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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


performance improvement

Hi, I would like to have information about means to improve... - Jun 13, 2002 by

Add this forum to your site! Click to learn more. SMTA

myfloh

#20343

performance improvement | 13 June, 2002

Hi, I would like to have information about means to improve the stability of a SMD-Line: indeed the variation of the daily performance is too important (between 50 and 250 ppm). What is the solution to reduce these variations? climatisation, other reflow profil?... I'm waiting for your help and answers. Myfloh

This message was posted Add this forum to your site! Click to learn more. the Electronics Forum @

reply »

#20348

performance improvement | 13 June, 2002

Most SMT lines are comprised of several pieces of equipment. Each piece of equipment, the components, the plant environment, and the people running the line can have / cause variations in their operation that each can affect product quality. Consider focusing on the element of your production process that is the major source of variation.

Understand, we are a very task oriented group. In order for us to help you further, in any meaningful way, we need information about: * Defects * Machine and machine set-up * Components, board fabrication, etc * Etc

reply »

kenbliss

#20355

performance improvement | 14 June, 2002

DaveF's request for more data of course is needed to be specific, but here are a few ideas and places to start. If the SMT line is causing the defects you need to identify where on the line the problem is being caused. What we see regularly is mishandling of boards coming off the line, ie. stuffed in totes, piled in bins, put in the old style flat racks that store boards on edge on an angle. The best way is to use the tray cart method that allows for a flat surface for the boards to lay on. This eliminates flexing and microfractures in addition to virtually eliminating ESD problems as the operators are handling the tray not the board except of course when they first take the board off the line. If misplaced components are the problem are the feeders in good condition, how are you storing them, they need to be stored properly in proper feeder racks to propect your investment and ensure they work correctly, consistently. If feeder changeover time is more than 20 minutes, you need to pre-stage feeders, if feeder exhaust during a run is taking more that 20 seconds you need to be pre-staging the known needed backup feeders at the machine. These two methods will typically gain you 1-2 hours per 8 hour shift of additional uptime. For more information on handling go to http://www.blissindustries.com and click on Bliss University.

I hope that helps

reply »

myfloh

#20359

performance improvement | 14 June, 2002

It is a FUJI Line with the machines: GP641 GL641 CP6 IP3 the oven is the smt Quatro peak. Moreover we have an AOI.

This message was posted Add this forum to your site! Click to learn more. the Electronics Forum @

reply »

#20363

performance improvement | 14 June, 2002

Ken,

With all due respect, you need to quit preying on these new guys. I mean cummon man........

reply »

Mark J

#20364

performance improvement | 14 June, 2002

The first question is, is this a high mix, low volume environment or vice versa? If it's the former then Kens solution is something you might want to look into. Depending on your feeder set up you might check to see if your operators have been trained to properly to use the CP6 Device table modes for change overs, if your budget allows you should get an extra set of MFU's for your IP3, that will also speed up change overs. You might also want to look at the line balance and program optimization of your placing programs. If you're low mix, high volume then start looking for stoppages and their causes. you can use the production reports or line monitor (depending on which programming software you are using on the Fuji's) these are pretty good tools to see whats happening on the floor. Assuming you didn't buy used equipment that is beat half to death, then the equipment shouldn't be a major problem. Fuji's are built like tanks and tend to be very reliable. So my guess at this point is that you may have some operational procedures that need to be streamlined. If you need a little more help with your Fuji's drop me a note offline and I'll see if I can help, I have a fairly strong background with their equipment.

reply »

#20367

performance improvement | 14 June, 2002

So, where do you want to focus our attention? What types of problems are you rejecting?

reply »

#20368

performance improvement | 14 June, 2002

Ken: At the end of arguing about how �myfloh� is calculating defect levels, reality sets in. �Between 50 and 250 ppm� is pretty darn good. Clearly, they have picked all the low hanging fruit.

reply »

Ken Bliss

#20376

performance improvement | 14 June, 2002

Dave, you are ahead of me again. I was thinking that this morning. however it clearly is a problem for him and I based the rest on my experience, anything can be improved but at what cost. If nothing else he can get more boards off his line at the end of the day and maybe improve quality a bit, admitantly a small bit. But if he can increse volume and maintain that level of quality his percentage of defects will go down and his cost per board would go way down, don't you agree?

reply »

#20377

performance improvement | 14 June, 2002

Hey Ken...my car won't start. Can your feeder racks or storage totes solve that problem as well? Sheesh....

reply »

#20379

performance improvement | 15 June, 2002

When picking fruit, the further you climb a tree, the more expensive it gets to pick that fruit.

reply »

Brian Doyle

#20385

performance improvement | 17 June, 2002

Guys,

Everyone who posts here works for a company. I get emails occaisionally from people who are not pleased with the way the forum is progressing lately. If someone has a problem and a person offers solutions why is that bad? If an answer is posted that helps the person who asked the question I think that's great. If it has a link to the company that provides the solution that's fine too. Its customer service in my opinion. Sure we get people who may abuse this by posting questions under dummy names and then just answering it themselves. If you feel that something is blatant advertising then email us at info@smtnet.com and we'll take a look at it. Perhaps we should make posting to the forum for registered users only (registration is free and takes about 60 seconds). That way there would be accountability. People have commercial problems and generally they can be handled with a commercial solution.

There are companies who have emailed me because they see posts about their equipment that is not factually correct. Wouldn't it be a great service to everyone who uses the forum if they came in and cleared that up? I'd think it would be. However many companies don't want to bother with that for fear of being flamed.

I see this forum as a great place where companies can interact with their end users. And that users can ask other users for unbiased opinions but still be offered with the facts straight from a service provider or manufacturer.

Do not drive the companies away. Use us to police the abusers, but invite companies to join in. I don't think you'll regret it.

reply »

kenbliss

#20394

performance improvement | 17 June, 2002

Dave, in theory I agree with you, but in the real world isn't the expensive fruit, the fruit you spend countless hours trying to fix a small defect that has a fixed cost, the cost of the board as a single unit. Versus the cost of running the factory. My point is as simple as the concept is that when quality is good the focus should be a maximizing uptime on the bottleneck machine which on the SMT line should always be the pick and place machine. I have only seen 3 plants in the last 10 years out of over a 100 that understood this. The down time from feeder exhaust is typeically about 3-5 minutes in just about every plant I have seen. You do this 20-40 times a day in a high mix or low mix environment and you are throwing away 1-2 hours a day or 10-30 percent of your lines capacity, this can be done so easily in about 20 seconds. I to see this as low hanging fruit but very few people get it. They all seem to think it is very hard to obtain. Xtra feeders are cheap if you can get the improvements, and you should be able to.

reply »

ianchan

#20400

performance improvement | 17 June, 2002

Hi mates,

ain't we all forgetting something?

this bloke who posted this thread, has not even specified what his "stability-variation" issue is all about? defect/quality variation? downtime variation? CPH variation? etc...

Just read this bloke's thread extract to ponder: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ I would like to have information about means to improve the stability of a SMD-Line: indeed the variation of the daily performance is too important (between 50 and 250 ppm). What is the solution to reduce these variations? climatisation, other reflow profil?... I'm waiting for your help and answers. Myfloh ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

So if we do not even know what this bloke's problem is, then what the heck are we all debating on???

Gentlemen in this thread I am for Dave F's question for most (if not all) posted questions to be explicit, and not face some ambiguity theory-based mumbo-jumbo shotgun scattering...

how you expect to meet someone's expectations if you do not even know what they expect? (like from the wife or boss?).

Come on Guys, the world is watching, lets' behave like the true professionals we all know each other to be, and respect each other as such.

Cheers!

reply »

#20402

performance improvement | 17 June, 2002

Ken,

Yes, when quality is good, increasing throughput of a bottlenecked resource [lowering costs] is a fine goal. [Wouldn't E Goldratt smile?]

In MY real world, quality is not as good as it should be. We have: * Suppliers that do not do as asked. * Employees that make mistakes. * Designers that think independently. * Bosses make unrealistic demands. * Bosses that accede to unrealistic demands from customers.

So, the fruit, in this real world, is the benefit of making a process improvement. Such a process improvement could be: * Increasing the throughput of a bottlenecked resource [as you say] � OR * Doing something so that you never have to see that small defect again.

When push comes to shove, making customers happy with the product they touch is more important than lowering costs, even if it increases costs. [EG would NOT be happy here.]

I will grant you that all things being equal, it would be difficult to imagine owning TOO MANY feeders. [EG's back to smiling.]

I stand by my original point, improving the quality of the product shipped costs more when improving from 1,000 dpmo to 500 dpmo than it does when improving from 10,000 dpmo to 5,000 dmpo.

reply »

myfloh

#20404

performance improvement | 18 June, 2002

Actually it is the number of defects which varies!!

This message was posted Add this forum to your site! Click to learn more. the Electronics Forum @

reply »

ianchan

#20405

performance improvement | 18 June, 2002

Hi mates,

finally! some sense to all this, so this bloke is concerned about "ppm = quality" variations...

looking back through all the muck in this thread, last posting before the s**t hit the ceiling was:

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Date: June 14, 2002 02:20 PM Author: Dave F Subject: performance improvement

So, where do you want to focus our attention? What types of problems are you rejecting? ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

so lets' rewind abit and start from here...

1) kenbliss: pls layoff the sales pitch for them carts, useful as they are (heard some good reports about TOC) they ain't relevant here in this thread, which is about quality ppm.

2) myfloh: so what do your pareto charts say as being the top defect(type) for this week? how about the defect types for past 6 week trend? you running HMLV? or you running LMHV? please share what is your current internal analysis and share it with us before anyone comments about "something".

reply »

Stephen

#20453

performance improvement | 20 June, 2002

If you don't already have one, the best investment you can make in your case is a good thermometer/humidity meter. They are not at all expensive. Then record the temperature and RH about 4 times a shift. After a few weeks (maybe sooner) you can see if there is a correlation with your PPM rate. There is debate about whether you need a reflow profile for each board or how many different ones you do need. If all your boards have the same PPM defect rate variety, then that is probably not your problem, if on the other hand you sort boards largest thermal mass to smallest and one or both extremes have the highest PPM defect rates, then there is your culprit. Another thing to consider is the actual measureing of your PPM defect rate. Last I heard people are 60-70% repeatable and AOI is 80-90% repeatable. You might sort your boards by colour of shininess of finish, and see if you can find any correlation. AOI might be finding more defects with some finishs more than others. Suspect all data. SMT magazine a few months ago had an article where they checked 1,000,000,000 joints. They found a reject rate of about 1,100 PPM. Everyone thinks they do better but the data didn't show it. Anyone with always better than 250PPM is missing a lot of defects, or is doing much better than almost everyone else.

reply »

Training Services

Electronics Equipment Consignment