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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


B.I.C. SMT Process Controls

delnosa

#17120

B.I.C. SMT Process Controls | 7 August, 2001

Looking at ideas for "best in class" process controls for SMT mfg. ops. Wanting to add improved p-controls at some of our suppliers and I'm looking at what others may or may not recommend. Already looking at the following: Solder Printing X-Y-Z axis measurement Stencil tension Stencil thickness Placement X-Y-theta measurement

Are there sites where this can be found ? Thanks.

This message was posted via the Electronics Forum @

reply »

Michael Parker

#17128

B.I.C. SMT Process Controls | 7 August, 2001

In addition, consider:

Solder paste selection, including clean or no-clean is also quite important. This can vary from application and is somewhat process dependant (what works best in the environment).

Relow profile. A must for many reasons.

And foremost, the design itself, does it lend well to the processes (DFM)?

Do an archive search of this site on each of the topics to glean "BIC" practices. BIC happens!

reply »

#17134

B.I.C. SMT Process Controls | 7 August, 2001

Sowaza BIC? Ru the ball point pen people?

Know no sites, like thet, but let�s see what we can do here. Wedabomb.

First, rather than defining the controls your suppliers must use to produce your product, why not specify a DPMO rate er suppin like thet and not get involved in micromanaging their process. * If they produce good product, good. * If they don�t produce good product, then you should complain and they should describe their plan to correct things. Yano, that's why the US military go out of the specification business. They realized they could either tell their suppliers: (1) give us good product OR (2) here's how to manufacture good product. NOT BOTH.

Second, let�s talk about each of your proposed performance measures ... * Solder Printing X-Y-Z axis measurement. Yagot me. Let�s see, is: �X� is the stroke. �Y� is the blade width. �Z� is the ah, umm, er � Listen here's the issue. How to get a BIG honkin� brick of paste on the board [of the correct volume and in the correct location] and have it stand ready to get its face smooshed-in by a component termination. So, it�s reasonable for you to want to understand your supplier's process controls used to print the correct volume of paste. Then when you see too little solder on your boards, wammo you can ask them why those controls went south and how they plan to prevent recurrence. * Stencil tension. If printing with a zero snap-off, which virtually universal today, stencil tension becomes a very obscure measure of success. We spoke about this in the archives a while ago regarding frameless stencils. The answer was ah, umm, er, so enlightening that you should read it yourself. * Stencil thickness. Are you concerned about the stencil wearing-out from too many prints? Most stencils are stainless steel. They aint gunna wear out. Chase something else. * Placement X-Y-theta measurement. Don�t know what this means. Among measures people use to help qualify the purchase of placement machines are placement accuracy and Cpk. But you are not buying machines. Listen here's the issue. How does your supplier put component terminations close enough to the pads so that when the solder melts, it sucks the component onto the pads and looks perfect. In real life, it�s reasonable for you to understand placement machine Cpk, like any other shop equipment calibration measure. Then when you see lousy component placement, wammo you can ask them why Cpk went south and how they plan to get it back on track.

Finally, some assemblers use designed experiments to determine which of the plethera of factors in a process have the greatest effect on proper output. Consider that your supplier may already know these critical parameters for each of their process and are in a good position to describe how they manage their process. Then again, they may be a dirt floor sweat shop in LA [ Lower Alabama] and they are grateful for every smidgen of knowledge you provide them.

reply »

delnosa

#17143

B.I.C. SMT Process Controls | 8 August, 2001

Understand your advise, but don't know if you've dealt with Asian (Taiwan, China, Malaysia, etc.) motherboard suppliers in the past. They have some of the most technically advanced equipment and some of the most challenging products to produce, but unfortunately process control doesn't appear in most of their vocabularies.

The problem isn't us telling them what their targets should be, it's them not hitting them and not showing the capability to make substantial improvements. (I won't go into the politics of changing suppliers, etc., you likely know that already) Believe me, I don't want to micromanage their processes, but my management direction is basically that. When you walk on their shop floor and don't see any sort of control indicators (charts, plots, etc.) this either indacates their real good, or they don't have any idea what to do. It's usually the latter...

Fundamentally I'm trying to get them to look at the process and understand the input variables and the effect on the output. Part of the purpose of taking some of the measurements is to instill the discipline to monitor variables. Most of the data being collected is to establish both Cpk and longer term Ppk, which again is unfamiliar territory to most of them. Some of the measurements are to build correlation between variables and determine the net effects. Even if the results prove nil, the point is to dig deeper and make them understand cause and effect.

Getting them in the mode of looking at variables and determining if they are meaningful or not is the real intent. Once they grasp some of these basics we move to DOE's, etc.

Thus I'm looking at other examples of what others are doing to both monitor & control their processes.

This message was posted via the Electronics Forum @

reply »

Michael Parker

#17144

B.I.C. SMT Process Controls | 8 August, 2001

Your reply with the reason for wanting the BIC data finally exposed the root cause of your problem. Good luck with trying to educate your suppliers. I have had copiuos amounts of experience with Asian cultures. First, you gotta know that they have a million words (in varying degrees) for yes and nothing for no. So when they say they understand, don't just believe it, act like you are from Missouri and say "Show me!". Once you learn to get past the cultural thing, you may have a chance. They are very traditional and resist all efforts to change. They will continue to hold a left handed hammer, backwards, in their right hand to pound nails rather than get a nail gun and do it faster and easier. If you want it done right the first time, come see me, I manufacture in the US of A. (We puts creosote on our dirt floors here, keeps the dust to a minimum!)

reply »

#17153

B.I.C. SMT Process Controls | 8 August, 2001

OK, so yer stuck wid dees sembly guys, yer bosses are moroons wid pointy heads, and you gotta fix it. snafu.

Paste printing: Measure shape, height, volume, area, registration, and paste bridging. Inspect/control solder paste deposits for fine pitch & 0402s. Component placement: Inspect/control placement position for fine pitch & 0402s. Reflow soldering: Monitor time over liquidous.

Do paste first, reflow second, and placement third.

reply »

Sean D

#17156

B.I.C. SMT Process Controls | 9 August, 2001

Hello everyone,

First when defining any criteria for process control, one should identify quantiatively that criteria. Then one establishes an effective method for measuring such a criteria. Once you have a test method, your gateway to a solving your production issues through data analysis is open.

The balancing act is going to be implementing test methods in a cost effective manner.

1. With all of your screen printing concerns, try offline or online paste inspection, the resulting data should better assist you in determining where your issues really lie.

2. For Placement of components you can implement on line placement inspections both before and after reflow.

a. the pre-reflow inspect will help to id issues in Pick and Place. b. The post reflow inspect will help to id issues in your reflow profile.

These inspections can be performed manually by operators qualitatively, but quantitatively, that's another question. Not to mention with repeatability. We all know the concerns of operator errors/training, etc.

If one steps back to examine the big picture, ie by inspecting the sum result of each process, you can then work your way back to the root causes for each issue.

There are companies that can provide these inspection tools. Personally, I'm a CyberOptics representative but I won't press the group here with any bias (Best Best Best Best Best.....woops sorry!).

One question to examine is if you force these quantitative criteria onto your supplier, what kind of costs will you eventually see come back to you as a result of establishing these new parameters?

Good luck with your suppliers,

Sean D.

reply »

IPC Training & Certification - Blackfox

One stop service for all SMT and PCB needs