 July, 2001

I was reading something in the paper this morning where the police have found that having a witness look at six pictures of suspects laid-out on a table [the six pack method in their parlance] is more error prone than just looking through the stack of pictures on at a time, like we'd do, if we were looking at Paulette's daughter's prom pictures.

That got my mind ta wurring. Our inspectors look at multiple boards peering at the portion of the boards as they sequentially proceed for one board to the next. How do your inspectors look at boards? What other novel techniques do they use?

reply »

mugen

#7321

Inspection Methods | 23 July, 2001

Dave. F.

Hmmmzzz....

Thats interesting, coz as a contract manufacturer, we try to optimize processes asmuchaspossible.... simple...yeah! that's the word "Simplify"....

We found that if you layout a production line, assume x10 operators, and one single product, we split it into x10 segment steps, and mass production run this single product, through all those x10 steps, each step-station completing their own process, and handover to the next flowdown step-station, its bit like "pass-the-parcel" game....

Now, we ask train & ask, each of those operators to individual, take each product unit, continually proceed through the whole processes, each operator taking the product through the entire flow, and each operator taking responsiblity for their own units, guess what?

The process time is shorter per unit timing, the operator builds the product unit, and caters for its conformance to specifications, and feels ownership for each individual produced unit...

No more : " it an'it my prob dude, it was the guy before me that smuckedup the-wats-it-call-it thingyabob that you are now barking at me for screwing-it-up....mommy!"

Bottom Line:

Test this out, if your boss has reservations, line out the guys, into two teams, folding paper squares, one line with process-steps broken down, and with one operator doing one folding action, and last operator placing each x4 corner-fold square, into a packaging box..... and, the other line, with each operator folding all corners, and then packaging into a box, after which each operator repeats the cycle, try, this out within a time frame, say 5mins, and see which team has the most correctly & neatly folded corner squares?

To all my fellows out there, consider this line arrangement, I would be most interested to find out your data findings.

Cheers :)

reply »

mugen

#7322

Inspection Methods | 23 July, 2001

Dave. F.

Hmmmzzz....

Thats interesting, coz as a contract manufacturer, we try to optimize processes asmuchaspossible.... simple...yeah! that's the word "Simplify"....

We found that if you layout a production line, assume x10 operators, and one single product, we split it into x10 segment steps, and mass production run this single product, through all those x10 steps, each step-station completing their own process, and handover to the next flowdown step-station, its bit like "pass-the-parcel" game....

Now, we ask train & ask, each of those operators to individual, take each product unit, continually proceed through the whole processes, each operator taking the product through the entire flow, and each operator taking responsiblity for their own units, guess what?

The process time is shorter per unit timing, the operator builds the product unit, and caters for its conformance to specifications, and feels ownership for each individual produced unit...

No more : " it an'it my prob dude, it was the guy before me that smuckedup the-wats-it-call-it thingyabob that you are now barking at me for screwing-it-up....mommy!"

Bottom Line:

Test this out, if your boss has reservations, line out the guys, into two teams, folding paper squares, one line with process-steps broken down, and with one operator doing one folding action, and last operator placing each x4 corner-fold square, into a packaging box..... and, the other line, with each operator folding all corners, and then packaging into a box, after which each operator repeats the cycle, try, this out within a time frame, say 5mins, and see which team has the most correctly & neatly folded corner squares?

To all my fellows out there, consider this line arrangement, I would be most interested to find out your data findings.

Cheers :)

PS : its has no direct linkage to Dave F's mind-probe, or does it??? *nyak* thin line baby....

reply »

#7347

Inspection Methods | 23 July, 2001

We are optimizing our process, it's just that we build to a daily or semi-daily demand schedule of product quantities of 50 er so.

reply »

Steve

#7359

Inspection Methods | 24 July, 2001

Interesting concept. Are the PCB's still coming off of the SM line one at a time, or are they being batched? It also seems as though you would need more duplicate tools since most stations are tool specific.

It does make sense though because whether they are built in a line or not, they will still take the same amount of man-hours (sorry, person-hours). Also, it would create a sense of ownership.

Thanks for the idea.

reply »

Steve

#7360

Inspection Methods | 24 July, 2001

Interesting concept. Are the PCB's still coming off of the SM line one at a time, or are they being batched? It also seems as though you would need more duplicate tools since most stations are tool specific.

It does make sense though because whether they are built in a line or not, they will still take the same amount of man-hours (sorry, person-hours). Also, it would create a sense of ownership.

Thanks for the idea.

reply »

Steve

#7361

Inspection Methods | 24 July, 2001

For SM inspection, AOI is the only way to go. Per component inspection, it's the best.

For in-line operations, split the inspection so that each sequential operator is inspecting the previous stations work. This splits the responsibility and is easier to manage.

reply »

Lori

#7378

Inspection Methods | 25 July, 2001

I have been re-evaluating the way we do our visual inspection process due to its apparent ineffectiveness. We recently tested this by introducing 6 obvious defects on one assembly and gave it to one of our "esperienced" visual inspection operators. 2 of the 6 were detected.

We have investigated AOI. However, I am looking for articles or texts that cover HUMAN visual inspection techniques as well as any other successful implementation stories from this forum or people you know.

Our shop makes about 1100 unique assemblies with build quantities ranging from 2 to 500 at any one time. Material flows into each "work cell" as it is made (not a traditional "batch" process).

Any reading or industry expert recommendations?

Thanks!

reply »

Lori

#7379

Inspection Methods | 25 July, 2001

I have been re-evaluating the way we do our visual inspection processes due to its apparent ineffectiveness. We recently tested this theory by introducing 6 obvious defects on one assembly and gave it to one of our "experienced" visual inspection operators. 2 of the 6 were detected. Experts have written that depending upon several factors, visual inspection is only 10% to 50% effective....

We have investigated AOI (analysis pending), and currently employ x-ray as well as in-circuit tests. However, I am looking for articles or texts that cover HUMAN visual inspection techniques as well as any other successful implementation stories from this forum or from people you know.

Our shop makes about 1100 unique assemblies with build quantities ranging from 2 to 500 at any one time. Material flows into each "work cell" as it is made (not a traditional "batch" process). We tend to make the "crap" that no one else will build - high workmanship content.

Any reading or industry expert recommendations?

Your input would be welcome. Thanks!

reply »

Steve

#7384

Inspection Methods | 25 July, 2001

AOI is very cost effective when you are dealing with that many assemblies. Once you do an ROI, you'll find that it probably will pay for itself within 12 months. The reality that many managers can't bring themselves to do is that if you bring on AOI, you will need to layoff several people.

AOI provides 100% coverage, is repeatable and consistent, is orders of magnitude faster than a human, and doesn't need to take a day off.

reply »

#7424

Inspection Methods | 26 July, 2001

Dave, We don't do any inspection. Before everyone says "bull...t". Let me explain. We are an OEM. I spent six years in subcontractor hell and wouldn't go back for all the tea in China. It took me two years to convince them when I first got here but I finally succeeded. Manufacturing gets final approval on design. Therefore I'm not continually trying to turn dog...t into honey. Every thing we do is RF, so at some stage it has to be finetuned before it goes into the box - this is were we find our faults. Pass rate hovers around 98% straight off the line. We work a 38hr week - no shifts. SM operators are one person - one line (all lines have two or three p & p machines). They print, place, reflow, stock control, paperwork, maintenance, the whole she-bang. I'm just starting to get some of the senior ones into programming and stencil design. In the end (I hope)they will look after "their" product from prototype discussions through to release for manufacture. I don't use passwords on anything and they all have access to the PCs etc. This is were the "ownership" and "built in quality" come from.( Go that feel good jargon!) The most important thing is to get good people who care and understand about what they're doing. And no, in case anyone is wondering, there are no jobs available at the moment. I guess my reply doesn't help you to fine tune your inspection - it's just the way we do it.

reply »

#7431

Inspection Methods | 27 July, 2001

Hi Darby,

you must have read my thoughts about the stuff we are dealing with. Think I�ve got to hide your message from my boss it may lead to the decision that I�m not allowed to read this and other forums anymore ;-). We still need optical inspection for all the outsourced things and believe me, it�s necessary, we do not have the best design, they do not have the best people :-( !!!

Useless to say that manual optical inspection is time intense, inspector dependend, and not detecting all defects. We concentrate on differnet techniques

1) look at general appearance - most obvious unsoldered ,shorts, wrong mountings are detected (needs a very good experienced eye and should be done only by the most experienced people you have who do the assembly job all day themselves and know about the standards)

2) look at specific critical parts with a microscope to manifest those first impressions (checklist)

3) are than (after a couple of boards) limited in our view to the most obvious symptoms that other things just slip through

4) have hot debates what we may accept and what is send back ( we hate to send it back, the same people do the rework that made that b......t first, do you think they can do that job better ? )

5) put up a meeting to discuss alternatives and cancel all reasonable measures due to NO TIME, TO EXPENSIVE, 6) cut the price for the delivered goods and everybody ??? is happy.

Now, IMO the best optical inspection takes place while doing the work and reacting immediatly to any deviation from the standards and that calls for operators and a workaround like you described. In other places differently organized and really mass production this won�t work and optical inspection seems to be a must to keep test engineers from getting bezerk.

For Dave:

The most important step in manual optical inspection is step 1). I can�t describe the way how to do it, not everybody is a "perfect witness" , you need to have the ability I would call the "unfocused look" ( for example: if you look at a star directly you won�t see it because it�s focused in the so called blind spot of your eyes, you have to focus a bit of target to get the image). Spot checking under the microscope after a checklist with pictures of good and bad at your side isn�t that challenging but boring, tiring and after some time questionable.

Remember, important things always happen in the corner of the eye and from that perspective they can be detected best.

Wolfgang

reply »

Lori

#7436

Inspection Methods | 27 July, 2001

My engineering couterpart has the task to evaluate AOI and we will be receiving one next month. In most of the articles I've read, AOI is deemed effective post paste and post pick-and-place (pre-reflow), but there are rarely discussions about post reflow inspection effectivity.

Here are the questions we are investigating with AOI:

How effective for inspection of post wave operations? We realize it detects component presence and polarity, but what of solder joint presence - specifically hand-soldered joints and non-standard kluge-type joints? (This is necessary since X-ray is not reliable in detecting solder joint efficacy on hand-applied operations.) How much programming time and maintenence time is necessary? Is there a high false fail rate (I'm sure based on programming...)?

I believe AOI to be an effective tool in certain applications - we just do not yet understand its fit in our extrememly high mix and craftsmanship environment. We have excellent pre-reflow and reflow results. Our primary areas of concern are post reflow.

If you have any experience in evaluating or implementing AOI, I would be interested in knowing your strategy.

L

reply »

PCB X-Ray Inspection

Solder Paste Inspection